INTRODUCTION

In most of the developed world, a split divides architectural practice, with activities related to design and those related to site construction separated for legal and economic reasons.² The result is that the values of the two territories also remain isolated: architects' concerns regarding aesthetics, materiality, or the impact of intangibles such as natural light are not ones which most contractors or manufacturers have been trained to address. Further, important communication between architects and their counterparts, characterized primarily by written and drawn documents, do not effectively convey the significance of these issues. Similarly, the contractor and trades hold concerns which are not communicated to the architect: staging, safety, ease of fabrication, or the changing economic climate are indicated only crudely, through pricing. On larger projects, Western architects attempt to address this split through the use of consultants and the accretion of personal experience – imperfect mechanisms in a volatile and increasingly sophisticated context.

Furthermore, while architects and fabricators or builders struggle to communicate their goals through imperfect means, the act of construction itself offers its own challenges. Alan Day notes in his recent book, Digital Building, that "The two key characteristics of the building industry are fragmentation and uncertainty"³ Uncertainty is a natural result of not only the complexity of contemporary construction processes and architectural equipment, but also unknown or unpredictable factors such as soil performance, weather, and the variable results of crafts such as concrete work or welding. Under such circumstances, "…it is impossible to make formal contracts that cover all problems and difficulties that might come up. Instead, technological cooperation must, to a large extent, be founded on mutual trust and expectations that unforeseeable problems will be solved in a cooperative spirit."⁴ Just such a cooperative spirit is found within Japanese practice.

Collaboration in design and construction, where the construction and design development phases are essentially collapsed, is not universal in Japan. However, it is recognized as a highly effective instrument, and it is common for leading architects and contractors to support using this approach for the design development of innovative structures. Design consultants are recorded as speaking of Japanese construction with rare affection. Cesar Pelli said simply "there is no more satisfying experience than building in Japan."⁵ Rafael Viñoly is quoted as saying that in the Japanese system “works incredibly well” and that "…design professionals and construction managers cannot isolate themselves from the process..."⁶ Kathryn Findlay, whose practice is based in Japan and the United Kingdom, describes her early experiences in Japan with warmth, saying, "When I first began working on the construction site, the process was relatively smooth, [and I thought] Japan’s contractors more accommodating, compared to England, where even a small building required detailed, difficult drawings to be completed… [In Japan] we’re able to get substantive changes implemented on the site, after construction has begun…"⁷ And Tadao Ando, reflecting on his recent experiences building in Chicago, and preparations for the addition to the Kimball Museum, is quoted as saying, " After all, Japan is blessed to be small and geographically isolated, I think. In architecture, too. The design consultant and the contractor work collaboratively, so it is easy. The contractor says ‘leave this to us’ about some area, and the architect is happy to do so."⁸ The leading Japanese architects and engineers I have interviewed almost universally declare that they have yet to construct a satisfactory building outside their home country.

These comments allude to some of the key factors which permit successful collaborations to occur, and result in the innovative and technologically sophisticated work Japan is known for. As Viñoly notes, architects are not isolated from construction – thus, architectural staff recognize and respond to construction-related problems. Similarly, contractors, and by extension subcontractors and manufacturers, learn the issues important to architects and thus can assist the designer in finding ways to achieve these ends. Findlay reinforces this point; the contractor responds to the architect. Additionally, as she notes, the process is more flexible and an architect is not confined to torturously drawn media. Rather, through models and countless hours of discussion, the architect and their counterparts cooperatively engage in design development. Further, as she also notes, there is flexibility during construction as well. Finally, Ando not only points out that designated design development is often encouraged, but also begins to suggest a reason that collaborations are successful. However, I would argue that it is not so much geographic isolation but rather economic isolation and the Japanese oligopolic system which is a key stimulus, as I will discuss.

In collaboration, there are clear benefits, as these architects indicate, but there are some very real disadvantages as well, and I will discuss these, too.

In fact, the line between architect and contractor can grow extremely blurry: in one project I researched during 1998, many of the architect’s site-based staff were actually on loan from the project’s contractors. In another case, a retired employee of a large contractor acted as an advisor to an architectural firm during a major project’s basic design. Although there were no guarantees that the contractor would be awarded the job, this was ultimately the case, perhaps because the former employee passed information back to his former employer. In the meantime, the architect’s staff received valuable advice as they worked out how to execute a challenging structural system. Such liminal staff are also found in Japanese manufacturing alliances, where the term "guest engineer" has been coined. As Jeffery Liker, an observer researching Japanese production methods, pointed out, these individuals serve as a "…two-way conduit of information flow…[and] may be able to influence specifications for the subsystems" in a way which benefits the contractor.¹¹

Most architects confess that they do not initially have a clear sense of the opportunities for overall collaboration on any given project -- even with contractors they are familiar with -- until supervising staff have had some experience working together on site. Architect-contractor interactions, rather than being defined as they are in the West, by a sense of what is normative or by legal constraints, are worked out over the initial months that design and construction staff work on site.¹² The benefit of this approach is that it allows the individuals directly involved in a building’s production to develop an understanding of the biases and experience held by their counterparts. Site-based staff, over time, grow to see themselves as part of a single organization, and because this organization represents a wide range of expertise, it becomes easier to, as a group, propose and work out innovations.¹³ By sharing experience across trades, the team is able to reduce uncertainty and innovative ideas can be more quickly developed, because one has access to technically accurate and reliable resources.¹⁴ As the expertise reflected by these teams expands, the pool of available information also expands, and important knowledge is more readily accessed. This, in essence, lies at the root of Japanese customization and innovation, not the remarkable manufacturing capacity available.

Jeffery Liker further noted activity in the Japanese auto industry, similar to the construction industry’s designated design development, which he dubbed "black box sourcing." As he outlined it, the practice "…has three characteristics; early involvement of suppliers, clear communication of the customer’s design requirements, and extensive design-related responsibility assumed by the suppliers."¹⁵ Elsewhere he noted that, "Japanese automotive customers often initiate the design process by giving the supplier a general conceptual description of the part of subsystem. This general description is followed by intense joint activity by the customer’s and supplier’s engineering staff."¹⁶

Takahiro Fujimoto found a similar situation, but placed it in a more complex context. In 1993 he surveyed parts suppliers to see how orders were received. For completely new car models, 18 per cent of parts were bid, based on detailed drawings provided by the car maker; 49 per cent involved development competition among two or more suppliers, based on looser specifications supplied by the car maker; and 32 per cent involved selection of a particular supplier, which then participated in product concept, planning and specification development.¹⁷ This reflects what I observed, although the percentages would be quite a bit different if all materials used in a building’s construction were counted. Nonetheless, while there were certainly times when building materials were ordered in a manner Western architects would consider conventional, by bidding from detailed drawings, significant areas of the building, particularly finishes, critical subsystems, and structural materials, were developed in partnership with one or more manufacturers.

Most design professionals in Japan told me that customization was not an automatic solution when faced with a problem or envisioning a new approach. Several argued that customization was a necessary, but time-consuming and expensive, response to a lack of available building materials, especially those of high quality. Additionally, off-the-shelf components offered greater reliability, and even without professional liability concerns, architects naturally favored this. In some cases, though, an architect’s own actions seemed to contradict this statement, as for example, when the staff in Fumihiko Maki’s office investigated new white tile glazes, even though a range of whites were already available. But an extensive search of available materials can also yield ones which can be used as a basis for customization, and manufacturers better prepared to respond to architects’ demands.

I found that collaborative partnerships with manufacturers were of two types. For a limited number of materials, those which were either essential to the overall success of the building design or representing previously untried technologies, cooperative design development would actually begin during the basic design stage, long before construction began. This also allowed the manufacturer to get government approvals when necessary; permission to use special materials takes six months and costs ¥800,000. But for a broader palette of building components, selection was more likely to occur after the construction/design development phase had been initiated. In the latter case, it was not uncommon for contracts to be offered to more than one supplier, for reasons I will discuss below.

Occasionally, the manufacturers made direct proposals to the architects they thought might be sympathetic to working out the details of a new material or use. Toyo Ito, for example, noted that he had been approached about working with an aluminum manufacturer, with the invitation to develop an all-aluminum structure for a suitable residential project. The development of another new material was initiated through connections to an author frequently published by the manufacturer’ s publishing arm, which led to him being informed about products in the development stage.

But generally, in the case of innovative materials or subsystems, architects sought out suitable manufacturers. The mechanisms used to find a cooperative manufacturer included everything from cold calls on up. Introductions were reported from materials brokers, other small manufacturers, engineers and even family friends. There seemed to be no consistent pattern, except that contractors were never used to find these manufacturers, in part because no contractor was as yet committed to the project. This sort of search is unusual in the United States; Kunio Watanabe, a renowned structural engineer, told me that when he made cold calls to pre-cast plants in the area of a Princeton University project, he was repeatedly met with surprise. American architects and engineers, working isolated from manufacturers, probably would have less confidence in the value of such a trip. Furthermore, a North American design professional would probably assume that suppliers were scattered, some at a considerable distance from the site.

Some materials suppliers are widely recognized as being supportive of innovation, and if an architect has a new idea for materials produced by these companies, partnership can be established quickly. However, most architects told me that they would have an idea and approach numerous manufacturers about working together, only to be turned away. Nonetheless, experience can assist in identifying likely partners: in some cases, particularly for customizations of crafted fabrications on publicly-funded work, the architect is required to make a visit to the plant to inspect material prior to shipping. Manufacturers inclined to collaborate use these visits as an opportunity to present examples of other special-order work they have produced, and to explain in some detail plant capability.

In either case, architects and manufacturers began working together prior to bids being put out on a project, and with no assurance that the manufacturer would ultimately receive the order. One design professional did make the point to me, though, that pre-bid customization was best limited to projects where there was greater confidence that they would ultimately be constructed, e.g., funding was already in place. Still, this serves as no guarantee that a building would be built, and some customizations are dropped in design development. Why would manufacturers offer support at this stage, when there was no contract and a great deal of effort might be required, potentially without remuneration?

While most manufacturers find such collaboration an expensive nuisance, some see it as product research, which can not only lead to new markets, but increase the manufacturer’s prestige. Products may even be offered for free, on the assumption that a leading architect can be an effective way to introduce a new material – as was the case in a heat-absorbing ceramic wall I saw completed (which was installed for free on a government-funded project), or the all-aluminum structure mentioned earlier. One advantage offered is that manufacturers work with the architect to develop specifications outlining production capacity (not product specifications) for these materials. Obviously, these specifications will include, if not favor, the advising manufacturer. Additionally, manufacturers may feel that offering advice gives them an edge when a supplier is selected, both because the manufacturer has greater technical information and considerable lead time, and also because the architect will want to continue the partnership during construction. It is far more common for an architect to jettison later customizations in an effort to maintain funding for those collaborations which have been in place for a while and appear to be successful. In the publicly-funded projects I observed, for example, budgets had been cut across the board by 15 percent, but customizations initiated earlier continued. Contractors, though, will often try to shift customization to suppliers who are more inexpensive, or suppliers having a long-term relationship to the contractor. At the very least, this can whittle down the scope of work available to the original manufacturer.

Innovations occurring while still at the basic design stage are limited to only the most crucial elements, and much component customization occurs after construction has already begun. Such adjustments are less significant than those noted above, but nonetheless account for the unique and refined characters of Japan’s best architecture. Here, contractors have a more active role in selecting manufacturers, and architects, who hold right of refusal, sometimes complain about the choices offered them. However, even under this system, manufacturers remain under pressure. Contractors and architects tended to employ an interesting mechanism to assure that each manufacturer remained competitive: production was contracted out to two or more manufacturers, with the understanding that if any one producer proved unsatisfying, they could be dropped from the project with little difficulty. Both contractors and architects frequently referred to this strategy. Additionally, the advantages of one fabricator would be used as a goad to others. Thus, for example, one producer might be able to make long sheets of steel, but not thin ones, while another could work with thinner steel. The architect and contractor would note a competitor plant’s strengths, suggesting that, if the manufacturer could not meet similar standards, work could be shifted to the competitor. And this message would be communicated to both producers, in the hopes of getting thinner and longer steel sheet. (I should note that through the mid-1990s, this often led manufacturers to invest in unnecessary equipment, contributing to the excess capacity available today.) Similarly, when multiple contractors, subcontractors, or manufacturers are involved in a project, the architect may begin by proposing a challenging idea to the counterpart most capable of its production, then indicate to competitors that the lead firm has begun working on the problem, thus bringing competitors on board.

As T.J. Pempel, a noted economist, summarized the situation, "Ultimately, power was concentrated…Contractors typically divided orders among several subcontractors to gain multiple alternative sources…They could shift contracts among subcontractors as rewards or punishments for past loyalties, current pricing problems, quality performances, or personal whim."¹⁸

Manufacturers respond to these pressures by bringing production staff into discussions right from the start. I frequently witnessed early-stage, on-site consultations involving both a sales representative and someone from production. The sales staff were responsible for keeping work within a target price, determined by manufacturer and contractor during initial discussions. The production staff were responsible for advising the architect and sales staff as to how the concept could be achieved within this price. If the architect’s demands seemed simple, or the discussion did not require production advice, further consultation might only involve the sales staff. But on those projects where initial consultations required extensive involvement on the part of production staff, this allowed the fabricator to concurrently design the component and its manufacture, by advising the architect regarding capability. These teams have a big impact on architects: one project architect made a point that even in selecting materials such as projectors and other non-customized equipment, she felt a preference for companies which had sent out the most knowledgeable sales staff.

Nonetheless, such strategies are expensive, and probably at least partially account for estimates that 87 per cent of all construction products are priced higher in Japan.¹⁹ While the overall ratio of imported to domestic materials is probably being affected by the current recession, most construction materials used in Japan continue to be domestically produced. Keiretsu ties between firms, both vertical and horizontal, can also maintain domestic demand, but it is also true that various barriers to trade have maintained an oligopic structure in the marketplace. I discuss the impacts of oligopoly, both positive and negative, below.

In the collaborations I observed, architects and their counterparts would initially come to a verbal agreement about their goals, without any contractual obligations. At this stage, the process was essentially exploratory; issues under consideration might include whether an idea was feasible, whether the manufacturer could produce the component, what materials or shapes would be most appropriate for fabrication, etc. In many cases, even on well-supported sites by leading architects, some of the ideas floated would later be dropped. The reasons for this varied: an architect decided that an aluminum floor was costly and wanted to use the moneys elsewhere, for example, while a contractor decided that a sophisticated process lifting steel floor panels into place was unwarranted. I did not find that abandoning a proffered idea affected the overall quality of the collaboration. Nevertheless, I can imagine that, where a reasonable level of support has been offered to the architect – lengthy exploratory discussions are the norm – and an idea then ultimately abandoned, this has the potential to affect cooperation. Furthermore, as architects are aware of the limited number of manufacturers willing to customize, they may be reluctant to jeopardize future opportunities to work together.

One salesperson, representing a large furniture manufacturer, told me he sees architects as being of two types. Some “use paper,” while others meet to communicate their needs. He described the use of written material as easier, but made the point that his company could then more quickly decide to refuse to put in a bid. With verbally negotiated work, the prices were generally higher, he said, but there was opportunity to remove production barriers and avoid excessive expense. But in Japan, drawings are not seen as fixed instruments. I was amused to hear a contractor, when pressuring an architect to make revisions to a set of documents, lecture his colleague that the original bid documents were not a “Bible.”

In Western architectural practice, contractual written and drawn materials are privileged, while verbal communications are often referred to as being “not worth the paper they are printed on” – the implication being that since they are not printed and verifiable, they have no value. As challenging material and building component ideas are often investigated prior to a contractual commitment in Japan, all parties do work verbally, perhaps with preliminary drawings or models. Verbal communications are considered essentially trustworthy, especially while the form or execution of an idea is still being developed. If a contractor says something is possible, or estimates that it will cost a certain amount, this is treated as relatively reliable information.

In this way, architects are able to test how various strategies, related to a specific component or subsystem, might impact cost, aesthetics, maintenance or other issues – without becoming invested in a particular approach. Similarly, contractors or manufacturers can begin to ascertain how committed the architect is to an idea, and which factors initially outlined are flexible. I repeatedly heard people note to each other: "this point is very important, but I can make changes to these and these." In one case, for example, the architect had a very clear idea about the color of a material he intended to develop, one not available on the market. Initially, he also hoped to use a relatively large size, but as it became clear to the manufacturer that size was not as important, the supplier switched the plant where production would take place, to one which could not work with larger materials but seemed more capable of achieving the desired color. Both plants made prototypes for the architect’s review, at different stages in the development process. The plants, incidentally, were ostensibly independent, but sold most of their product to one of the giants in the field. This relationship gave them little choice when assigned customizations, and put pressure on them to respond successfully.

Jeffery Liker has researched how teams function in the Japanese automotive industry, and he refers to them "being flexible within clear boundaries."²⁰ For architects who have a clear formal agenda (thinner, more transparent, etc.) or easily articulated goals, it is relatively simple for those they work with to begin to determine how these intentions affect the piece under discussion. In this case, the contractor, subcontractor or manufacturer will be able to offer suitable guidance on available materials or forms that can comply with the architect’s overall intentions. The architect’s goals do not have to be communicated verbally, but when one of the functions of negotiation is to determine boundaries, the back and forth of conversation can be quick and effective.

In the West, quality is defined by specifications, which outline performance. However, in collaborative development of new materials or assemblies, the final outcome remains unknown, and architects must define the production capacity of the plant instead. (The tolerances and norms which have been accepted as standards are also crucial to this flexibility, as they set maximum variations in measurements and other criteria often written into Western specifications.)

After writing a specification that broadly outlines the capabilities of suppliers, architects will approve one or more manufacturers from a list proposed by the contractor. Over the course of 1998, many related to me that offshore production is becoming an impediment to verbal negotiation, although it is certainly possible with some materials, such as steel and aluminum panels, to spilt collaboration and production. On major components, where the architect felt collaboration was crucial and would be affected by off-shoring, they simply dug in their heels and refused to approve overseas production plants. But some have found that when Japanese plants off-shore only some activities, such as shop drawing production, it can affect modifications initiated later in the collaboration. This is not always clear unless a problem arises. Then designers may discover that collaboration during production is not an option. For architects who rely on negotiation and advice over written material, this actually offers a genuine challenge.

Verbal communication, as I noted above, is usually accompanied by the use of models and sketches. One architect, for instance, told me that the final form of a cast steel connector was first established in a full-scale, rickety model her office produced, and the steel plant sent a back a neater model for confirmation. Upon approval, this model was used to make molds for the part. Drawings were secondary to the models, and lacking in detail.

On particularly complex or important components, architects, contractors or even governmental regulations may also require mock-ups or prototypes. These offer an opportunity to check actual production against the more flexible instruments employed in collaboration, and may even replace drawings. (I was amused by a mechanical engineer who said to the team of subcontractors, “If it doesn’t look like we can building this as drawn, we need a mock-up, right? His proposal, however, fell on deaf ears.) Often the designers’ expectations can be significantly different than the first mock-up produced. In the case of an architect who intended to use precisely formed pre-cast concrete columns as an aesthetic baseline in a project, for example, chamfered corners, evident on the first mock-up, became a significant point of contention between architect and fabricator. The architect assumed that details drawn with squared corners would be understood as requiring higher care in production, while the production plant had assumed this was a simplification, and that easy-to-fabricate chamfered corners were perfectly acceptable.

Thus the prototype offers a way to unearth misconceptions, a natural part of any construction experience. In particular, contractors rely on them as a communication tool; when two or more manufacturers supply materials, the mock-up can be used to confirm that suppliers produce comparable materials. Manufacturers also indicate that they appreciate the opportunity to confirm particulars. And when mock-ups are located on site, architects sometimes attach models to them, trying out ideas related to later stages of construction.

Various Japanese words are used to discuss these prototypes, and each suggests differences in fabrication complexity, expense and purpose. Three of the most common nouns used are jikken, mokku’uppu and shisaku. Jikken literally means “experiment” and is generally applied to simple full- or large- scale models which function like the proposed component. One example I saw was a foam-core board model of a lighting grille, with a strip of small lights behind, which the architectural staff waited until after dark to observe in a context similar to proposed use. The mokku’uppu is a test construction and, as far as possible, actual materials are employed; this involves both the manufacturer and relevant trades. In some cases, simpler models will precede the mock-up; this depends on the relative level of customization and whether the mock-up is intended to confirm ideas or demonstrate execution. Finally, shisaku refers to a test piece, sometimes manufactured with variations in connections and configuration, to allow the architect and manufacturer to work out details. Although it would seem possible to maintain a sample, which could be used in this way, the test pieces I saw were left on site, suggesting they were specially made.

Mock-ups and test pieces are considered "free", but designers noted that there were costs. A woman designing custom fabrics made a point that weaving designs are supported only where high-end raw materials were used, thus assuring the manufacturer a higher mark-up. And when one engineer, as part of a group, told me that the mock-ups produced by three manufacturers (to assist in selecting which one would fabricate a combined curtain wall and structural assembly) were free, his colleague responded “somebody pays!” And all present agreed that the costs associated with customization were spread over the manufacturer’s other work.

Liker also raises the concept of "time windows," periods when changes are acceptable and accommodated.²¹ He noted that, although suppliers were continuously fine-tuning their production technologies, there were specific periods where these new technologies were introduced to automobile manufacturers. Although the timing differs, this roughly correlates with what I witnessed in architecture and construction.

In most cases, openness to modifications in a material or subsystem declined over the life of the collaboration. Final changes, often no more than fine-tuning, were accepted only through the period shortly after review of the final prototype, just as the material was to go into production. One can determine the length of the collaboration by working backwards: the point of installation is based on the overall construction scheduling, and manufacturers need time to fabricate the component to meet this deadline. Simple customizations, such as doors and windows, generally took only one month, but more complex ones, such as the tube-like structure for Sendai Mediatheque, were planned to take more than a year; in this case there is continued modification during fabrication and construction, a point I will return to below.

As the parameters for customization are defined, they close off areas for potential change, narrowing the range. In general, the architect and their counterparts will come to agreement, but occasionally the architect will hold fast to an approach the contractor or manufacturer advises against. (In this case, the architect is sometimes asked to guarantee the relevant construction.) However, reviews – whether signaled by the prototype or by shop drawing approval – allow some opportunity to re-assess the customization. If problems are identified, or reasonable arguments emerge, major changes can be effected around this time. The parties initiating the change, though, are frequently warned that this will create delays and expense. But if egregious problems are identified, even during installation, some modification may be attempted in the field. To effectively negotiate a major change late in collaboration, it is best to have something to offer to compensate. In many cases, simplifying fabrication or assembly is used as a lever, but I have also heard architects prepare for announcing a major change by notifying the contractor that they have decided to give up some other piece in the building, one which would have required extra attention.

Sometimes, the scale or volume of customized work will mean that it is produced over a long period of time. In this case, all parties become aware of small changes that could make fabrication or assembly easier or safer. I did see situations where work in the field led an architect, client or contractor to suggest that a modification might be made. Rarely did this go beyond a simple question of whether change was possible. However, manufacturers indicated to me that they were continuing to work out process technologies during fabrication, and this was reflected in some changes to the final building materials, so that variation in the product run might be apparent. In these sorts of cases, the initial production run may still have some problems, and a wise architect is one who locates the earliest use or installation of custom work in inconspicuous locations.

Architectural staff regularly indicate when they are involved in working out detailing or material selection for specific areas of a building, and invite the contractors’, subcontractors’, or manufacturers’ input. Thus, they initiate the opening of a “time window.” Contractors or manufacturers would also alert architects, and architects remind their counterparts, of a target date for concluding adjustments, based on the construction schedule and approval deadlines. (Generally, approvals were sequenced to include subcontractors, contractors, architects and occasionally clients’ representatives, especially in public work.) It would appear that a cushion was built into the scheduling of most approvals. Frequently there was some slippage in deadlines, as all parties struggled to keep up with a tremendous load. On one 34,000 square meter site, an architectural staff of thirty (many inexperienced), were responsible for approving 86 items over a five-day period, virtually all involving at least simple customization. These approvals included the measurements, configuration, colors or other important decisions about everything from window sizes and frames, signage, furnishings, and plantings, as well as selecting paint colors or stone. The period before or after was similarly crowded, and the architects had a severe backlog of materials where the approval deadline had passed.

In response, architects will wisely attempt to limit the areas where they take advantage of customization, and when earlier stages in the collaboration have yielded a satisfactory result, there is little temptation to tinker further on any but the most critical pieces.

Around the time the structure is going up, or shortly thereafter, the architect will begin to fine-tune and make modifications to doors and windows, finish materials and fixtures. This may become a point where struggles emerge. The architect can inadvertently or unthinkingly create construction delays either by making copious modifications, even if small, or by letting too many approval deadlines pass. The flurry of changes can also cause problems between architects and clients. Public clients, responding to recent corruption scandals, resist changes which create expense or those which bring in new suppliers. In at least two cases, both at roughly this point in the sequence, I witnessed clients attempting to put a moratorium on changes, while the architects saw their work as only half done, potentially yielding a hodgepodge of refined and unrefined components. But flexibility, so much a part of many interactions on the site, exists here as well: when told there would be no further changes, one architect joked that he had only four changes left and would have to use them well.

Neoclassical economists argue that government protection for oligopolies reduces pressure to modernize, to engage in research and development, or to improve the quality of materials produced. In short, the expectation is that market demand forces manufacturers to behave competitively. The Japanese governmental has established incentive programs intended to counter these tendencies, by promoting the introduction of new technologies, upgraded facilities, and higher levels of research and development than might otherwise exist.²² As a result, the technological sophistication in declining or highly competitive industries is relatively high.

In my discussion of the interactions between architects and fabricators, I outlined how competition is introduced into the system, by ordering from multiple suppliers. However, as Liker notes, "…suppliers would not be willing to invest the amount required, give in to customers’ demands for aggressive target prices, and respond to the pressures of aggressive development cycles if the customers could choose from a large group of competitors at any time."²³

Most economists agree that in Japan’s basic materials markets, especially those crucial to the construction industry (including concrete, steel, aluminum, ceramics, and pane glass), there are clear oligopolies, with at least 90 per cent of the market dominated by five or fewer firms.²⁴ Furthermore, many of the larger firms are linked to “cooperative” subcontractors, who make up much of the "competition" by smaller firms. In the case of a small craft tile company I visited, for example, it supplied over 90 per cent of its product to Inax, one of Japan’s largest tile producers. In this way, Inax fielded out less profitable production, while standardized tile shapes were made in a large, fully automated plant. It is the fact that both contractors and many supply sectors are oligopolies, with a limited number of dominant agents directing demand, which keeps this system in balance. This also allows architects and manufacturers to develop smooth, co-operative relationships over time.

Customization is particularly found in declining industries, where the producers as a whole are in a relatively weak position. Demand for many of the materials associated with construction has been affected by the bursting of Japan’s speculative “Bubble” and the consequent drop in construction orders, and also by a drop in domestic automobile production, which has reduced demands for steel and glass. (As I write this, the Financial Times reports that construction orders dropped 21.3 per cent in November 1998, the 11^th consecutive month orders dropped.²⁵) Although it is beyond the scope of this discussion, I found that customization was more evident in declining -- and thus diversifying -- industries, and almost non-existent in industries such as plastics, which were economically more viable and had a broader customer base. While architects had more mixed feelings about collusion, which could limit the scope of work or even cause clients to decide not to build, several engineers told me that they thought oligopolies and dango a good thing.

As far as customization is concerned, weak demand and excess production capacity are a good combination. Often manufacturers took on the risk associated with customization because they wanted to keep employees working, and were reluctant to lose them to layoffs. I found that customizing plants had a mixture of FMS equipment and skilled crafters. They hoped that customization could be achieved within the current production process (and for minor changes it generally could), but the presence of under-utilized skilled labor offered insurance that even more tricky problems could be resolved.

One of the most significant disadvantages of the kind of collaboration I witnessed was that it was extraordinarily time-consuming, redundant, and thus costly. On one site, each week roughly thirty people – representing the architects, structural and mechanical engineers, contractors, and various subcontractors – would sit down to a meeting which routinely took eight hours. While those who were not directly concerned with the work at hand might occasionally wander off to take a call, most remained in the meeting regardless of whether they were directly involved in the discussion at hand – in part because there was an expectation that their input might be needed. The involvement of large numbers of staff is directly related to the fact that the problem is not tightly bounded, and most observers consider this inherent in the process of collaboration. This architect, though, made some effort to contain the amount of customization, and the architectural staff devoted to it; in at least one case, I had the genuine sense that the architects, in pursuing many opportunities for customization, had allowed themselves to be overwhelmed by the process.

While construction has never been considered to have high productivity, this sort of collaboration will certainly work against contemporary efforts geared at increasing efficiency. And although some observers still claim that the Japanese corporation continues to rate customer satisfaction as more important than profitability, the current economy has certainly rendered this a luxury.²⁷ As Liker notes  …it is increasingly recognized in Japan that a customer’s focus can lead to unreasonable demands and sacrifice on the part of employees..."²⁸ For many manufacturers and contractors, where prestige is less directly tied to innovation, the high costs of collaboration, such as the stack of A1 drawings more than 8" deep produced by the supplier of blinds for one site, make customization simply too expensive. Thus has been aggravated by recent legal shifts and increased liability.

Paul Herbig, another researcher of Japanese management practices, observed, "The Japanese innovation generation process has an unusually high cost associated with it in terms of the generation of problems and solutions, [a] high degree of social interaction…human exhaustion and overwork…mental exhaustion, and burnout."²⁹ The hours of effort I saw architectural staff put into a project struck me as nothing short of Herculean, and certainly entailed great personal sacrifice.

Furthermore, most recognize that innovations cannot be thoroughly tested within the context of Japan’s rapid construction. As a pragmatic response to that, regulations prevent customization of many mechanisms, especially those that hold weight, and of electrical supply materials.

Stable relationships among architects, contractors and manufacturers have their own set of efficiencies, though, in that the goals and abilities of each actor are better understood by their counterparts. In Japan, I am regularly impressed with the sensitivity to architectural issues articulated by contractors’ employees, and by the awareness of production constraints expressed by architects. Yet at the same time, architectural staff often had no applicable experience which could be relied on in addressing the customization at hand. Rather than attempting to develop a comprehensive body of knowledge, architects’ efforts were contrived to work effectively at identifying and accessing useful information held by their counterparts.

In this way, architects were able to propose and execute innovative construction, often in many of the major subsystems of a building, simultaneously. Leonard H. Lynn, in a study on innovation, calls this the "garbage can model” of innovation: rather than a clearly established problem-solving sequence, the process allows for "…a confluence of changing streams of problems, solutions, participants, and change opportunities.”³¹ As a result, the architectural staff I observed generally seemed more confident, more secure, and more willing to take risks, because they saw the entire construction team as committed to innovation, rather than feeling pressure to personally embody all necessary knowledge.

Thus, architects in Japan will often propose the use of entirely new technologies or materials with complete – and to the outsider, audacious – confidence. One project which I observed, the Saitama Arena, includes a curved, 15,000-ton moving wall, which can be adjusted to allow the arena to take on various configurations. Several years ago the American partners on the project (who were primarily involved in basic design) stated, "Probably the most unique thing about this building is the money being spent on technology that really doesn’t exist…Technology is being developed especially for this building."³² Today, the necessary technology exists.

During the year, I met with and interviewed the following professionals (architects, unless otherwise noted) on the dates given. (Persons who are office staff or contractor’s staff and were interviewed as a part of site visits are not listed.)

• Kazuyo Sejima (January 29, December 15)
• Hajime Yatsuka /UPM (February 4, July 8)
• Toyo Ito (August 14, December 9)
• Tei’ichi Takahashi/Dai’ichi Kobo (February 13, November 16, December 10)
• Yasufumi Iseki, Engineer, Kajima Construction (February 24)
• Kojiro Kitayama (February 25)
• Ryoji Suzuki (February 27)
• Shigeru Ban/ Shigeru Ban Architects (March 2)
• Gary Kamemoto /Maki office (March 2)
• Fumihiko Maki (February 26, December 10)
• Makoto Yokomizu Toyo Ito Architects (March 9)
• Riken Yamamoto (March 11, July 31)
• Kunihiko Hayakawa / Kunihiko Hayakawa Architect (March 20)
• Yoko Kinoshita and Makoto Watanabe/ADH (April 14)
• Jun Aoki /Jun Aoki & Associates (April 17)
• Toshitaka Kawamura/Dai’ichi Kobo (May 21)
• Kisho Kurokawa (May 25)
• Astrid Klein/Klein Dytham, Tokyo (July 1)
• Kazuo Kishi/ ISO 9000 series, JQA (July 27)
• Tadahiko Okumura and Osamu Inoue/Engineering, and Shinji Matsumoto, Head of Institute of Technology/Shimizu Construction (July 28)
• Yasuo Tadano/Nikken Sekkei (July 30 and September 3)
• Yasufumi Iseki and Toyoaki Naito /Kajima Construction (July 30)
• JSCA members: Masato Araya, Yoshiharu Kanebanko, Toru Kobori, Kunisuke Kusugawa, and others (September 9, November 12)
• Toshiyuki Yamada, engineer/Shimizu, Kinya Ando, engineer/Fujita Corporation, and Hiroshi Furukawa/AKI Structural Design Engineers(October 4)
• Kunio Watanabe, engineer/SDG (December 18)

Visits to plants and construction sites (generally all-day affairs) were conducted on the following days.

Particularly helpful manufacturers and consultants (not an exhaustive list) met on site came from the following firms:

• Kumiko Taguchi, custom carpet design, Saitama Nursing College
• Akane Yamazaki, A Works, custom signage and furniture coordination, Saitama Nursing College
• Shimadzu, Fixtures and furniture, Saitama Nursing College
• Tachikawa Blinds, Saitama Nursing College
• Hiromura Design Office, Custom furnishings, Saitama Nursing College
• ACC Engineering, DPG system, Saitama Nursing College
• Dekoratekku Japan, Hisai bus stop
• 3M, Hisai bus stop
• NKK, Hisai bus stop
• ToTeku, Snow Research Center
• Inax, Nagaoka Folly
• Armec, Nagaoka Folly
• Figla Co, Ltd, Nagaoka Folly
• Mutsuaki Saskai, Structural engineer, Sendai Mediatheque
• Masamitu Ikeda, Ikeda Architectural Research, Structural engineer, Sendai Mediatheque

Footnotes: