One Architect, Three Approaches: Bertrand Goldberg’s Early Experiments with Prefabrication, 1937-1952

Bertrand Goldberg embraced an industrialized approach to building at a time when architects working in the United States were rarely involved in the construction process. Although he was certainly not the only architect to show an interest in prefabrication in the early twentieth century, it was an unusual pursuit since “design-build” was not yet a popular concept. The American Institute of Architects, according to Goldberg, highly discouraged architects from becoming involved in the construction of their designs for fear that it would lessen their attention to the design aesthetic in favor of profit making.¹ Thanks to his training both as an architect and engineer—Goldberg studied architecture at the Cambridge School of Landscape Architecture (Harvard) and at the Bauhaus with Mies van der Rohe in Germany, followed by further studies as an engineer at the Armour Institute (Illinois Institute of Technology)—he had a distinct advantage in understanding the ways to combine the design and structural aspects of prefabricated housing to create a streamlined final product without compromise. His interest in prefabrication was triggered by the changing needs of pre- and post-WWII society. Following the 1930s economic depression in the United States, architects could no longer rely on commissions to build private homes for the wealthy. Instead, they were forced to turn their attention to new kinds of projects: government-supported public works such as mass housing, hospitals, and schools.² Gone were the days of creating great works of architecture that did not contribute to a larger purpose. Post-war architecture needed to benefit society.

Upon starting his own office in Chicago in 1937, Goldberg decided that the best way to address the local economic crisis was to put people to work, not with their hands, but with machines. More specifically, through building prefabricated housing.³ Over the course of what he referred to as his “fifteen years of enthusiasm” for prefabrication, Goldberg developed two distinct approaches.⁴ The first was a method of assembly that included the manufacturing of panels and other small pieces that could be quickly constructed on site, as demonstrated in his Standard Houses (1937-1943). With the increasing demand for mass housing during and after WWII, Goldberg developed a second method of fabrication based on the lessons learned from his Standard Houses. The Standard Prefab Bathroom (1946-1947), Unicel Prefab Freight Car (1949-1952), and Unishelter (1952) exemplify this second method. Here the entire building module would be completed in a factory and shipped to site ready to use. Over the course of fifteen years, Goldberg refined the design and even his philosophy of what a prefabricated house represented. His experiments with assembly and fabrication ultimately led to a third, and more innovative, approach to evident in his later work—building with geometric modules.

ASSEMBLY: Standard Houses

In the early part of the twentieth century, there were not many government agencies that paid attention to the nation’s housing problem.⁵ For that reason, Goldberg began experimenting with prefabricated housing through a study conducted at Purdue University (1937). His design was part of a neighborhood of demonstration homes built in Lafayette, Indiana. The five-room model had two bedrooms, a living room, kitchen, and bath. The exterior was built of “prefabricated plywood sections,” covered by a pitched roof. Goldberg used techniques such as standardizing windows and wall panels, reducing interior partitions (such as the elimination of corridors between rooms) and millwork to save in construction costs.⁶ Considering Goldberg’s training at the Bauhaus, his first attempt in designing a prefabricated home was far from modern. The traditional features, such as the pitched roof, were simply a matter of economy. Flat-roofed homes, as he pointed out, would have been much more expensive to construct and did not provide the extra storage space that was possible with pitched roofs.⁷

After completing a successful prototype, Goldberg was inspired to continue experimenting with prefabrication on a slightly larger scale. At the time, real estate and government experts estimated that there was an immediate need for several hundred thousand new homes in the Chicago area. Goldberg partnered with his associate Gilmer Black, along with a lawyer Edwin Ashcraft III and real-estate developer Ross J. Beatty Jr., who had a lot of property but no buyers. The group agreed that prefabrication was exactly what an industrial city like Chicago needed to fill the housing gap. Together the partners founded the Standard Houses Corporation, which was dedicated to producing low-income housing in the $3,900 to $6,000 range using the newest materials and practices.⁸ Their first project was a model community based off of Goldberg’s prototype at Purdue University located on a small plot of land in Melrose Park, Illinois.

The Standard Houses at Melrose Park reflected the dream of creating an idyllic single-family home at an attainable price for working class families. Five one-story model homes were built, with slight variations to accommodate the needs of different families. There were five rooms (living room, kitchen, two-bedrooms, and a bathroom) that included amenities such as insulated hardwood floors, sliding doors, porcelain combination tubs and showers, heating, and hot water.⁹ Goldberg and Black placed a strong emphasis on the arrangement of the buildings on the site, staggering them at incremental distances away from the road to avoid the monotony of a “row-of-houses.” The model homes opened on September 24, 1939. Advertisements for the homes touted that the designs were a result of two and a half years of research, including Goldberg’s prototype for Purdue University, and represented some of the first low-cost housing in the Chicago area that was insured by the Federal Housing Association.¹⁰ The houses sold for $2,995, a price so low at the time that it drew crowds, and all five homes were purchased by the end of the day. With payments of less than $25 a month, the homes at Melrose Park were a true manifestation of affordable housing.¹¹

The Standard Houses Corporation’s first venture was a smashing success, creating a demand for more prefabricated homes just like it. However, the team was not properly equipped with the resources to scale-up the project. Goldberg admitted that the houses at Melrose Park were realized in a way that “could only be done once.”¹² Theoretically, the house could be reproduced with ease, but the team had not yet perfected the design to make the idea of prefabrication a reality. The shell of the first Standard House took a day and a quarter to be assembled, followed by another week of work by subcontractors to install the plumbing and electrical wiring. The project was fueled by the dedication and excitement of creating a new kind of housing, with considerable personal investment by Goldberg and his colleagues to pull the project off. Each of the four partners put in $1,000 of their own money to build the houses. They convinced suppliers to loan them the materials, extending the credit to 60 days after the houses were completed. The factory space and machinery used to build the five houses had been rented. The houses were priced so low that they brought in a profit of only $100 each. The Standard Houses Corporation, at this point, was more interested in demonstrating an idea rather than starting a housing business. To repeat the project a second time would have required a more serious investment.¹³

Despite Goldberg’s insistence that that the Standard Houses could not be replicated, he was able to continue experimenting with prefabrication when the Federal Works Agency sought out the Standard Houses Corporation to contribute to a housing development near the site of a defense industry in Indian Head, Maryland.¹⁴ The government offered the Standard Houses Corporation funding for fifty-seven houses at an average cost of $3,000 each. There were very few design guidelines provided by the government for the housing development. The project had to include a range of one- to three-bedroom homes, as well as the supporting utility systems. In addition, the Standard Houses Corporation had to prove that their proposed design could be dismantled, moved to a different site, and reconstructed within one day.¹⁵ Goldberg’s dedication to fully understanding the amount of pieces used in the assembly and disassembly process was meticulous. He personally drew over 1,100 pieces that were involved in building each house. The Standard Houses Corporation rented a local factory where they produced houses at the rate of three a day, with the hope of increasing output to ten houses a day.¹⁶ Because of the significant delay caused by subcontractor work on the Melrose Park, Illinois houses, the Indian Head houses were manufactured with the plumbing, wiring, and insulation already installed at the factory, so they could be assembled quickly on site.

After the success of the development in Indian Head, the Standard Houses Corporation won another government contract to build a community in Suitland, Maryland. The Suitland development was larger and more elaborate than the company’s previous projects, although the exact number of houses built remains unclear. Goldberg moved to Richmond, Virginia to design and oversee a new factory for the Suitland housing development. The layout of the factory was designed to encourage the workflow of prefabricated components, with separate departments for plumbing, electrical, painting, lumber milling, and assembly. The new factory greatly increased the output of the Standard Houses, and was able to manufacture the components (including wall panels, pre-assembled bathrooms, and kitchens) for ten houses a day. The finished pieces were then shipped to Maryland and assembled on site. Always looking to improve upon earlier models, Goldberg teamed up with Eero Saarinen to do color studies for the homes. The studies added a new dimension of variation in the homes, as the exteriors were painted in different shades of brown, grey, and white, and had blue chimneys. However, the interior of the homes remained rather conventional.

With the establishment of a factory in Richmond dedicated to producing Goldberg’s designs, his role in overseeing the Standard Houses began to seriously blur the line between manufacturer and architect. At times Goldberg mentioned that he felt more like a factory supervisor than an architect. It was a distinction that worried him, and caused him to wonder if his industrialized experiments could really be called architecture. In 1942, while Goldberg was meeting with the Virginia Pulp and Paper Company, a potential client, about a new project to support their Richmond factory, one of the paper factories exploded and fatally injured twenty people.¹⁷ Goldberg was one of only two people to survive the accident, and it took him over a year to recover. He continued the development in Suitland, Maryland while recovering, but the accident brought an end to the Standard Houses Corporation and signaled a new direction in his career.

FABRICATION: Standard Prefab Bathroom, Unicel, and Unishelter

Shortly after recovering, Goldberg enlisted in the Office of Strategic Services as an opportunity to try something different: to break away from the more traditional design of the Standard Houses by designing new products in entirely new contexts.¹⁸ After several years of working with a prefabricated panel assembly system, Goldberg determined that he had the wrong approach because the process required too much field labor.¹⁹ He began working on an alternative approach to prefabrication, one in which the final product could be fabricated completely in the factory and then shipped to site. Even better, the unit could first be used as a shipping container, holding furniture, building supplies, or any other necessary materials. His initial tests with this method included a Mobile Delousing Unit (1942), a Mobile Penicillin Lab (1943), and a Convertible Gun Crate (1943). The Mobile Delousing Unit and Penicillin Lab became unnecessary because of advancements in medicine, and therefore the two designs were never built. However, about 250 to 500 of the Convertible Gun Crates were manufactured for the Army to deliver arms to site, after which the crate could be converted into a bunk or field office.²⁰

With the end of WWII in 1945, Goldberg’s concerns began to shift toward urban architectural problems. He was convinced that “architecture for the individual was absolutely wrong, socially wrong, and that the only way in which architecture could develop would be through industrialization of components.”²¹ He realized that there remained many urban houses that did not have indoor plumbing at the time.²² A single plumbing appliance could be a cost-effective and easily installed addition to modernize an older home. Goldberg had already produced preassembled bathrooms for the Standard Houses made out of separate pieces of plumbing, but had never attempted to condense the amenities of an entire bathroom into a single, streamlined unit. He raised the equivalent of a half-million dollars at the time to start a new company and manufacturing plant in Chicago, the Standard Fabrication Corporation, and began working on a design for a prefabricated bathroom in 1946. The engineered unit saved space for homes where extra square footage was a precious commodity, and drove down the cost of purchasing and installing separate pieces of equipment. For the price of $495, a family could have a bathtub, shower, washbasin, toilet, and storage.²³ Testing estimated that the unit would save at least two days of labor per installation, allowing for 70% more plumbing units to be installed in homes each week. The Standard Prefab Bathroom was designed for maximum living and minimum care, and as Goldberg proudly pointed out, had been called the “first new thing in bathroom plumbing since the Romans.”²⁴

Despite tackling some of a homeowner’s biggest concerns, Goldberg did not foresee the hurdles he would face in distributing the unit. There was the problem of getting approval for more than 1,500 different plumbing codes in the US and working with various unions. The bathroom units veered from the standard manufacturing practices and required different training to install, even though its four connections—a sewer connection through the floor, a vent through the ceiling, and hot and cold water hookups—were supposed to simplify the installation process. About 2,000 of the units were produced and sold to Montgomery Ward and other local distributors.²⁵ Desperate to sell the project, Goldberg approached John Snyder, president of the Pressed Steel Car Company, to produce the Standard Prefab Bathroom on a larger scale. Snyder declined, but was impressed by Goldberg’s work. Instead, Snyder proposed that Goldberg design a new freight car built out of plywood as an alternative to steel, in response to the shortage of metal available after the war.²⁶ He thought that the freight car industry had fallen behind innovations of other manufacturing sectors and that it was time for an update.

The Unicel Prefab Freight Car was the newest freight car to be designed in half a century by the Pressed Steel Car Company, and was to be produced exclusively in Chicago.²⁷ The design was revolutionary because it was made of laminated super-strength plywood that could be bonded together to form a structural tube. The freight car’s name, Unicel, was a reference to the single, cellular structure of the freight car. Plywood roof and side panels laminated to wood ribbing spaced on 16” centers formed the core of the freight car, which was finished with curved laminated corners.²⁸ Extensive testing went into proving the merits of Unicel. Two years of engineering research, completed with the help of the structural engineer Clarence Plisky, showed that the material was less expensive, lighter in weight, and performed better than steel. In 1950, the product was unveiled to over 800 of the nation’s shipping and railroad executives in New York at the Waldorf Astoria, as well as at the Merchandise Mart in Chicago.²⁹ Unicel garnered significant interest and looked like it was going to be a successful venture. The final hurdle was to gain the approval of the Association of American Railroads (AAR) before Unicel could be used for general railroad service. In April 1951, the AAR technical committee tested the performance of the freight car in more than 32,000 miles of travel and found no structural failures.³⁰ However, following pressure from the AAR executive committee (in which a majority of the members were associated with the steel industry), the design was not endorsed.³¹

Somewhat disheartened by the failure of Unicel to break into the railroad industry, the Pressed Steel Car Company still saw potential in the product they had been developing. Goldberg, no doubt influenced by his earlier work, decided that the freight car could be repurposed as prefabricated housing.³² The new product, Unishelter, was essentially the same design as Unicel, but with the addition of interior partition walls, windows, and doors. Because of the rigidity of the building module, Unishelter could be fabricated at the factory with all of the furniture included inside and then shipped to site ready to use. Goldberg thought of the building module as a large brick that could be arranged in different configurations to serve different purposes.³³ The basic Unishelter module included one bedroom, one bathroom, kitchen, and living area. A second module included two bedrooms that could be combined with the basic module to create a three-bedroom home, or even a five-bedroom home out of three modules. The Unishelter could be adapted for civic and commercial needs as well, with suggested configurations designed to best fit each use. Shopping centers could be arranged like an accordion, hospitals in the shape of an “H,” and schools in the shape of an “E.” In theory, an entire town of Unishelters could be erected overnight since the design required little on site work other than adding a minimal foundation and joining the modules in the desired configuration.

The Pressed Steel Car Company found the perfect outlet to test the newly-repurposed Unishelter design when, in 1952, the Relocatable Defense Housing (RDH) sponsored a study to determine the feasibility of using high-quality re-locatable housing as an alternative to temporary housing in defense areas. The designs had to be comparable in quality, livability, and cost to fixed site housing, and meet the standards prescribed by the Housing and Home Finance Agency (HHFA) research division, the Public Housing Administration, and the Federal Housing Administration. Only eight out of more than thirty-five submissions were chosen by the RDH task force to be built as prototypes. Unishelter was among the eight chosen.

The testing for the prototypes consisted of taking each house from its original foundation, loading and hauling the house through an urban area, and installing the house in new location. Six of the designs were variations on the prefabricated panel system, including a folding (hinged) panel house and another that used a basic frame and curtain wall construction. Only one house could be shipped in its entirety, which awkwardly took up the width of the road and required a special driving permit to relocate the house. Thanks to Goldberg’s years of experimenting, the Unishelter prototype was a unique and more refined design. The Unishelter could be moved 100 miles, from start to finish, in 122 hours. At the price of $7,700 for three modules, which could be relocated for 8.1% of the original cost, Unishelter was predicted to make a “sizeable splash” in the low-cost housing market.

The HHFA picked up the Unishelter for a larger-scale test community in 1953 and built 258 houses at Camp Stewart, a Georgia Army post.³⁴ The units were manufactured at the Pressed Steel Car Co. in Chicago, shipped to Georgia by rail, and assembled on site in groups of three to form a H-shaped unit. In addition to the Georgia Army post, the construction of 100 prefabricated homes was supposedly underway in Nevada, along with three other planned locations for Unishelter communities. A severe cut in the HHFA’s budget for relocatable housing in mid-1953 brought an uncertain future for the Pressed Steel Car Company’s success. Up until this point, the Unishelter had only been designed for the Army, and had not been tested for a wider market.

According to Goldberg, John Snyder decided to build an upscale version of a Unishelter home on his property in Shelter Island, New York to encourage a non-military application for the Unishelter. Completed in 1952, the Snyder Residence included seven Unishelter modules arranged around a central living space, also designed by Goldberg. One of the modules functioned as a kitchen, one as a utility room and maid’s quarters, one as a recreation space, and the rest as bedrooms. The guest bedroom was the most dramatic feature of the home—cantilevered thirty feet over the ocean, to demonstrate the enormous strength of the prefabricated plywood construction.³⁵ The modules were completely furnished with all bathrooms, kitchen equipment, plumbing, heating, air conditioning, and furniture installed at the factory in Illinois. They were then shipped by rail, boat, and truck to demonstrate their maneuverability. The Snyder Residence used rich, heavier materials for the portion of the house built on site, such as an oversized stone fireplace and mahogany floors for the living room, and a lighter mahogany cladding on the Unishelter modules. The combination proved that standardized building modules could be transformed into a luxury home. Snyder predicted that “similar wheel-less railroad coaches [would soon be] sticking their vestibules into the water at every seaside resort in America.”³⁶

For Goldberg, the Snyder Residence was a turning point that caused him to give up on designing single-family homes and focus exclusively on multi-family housing. Although he regarded the Snyder Residence as his greatest triumph in his career up to that point, he became disenchanted with the possibilities of what prefabricated single-family housing could achieve. Under Snyder’s direction, Goldberg saw a future where the Unishelter would be repeated in subdivisions and resorts across the United States. In the process of devoting his attention to perfecting the design of prefabricated housing, he had left behind social and community planning aspects that were equally important to him. With the realization that each year the US government was spending more than $2,000 per family to develop suburban life and less than $100 per family in cities, Goldberg turned his attention toward reinvigorating urban life by building larger scale housing developments.³⁷ For too long, he built what the government, bankers, and factories were willing to pay for, without considering how the projects aligned with his own vision. Perhaps because of Goldberg’s disinterest in continuing the project and lower than expected sales, John Snyder closed the Unicel/Unishelter freight car plant in 1954.³⁸ The Snyder Residence was the last single family home to be built by Goldberg.³⁹

GEOMETRIC MODULES: Goldberg’s Later Works

Goldberg’s approach to prefabrication, in many respects, was far ahead of the times, and fell to the wayside due to a lack of acceptance by the industry and the public in general. Yet, by the time public opinion caught up with the wave of prefabrication, Goldberg had moved on to pioneer other ideas in housing. He was able to distill the lessons learned from his prefabricated single-family houses and apply them to a new kind of social housing. With Drexel Home and Gardens (1954-1955), Marina City (1959-1967), and Raymond Hilliard Homes (1963-1966), Goldberg no longer thought in terms of individually fabricated modules that could be grouped together, but instead in terms of geometric modules. Goldberg’s geometric modules were different from individual modules because they shared a single adjacent wall. This removed the redundancy created when two individually prefabricated modules overlapped, such as the double-wall created when two Unishelter modules were joined together. The geometric modules were more than a way to minimize redundancy in material; they offered new structural opportunities, as well. Much like the bent plywood tube that comprised the basic structure and enclosure for the Unicel/Unishelter freight cars, Goldberg thought of each geometric module as an enormous brick that could be combined in an infinite number of ways. He pioneered the use of geometric modules with a single shared CMU wall in the Drexel Home and Gardens, then through the radial column grid of Marina City, and finally through the Hilliard Homes where each unit became a structural module.

The Drexel Home and Gardens apartment complexes, a low-cost public housing development in Chicago, were the most straightforward (and least innovative) interpretation of the module as an enormous brick. The housing blocks were made of rectangular units stacked on top of each other: four on the bottom, four on top. The plan for Marina City, arguably Goldberg’s best-known work, was made up of 16 repetitions of the same module arranged in a circular form. Although the modules appear to serve a structural function, in reality the core of the building provided much of the support. It was the Raymond Hilliard Homes that demonstrated the greater variation and flexibility possible when the modules themselves provided structural support. Free from relying on the core, the “shell structure” of the Hilliard Homes allowed for more playful arrangements of modules grouped for high-density living. Goldberg explored different social benefits associated with each configuration. The circular towers of the Hilliard Homes were thoughtfully orientated for elderly living. The direct relationship of each unit to one another, with a shared common space in the center, forced residents to check in on their neighbors and encouraged community relationships. The slightly curved form of the housing towers for low-income residents were more suited for the needs of a family because they provided more space. The shape of two towers had the added benefit of looking down onto a “Tot Lot,” a place where children could have the freedom to play outside while still being within the range of supervision of their parents from the units above.

In a 1973 interview that reflected on Goldberg’s earlier work, the interviewer, John Cook, suggested “the repetition of the same unit over and over again in a housing area offers little opportunity for variety.” While Goldberg acknowledged that prefabricated housing could be monotonous, it was more of a self-critique of his earlier work than of modular building in general. He refuted Cook’s position by pointing out “there were different possibilities of combining the units…industrialization in architecture means almost limitless possibilities.”⁴⁰ And therein lies what was perhaps one of Goldberg’s biggest strengths as an architect: his ability to look past the stigma associated with standardized housing and to explore the limitless possibilities inherent in prefabrication. His initial “fifteen years of enthusiasm” for prefabrication, and continued interest in repetition in his multi-family housing, resulted in a truly revolutionary approach to urban living.