Experimental Residential Design: The Five Houses Project

Written and reviewed by Rice Design Alliance Editorial Team.

In 2011—three years after artist Ai Weiwei and the Swiss practice HHF conceived their Five Houses project—a Kunsthaus Bregenz exhibition record listed its location as “still unknown.” Yet HHF’s own project page now carries metadata placing it in Long Island, United States. That unresolved gap between record and reality is part of what makes the project worth studying. Experimental residential design works best as research-by-design: a disciplined way to test spatial, social, and environmental ideas before, or even without, pouring a foundation. Five Houses, conceived as a component of the broader Ai Weiwei House concept, proposes five related residences generated by repeating and varying a single well-proportioned unit. Reading that proposal alongside built and speculative U.S. precedents reveals how experimental house plans and avant-garde home design survive contact with empirical adaptability research and American building codes.

What Is the Five Houses Project in Architecture?

The Five Houses Project is a 2008 architectural proposal and exhibition concept by artist Ai Weiwei and the Swiss architecture practice HHF (Herlach Hartmann Frommenwiler). It explores repetition with variation across five named residences—Low, Secret, Introverted, Theater, and View—developed within the larger Ai Weiwei House project. Its built status is unverified, and its site record remains unresolved.

That distinction matters. This is not a confirmed completed residence with a documented client, budget, and certificate of occupancy. HHF’s project record describes it as an architectural proposal, and the office’s selected-projects portfolio labels it “Project, 2008,” “Ongoing, 2008–,” and “Houses and exhibition.” Those are the labels of a conceptual housing design, not a finished home.

Project Facts & Disambiguation

• Authors: Ai Weiwei with HHF (Herlach Hartmann Frommenwiler)
• Date: 2008 design; catalogued as ongoing from 2008
• Program (per HHF metadata): “Exhibition and Series of Houses”
• Location: HHF’s current metadata lists Long Island, United States; the Kunsthaus Bregenz press record from the 2011 exhibition lists the location as “still unknown”
• Unverified: final construction cost, dimensions, permit history, and built status

The honest way to reconcile those sources is to treat the Long Island reference as current office metadata rather than independent proof of construction. Until primary documentation surfaces, the project’s physical realization stays an open question.

One more caution: the name travels. Nommo Arquitetos’ Five Houses (2020) is an entirely separate built project in Curitiba, Brazil, arranging five units on a corner lot with varying plans, yards, cutouts, and skylights. And in Trondheim, Norway, Nøysom Arkitekter developed a five-house row at Svartlamon through a participatory self-build process. Same naming neighborhood, three unrelated projects. The Ai Weiwei–HHF proposal is the subject here.

Spatial Architecture Concepts Behind the Five Named Houses

Strip away the celebrity authorship and what remains is a rigorously legible parti—the concise organizing diagram that drives a design. HHF’s rule is disarmingly simple: take one well-proportioned unit and repeat it sequentially, letting controlled variations in orientation, enclosure, and adjacency generate five distinct houses. Architects call this repetition with variation, and it’s closer to musical structure than to industrial modularity. The unit is the theme; each house is a variation that never abandons the theme’s proportions.

The project’s analytical vocabulary comes from two crossed spatial ideas. HHF describes perpendicular “axes of privacy” running against a contrasting primary “axis of generosity.” Translated into plan-reading terms: each house organizes its sheltered, inward-facing rooms along one directional line, while a perpendicular connective line opens the houses toward shared space, landscape, and one another. Privacy isn’t a wall; it’s a direction. Generosity isn’t a room; it’s an axis that stitches separate dwellings into a composition.

The concept also extends below the scale of architecture. HHF describes modular tables designed to combine into a single long entity running through each house, so the repeated-unit logic continues into furniture and daily inhabitation. This is where the broader Ai Weiwei House framing becomes relevant: as a Gesamtkunstwerk—a “total work of art” integrating architecture, objects, and use—the proposal treats the dining table and the site plan as expressions of the same idea.

Low, Secret, and Introverted Houses: Enclosure and Domestic Separation

Read through HHF’s naming and unit-variation brief, the first three variations work the privacy end of the spectrum. Low House implies a profile pressed toward the ground, trading vertical drama for shelter and a compressed horizon. Secret House suggests a manipulated threshold and orientation where arrival reveals little and the repeated unit turns inward. Introverted House pushes furthest along the axes of privacy, organizing domestic life around an internal focus rather than outward views. Together they illustrate how rotating and re-thresholding a single module can calibrate domestic separation.

Theater and View Houses: Framing Landscape and Collective Openness

Theater and View Houses reverse that charge in HHF’s sequence. As their titles suggest, Theater House recasts the repeated unit as a domestic stage aligned with the connective spine of the axis of generosity, while View House directs the module toward outward sightlines and landscape. Where the first three concepts thicken thresholds, these two dissolve them. Across all five names, the proposal outlines a public-private gradient—from maximum enclosure to maximum exposure—generated by varying a single repeated unit.

Positioning Five Houses Within Modern Residential Typologies

To see what Five Houses is and isn’t doing, it helps to place it inside a formal typology framework. The University of Texas at Austin’s 5X9 Urban Housing Typologies primer catalogues 45 examples of socially relevant urban housing across nine organizational categories—linear block, mat housing, cluster, transparent block, and others. Five Houses doesn’t sit neatly in any one of them. Its serial repetition of a unit echoes the linear block’s logic, while its inward-turned variations behave like a cluster. It’s best understood as a typological hybrid: a row that behaves, at moments, like a courtyard field.

The more instructive comparison is methodological. One Colombian urban housing case study in the 5X9 primer, La Playa, begins with a basic 5 × 10.5 meter unit that can be subdivided in multiple ways to produce different bedroom configurations. That’s the same wager Five Houses makes—that a disciplined, repeatable module supports spatial diversity rather than suppressing it. The difference is intent: La Playa’s unit serves affordability and flexibility in dense urban housing, while HHF’s unit serves an exhibition argument about domesticity itself. Same tool, different research question.

The natural American touchstone is the Case Study House Program, launched in 1945 by Arts & Architecture magazine to explore efficient modern homes using new materials and industrial techniques. According to the Charles & Ray Eames Foundation, Case Study House #8—the Eames House—was designed in 1945 and completed in 1949: a built, inhabited prototype. The Los Angeles Conservancy counts 35 published Case Study plans, of which 25 houses and one apartment building were actually constructed in California and Arizona.

That built/unbuilt split is the key contrast. The Case Study program was a postwar industrial project—its experimental house plans existed to demonstrate production-ready ideas, and most of its designs were meant to be constructed. Five Houses is a conceptual and exhibition project; its value doesn’t depend on construction at all. Both use repetition and prototype logic, but one tested manufacturing optimism while the other tests how far a single spatial idea can stretch across five temperaments of dwelling. Confusing the two genres leads to bad criticism—judging an exhibition proposal by production metrics, or a production prototype by conceptual purity.

Innovative Spatial Planning: Privacy Gradients, Section, and User Data

Five Houses organizes privacy horizontally, through plan. Contemporary practice offers a useful counterpoint: organizing it vertically, through section, and socially, through data.

MOS Architects’ House No. 22 (designed for a private site whose location is withheld at the client’s request) inverts the conventional domestic stack. Collective spaces sit above ground while more private rooms occupy the level below, arranged around a sunken courtyard animated by nine skylights that double as solar chimneys. The void does three jobs at once—it admits daylight into the buried level, drives natural ventilation, and creates a buffered outdoor room that belongs to the private zone. It’s a sectional answer to the same question Five Houses asks in plan: how do you graduate openness within one dwelling?

The MOS project archive also documents Housing No. 16 in Asunción, Paraguay, which replaces existing dwellings for their current residents. The office recorded household information—ages, genders, family ties, disabilities, illnesses—and let that data shape unit planning. This is experimental housing whose variable isn’t geometry at all; it’s the precision with which space fits specific human lives.

A built American comparator completes the triangle. AIA New York’s profile of Green House in Bridgehampton, New York (designed by Roger Ferris + Partners, 2013), describes two interlocking rectangular volumes: glazed social spaces on the first floor, more enclosed private rooms clad in recycled fiber-cement panels above. The public-private gradient is legible from the street—transparent where the house hosts, opaque where it shelters.

Read together, these three projects and Five Houses map a toolkit for unconventional layouts: horizontal axes (HHF), sectional stacking with a courtyard void (MOS House No. 22), resident-specific data (MOS Housing No. 16), and interlocking glazed-and-solid volumes (Green House). None relies on eccentric form for its own sake. Each makes privacy a designed sequence rather than an afterthought—and each shows how courtyards, voids, and level changes can deliver light, air, and separation simultaneously, which is precisely what an open plan alone cannot.

Evaluating Avant-Garde Home Design: Performance, Adaptability, and Lifecycle Tradeoffs

Avant-garde home design earns its keep only if it performs beyond the photograph. A rigorous evaluation separates three things that are often blurred: verification status (is it built?), user model (who shaped it?), and environmental strategy (what does the form actually do?). The AIA Framework for Design Excellence offers ten principles for exactly this kind of screening; three are especially sharp here. Economy asks whether first cost and operational cost are honest. Resources interrogates material choices over their lifespan. Change asks the hardest question for sculptural homes: can this plan adapt when the family does?

Open plans and double-height volumes—the signatures of much radical domestic architecture—deserve particular skepticism under that lens. Double-height spaces complicate heating and cooling stratification, acoustic control, and future subdivision. Fully open plans erase the door openings and defined traffic zones that make later reconfiguration cheap. None of this argues against drama; it argues for drama that survives a second owner.

Comparative Matrix: Exhibition, Speculative, and Built Case Studies

The matrix exposes a spectrum: high-tech engineered performance (Green House) versus low-tech participatory performance (Svartlamon), with speculative and exhibition work supplying the ideas each extreme later absorbs.

The Chalmers Adaptability Test for Experimental Floor Plans

Researchers at Chalmers University of Technology analyzed 3,624 contemporary Swedish apartment kitchens and identified the spatial variables most associated with adaptive capacity: windows, door openings, traffic zones, service shafts, and room typology. Although derived from Swedish multi-residential apartments rather than U.S. single-family homes—and paired with 2024 Chalmers circular housing design research warning that experimental layouts often neglect end-user preferences—the spatial test translates cleanly to custom house plans. Are shafts positioned so wet rooms can move? Do circulation zones eat the plan or liberate it? Does the fenestration rhythm permit repartitioning? A concept that fails these questions is sculpture; one that passes is adaptable housing.

Translating Conceptual Housing Into U.S. Site and Code Reality

Admiring a proposal like Five Houses is easy. The professional question is what happens when a client in, say, Bridgehampton or Chicago wants its spatial intelligence in a permitted house. The answer is a feasibility sequence, not a leap of faith.

1. Extract the principle, not the form. Identify what the concept actually delivers—a privacy gradient, a repeatable module, a courtyard section—and treat that as the design brief. R-3 House shows how a speculative project can carry passive strategies and a clear parti on an ordinary quarter-acre lot without pretending to be a built artifact.
2. Confirm the regulatory baseline. The 2024 International Residential Code covers one- and two-family dwellings and townhouses up to three stories with separate means of egress. If your concept fits that scope, the IRC is your baseline model code; outside it, you’re in different regulatory territory.
3. Verify local adoption and amendments. The IRC is a model, not the law itself. States and municipalities adopt it—often with energy amendments, fire and life-safety modifications, and local egress interpretations. The applicable code is the one your jurisdiction actually adopted, in the edition it adopted.
4. Screen the site before the plan. Zoning governs setbacks, height, lot coverage, and use; none of that appears in a conceptual drawing. A house organized around outward view axes dies on a lot where the only view is the neighbor’s wall.
5. Stress-test the unusual moves early. Double-height spaces, sunken courtyards, and glazed social floors each trigger specific structural, energy, drainage, and safety reviews. Price them and detail them before falling in love with them.
6. Run the adaptability and lifecycle screen. Apply the Chalmers variables and the AIA’s Economy, Resources, and Change principles to the working drawings, while changes are still cheap.

Can avant-garde designs meet U.S. codes? Routinely—Green House is proof that interlocking volumes and aggressive glazing can clear permitting in New York. The trick is that code compliance is achieved in the translation, not imported with the concept. Because adoption and zoning vary by jurisdiction, treat every municipal claim as something to verify locally rather than assume.

Conclusion: Lasting Lessons from the Five Houses Proposal

The Five Houses proposal matters precisely because it may never need to exist as buildings. As a spatial laboratory, it demonstrates that one well-proportioned unit, repeated with disciplined variation, can generate an entire spectrum of domestic character—from the inward hush of Secret House to the exposed stage of Theater House—along crossed axes of privacy and generosity. The second lesson is methodological: pair that formal intelligence with empirical adaptability criteria and honest lifecycle evaluation, or the experiment ends at the rendering. The third is jurisdictional: in the United States, conceptual ambition becomes enduring architecture only through the discipline of code, site, and verification. For your own work, the next step is concrete—take one concept you admire, extract its organizing principle, and run it through the feasibility checklist above before drawing a single wall.