Start a conversation on the subject of ‘flexibility’ with architects and you will inevitably find yourself caught up in a lexical field that is as rich as it is confusing. Is it really flexibility we are talking about? Or rather adaptability? Or perhaps versatility? In fact, isn’t it rather reversibility that is at stake? And what about modularity? And isn’t all this ultimately linked to prefabrication?

In reality, these different terms apply to different operations and maintain a hierarchy between them. They can concern multiple spatial scales and occur together or separately, at highly variable temporal rhythms. The book Flexible Housing, by Jeremy Till and Tatjana Schneider, is a valuable resource on the subject, providing some clarity on the possible definitions and permutations of ‘flexibility’ in the field of architecture, particularly in housing, where this concept has found fertile ground for its development.¹ The term is currently making a strong comeback, offering a salutary promise in the face of the rapid and untenable obsolescence of our built heritage. 1 Jeremy Till and Tatjana Schneider, Flexible Housing (Elsevier, 2007).

But what exactly are we talking about? Till and Schneider propose to understand flexibility, in the field of housing, as the ability of housing to adapt to changing needs and contexts, whether before, during, or after the building is occupied. Adaptability is also understood by these authors as an acceptable synonym. Other authors, such as Stewart Brand in his curious How Buildings Learn, go in the same direction by evoking the idea of ‘scenario-buffered buildings’, buildings that can withstand any scenario that might have been thought of in advance.² Brand also points out, not without irony, that buildings did not wait for architects to put into practice their adaptability, which most often occurs under the pressure of local contingencies: ‘Almost no buildings adapt well. They’re designed not to adapt; also budgeted and financed not to, constructed not to, administered not to, maintained not to, regulated and taxed not to, even remodelled not to. But all buildings adapt anyway, however poorly, because the uses in and around them are constantly changing.’³ All these authors agree on the major challenge of flexibility: integrating the notion of time into architectural design, a challenge in this discipline where objects are most often considered – and shown – as static. 2 Stewart Brand, How Buildings Learn: What Happens After They’re Built (Penguin, 1994). 3 Brand, How Buildings Learn, 2.

These definitions nevertheless come with a distinction: among all these terms with which our architects, still busy conversing, are getting tangled up, there emerge, on the one hand, strategies (versatility, reversibility) and, on the other, tools (the use of modularity and prefabrication). This classification – which, due to the sprawling nature of the concept, is not fixed – varies according to the scales at which flexibility is sought, whether an entire building, several housing units or several rooms within a building. Attempting to create the conditions for potential future adaptability imperatively demands anticipation and foresight in basic typo-morphological and constructional choices.

Hardware and software 

An initial design strategy for implementing flexibility in a (new) building, intended for housing, consists in creating an envelope that will facilitate the subsequent transformation of the building’s spatial structure and/or the housing units that compose it. This strategy, which is deployed at the building level, is directly linked to questions about its overall geometry and constructional principles. How thick should it be? What ceiling height? What structural articulation? What type of envelope and what kind of integration for the vertical and horizontal technical flows? For what future life and programmes?

French architect Anne Démians has attempted to provide a universal response to these challenges in what she has called the Immeuble à destination indéterminée  (IDI, the building with an indeterminate purpose). The IDI is a superstructure that can accommodate collective housing, offices or a hotel, as its articulation is designed to respond to all the points raised in the above questions. This approach involves a conceptual separation between what could be called the hardware and software of the building; the former relates to spatial and structural proposals, studied in advance, to enable the latter, which is reversible, to evolve.

These reflections on hardware and software bring to mind the famous experiment carried out in Brussels by the Simone and Lucien Kroll studio at La Mémé in Woluwe-Saint-Lambert (Brussels) in the 1970s: students who were then moving into its large open-plan spaces were invited to change the location of the partitions, which were specially designed to be moved around as desired, independently of the structural elements and in a way that was completely reversible. This example reflects more broadly an era of emulation – which we are still heirs to today – around the idea that ‘superstructure’ (or hardware) and ‘filling elements’ (or software) can be thought of along different timelines.

Two recent examples illustrate this concept of reversibility on a smaller scale: the Tina building by architects Sozia (Breisach, 2022) and the SWCS building in Charleroi by architects Goffart-Polomé – Reservoir A – Meta. In both cases, the architects set up / relied on a relatively unrestrictive structural articulation (open plan with posts and beams) and a generic envelope that can accommodate a range of configurations and divisions of the plan. Service areas (distribution and sanitary facilities) are grouped together and/or pushed to the edges of the open plan.

A comparison between the IDI and these two examples shows that we must remain mindful of the pitfalls of an architectural response that is impervious to the local specificities of a given project. As shown by the contextual project of Goffart-Polomé – Reservoir A – Meta, it is quite possible to integrate the hardware/software strategy without this leading to a generic and universal architectural formula.

Architectural noughts and crosses

Another strategy, which can complement the first, is to propose stable spatial configurations, designed, dimensioned and configured to allow flexibility of use, in a given space and over a long period of time. It is then the layout, size and distribution of the rooms that will allow uses to vary and follow one another. Before becoming a subject of experimentation for architects, this logic was that of many vernacular dwellings, and in Brussels, the one we have inherited in our terraced houses, whose almost infinite adaptability was highlighted by Gérald Ledent and Alessandro Porotto in their book Brussels Housing.⁴ Applied to new buildings, this strategy requires the acceptance of a certain indeterminacy of spaces, or even their unfinished nature, when it is up to the resident not only to occupy them, but sometimes to complete them as they see fit. This ‘box to fill’ strategy can be found in its ‘stable’ form – as in Bruno Taut’s Hufeisensiedlung (1933), where all rooms are equipped with a water supply, making it possible to change their use – or later, in a more unexpected form, in the famous Quinta Monroy (2003) designed by the firm Elemental, which literally offers appropriable voids. 4 Gérald Ledent and Alessandro Porotto, Brussels Housing: Atlas of Residential Types (Birkhaüser, 2023).

The same strategy recurs in yet another form in many explorations of the combinatorial possibilities of identical square rooms with interchangeable uses, whether in the work of Andrea Palladio (Villa Rotonda, 1571), Jean-Nicolas-Louis Durand (Précis de leçons d’architecture, 1802) or more recently in the recurring obsession with the ‘nine square grid house’ by John Hejduk, Shigeru Ban Pezo von Ellrichshausen, and, in Belgium, Office Kersten Geers David Van Severen with their iconic Villa Van Buggenhout (2012).

Another interesting example is the reflection carried out in Brussels by the architects of Multiple, a2o and Koen Van Synghel on the conversion of the Proximus towers into housing. (Fig. 11) Using typical office hardware, the proposal explores housing typologies with generic rooms (largely inspired by the architects Peris+Toral). The result is flexible, compact (no corridors) and economical flats, hence their nickname, ‘Colruyt flats’. As we can see here, these different variations on architectural noughts and crosses, or boxes to be filled, yield housing that accommodates a degree of indeterminacy: for Till and Schneider, these configurations, based on a permanent spatial structure, are the most sustainable, because not only do they leave some decision-making to the users, but also because the variations that run through them do not require any materials to be moved.

More recently, the value of these forms of indeterminacy has been highlighted from a feminist perspective. Among other voices advocating for a gendered approach to housing, the Angela D. collective has published a practical guide which states:  ‘Modularity in housing allows it to adapt to the different stages and phases of life, particularly changes in household composition: single-parent or blended families, single women, etc. Making housing flexible means giving residents control over their living space and encourages their autonomy and independence. A living room can become a bedroom and vice versa.’⁵A form of equivalence between rooms, and their de-functionalization, makes it possible to tend towards these needs. 5 Projet d’habitat collectif, Angela.D dans CALICO (CAre and LIving in COmmunity), Une approche féministe du logement: Guide pratique (2018), 50.

The door of Marcel D.

Flexibility has also been explored in relation to interior spaces: the tools used in this regard consist of mobile, removable, pivoting elements, etc. Like the door of Marcel Duchamp (which must be moved from one frame to another to rearrange the space), flexibility is achieved through the physical alteration – whether slight or not, reversible or not – or the relocation of an architectural element. This strategy, which applies on a smaller scale, is based on a certain confidence in technical innovation, a confidence that accompanied the beginning of the twentieth century and research into minimum housing: the smaller a home is, the more it is necessary to provide for overlapping or interchangeable uses. From then on, experiments with sliding and pivoting walls, removable furniture and mobile floors multiplied, sometimes leading, paradoxically, to an overdetermination of spaces; the silent film The Scarecrow (1920), contemporary with these experiments, hilariously shows how this way of life can quickly become hell.⁶ 6 The Scarecrow, directed by Edward F. Cline and Buster Keaton (Metro Pictures, 1920).

Nevertheless, there is no shortage of good examples: from the ‘Swiss Army knife’-type movable walls of Gerrit Rietveld’s famous Schröder House to Steven Holl’s homes in Fukuoka (a reinterpretation of Japanese sliding Shoji), numerous projects show that flexibility can also be achieved with simple, domestic means. More recently, Sophie Delhay’s innovative research and Juliane Greb’s radical experiments have effectively combined generic rooms and large sliding panels. Their Unité(s) and San Riemo projects offer malleable configurations, making numerous socio-spatial aggregations possible.

In the meantime, our architects have not finished their conversation and will continue to debate ‘flexibility’ for a long time to come. They do have their doubts: most of the examples put forward are an accumulation of paper architectures and prototypes that have rarely gone beyond the one-off stage. Why so? Isn’t flexibility just an ‘architect’s idea’? Does it really meet the desires of residents, or is it just a specialist’s fantasy? What seems to be at stake, however, and what makes the concept so beautiful, is the idea of seeking and welcoming a healthy dose of uncertainty and humility in architecture, as Herman Hertzberger already invited us to do: ‘Flexibility means – since there is no single solution that is preferable to all others – the absolute denial of a fixed, clear-cut standpoint. The flexible plan starts out from the certainty that the correct solution does not exist, because the problem requiring solution is in a permanent state of flux, i.e. it is always temporary.’⁷ 7 Herman Hertzberger, Lessons for Students in Architecture (010 Publishers, 2016); see in particular the chapter titled ‘Functionality, Flexibility and Polyvalence’ (146–49).

The profusion of the lexical field of flexibility and the confusion it creates therefore reflects a rich concept that has permeated the history of architecture. It seems that even today, true to its history of heroic but isolated experiments, ‘flexible housing’ appears and disappears on our architectural horizons, like the Loch Ness monster, a distant fantasy, with undulating appearances and an uncertain existence.