Abstract
Histories of housing commonly explain architectural differentiation through settlement, climate, class, privacy, gender, household structure, and technological development. All of these matter, and none of them is set aside here. Yet they share a common starting point, the waking uses of buildings. This article asks what becomes visible when the same history is reconstructed from the several hours each day during which inhabitants largely cease using buildings in the conventional sense and instead depend upon them most completely. Domestic architecture, on this reading, is the changing multimaterial coordination of a recurrent biological fact: every human being must periodically become relatively immobile, less responsive, and dependent on conditions arranged in advance. Cooking, bathing, work, worship, sociability, storage, childcare, and sanitation can all be displaced beyond the dwelling. Sleep can be displaced too, but when it moves, the dwelling relation moves with it. Using Norbert Schoenauer's comparative survey of world housing as an empirical spine, while setting aside its developmental ladder from simple to complex societies, this article traces how the work of making sleep possible has been distributed across bodies, bedding, fire, rooms, and cities, from the minimal shelters of foraging societies to the residual bed-alcove of twentieth-century collective housing. A world history of the house, on this reading, is at its deepest level a history of where human beings put themselves when they can no longer keep watch.
I. What Remains When Everything Else Leaves the House?
Begin not with the bedroom but with an ordinary contemporary dwelling, and try a simple experiment in subtraction. Remove the kitchen. The dwelling survives the removal, since people can eat in restaurants, cafeterias, communal kitchens, food stalls, or from deliveries. Remove the bath. The dwelling survives that too, as Roman urbanites, users of the hammam, hostel residents, and countless others have bathed outside their own domestic units. Remove the workplace, and the dwelling obviously survives, since most dwellings in history have never contained one. Remove formal worship, entertainment, storage, laundry, childcare, even the toilet, and historical dwellings have existed without every one of these in turn.
Then ask a different question. What happens if every viable possibility of sleeping there is removed? Something categorically different occurs. An office containing a kitchen, a toilet, and a shower is not thereby a dwelling. A bare one-room shelter in which a family regularly sleeps can be one, whatever else it lacks. This asymmetry suggests a test worth stating plainly: a built space may lose almost every conventionally domestic function and remain a dwelling, but if it ceases to afford recurrent sleep, its status as a dwelling becomes unstable.
This needs one immediate qualification, because sleep plainly is outsourced, constantly. Hotels, barracks, hospitals, dormitories, monasteries, and a friend's spare room all accommodate sleep perfectly well. But these cases prove the point rather than defeating it. When sleep is outsourced, dwelling is temporarily relocated along with it. Eating in a restaurant does not turn the restaurant into anyone's temporary home. Sleeping at an inn does, for as long as the sleeping continues, which is why the inn's most consequential room has always been organized around beds rather than tables.
Architectural history already possesses histories of houses, of vernacular dwelling, of beds, of bedrooms, of privacy, of domesticity, and of the night itself. What it has not generally asked is how the history of domestic architecture would look if reconstructed from the sleeping body outward, rather than assembled from styles, materials, climates, and social structures with sleep folded in as one topic among others. This gap is not a matter of oversight so much as a matter of default perspective. Architectural historians, like most scholars, write from within their own waking lives, interviewing waking informants, reading waking-produced documents, and photographing buildings in daylight. Sleep leaves comparatively little behind for this kind of method to find, which is one reason a subtraction experiment of the kind proposed here, rather than a search for more archival evidence, is the more promising way into the problem.
Norbert Schoenauer's own survey, 6,000 Years of Housing, calls itself a history of dwellings from early settlements to twentieth-century housing, and insists that humble domestic buildings belong within architectural history alongside monuments and palaces. Sleep, however, is not his organizing principle. His book moves from ephemeral shelters through temporary and permanent dwellings, the ancient urban house, the Islamic and Asian house, medieval Europe, and on to the modern apartment and collective housing, along a developmental sequence he explicitly frames through stages of socioeconomic development and what he calls environmental determinism, language that has not aged well and that this article does not adopt. What follows uses his plans, his descriptions, and his extraordinary comparative range while reversing his explanatory logic. The question is not how housing climbed from simple to complex. It is how the work of making sleep possible has been deposited, redistributed, and occasionally concentrated again, across bodies, bedding, rooms, and cities.
This requires one theoretical distinction stated clearly at the outset, since it will do a great deal of work throughout. Embodiment names the sleeping organism itself, thermoregulating, adjusting posture, and periodically half-surfacing into partial wakefulness. Being-with names who sleeps beside, guards, warms, disturbs, or listens for whom. Multiversal dwelling names found, nonrecursive ground, night and daylight, temperature, seasonality, rain, and terrain, none of it built by any household and none of it alterable by one. Multimateriality names made ground, the hut, the tent, the fire, the mat, the bed, the roof, the curtain, the door, the heating system, and the lock. Multisymbolization names the rules attached to all of this: who may share a bedtime, what a room is called, who owns it, and which sleeping arrangements are permitted or expected.
The crucial sentence for everything that follows is this: architecture does not produce dwelling in the strict sense. Architecture is multimaterial forming through which living beings alter their own exposure to conditions of dwelling they did not make and cannot alter. The house at night is made ground, mediating between a temporarily less available body and a found world that owes that body nothing. What remains when everything else leaves the house, in short, is exactly what this article sets out to follow.
II. Before the Bedroom, Before the House: The Architecture of Sleeping Vulnerability
The temptation, in a history like this one, is to start with permanent houses, since they are the best documented and the easiest to draw. The better starting point is the minimum architectural problem sleep poses, stripped of every convenience later ages would add.
A sleeping body stops locomoting. It greatly reduces deliberate environmental surveillance. It cannot continuously adjust its own architecture in response to changing conditions. It becomes thermally dependent in a way a moving, working body is not. It becomes vulnerable to animals, insects, weather, and other humans in proportion to how far its surroundings fail to compensate for its own reduced vigilance. None of this means the sleeper becomes nonrecursive. Selfrecursivity persists, in thermoregulation, in postural adjustment, in the small movements that keep circulation going and pressure off the skin. Interrecursivity persists too, in the responsiveness of one sleeping body to another nearby. What changes is bandwidth and availability, not the presence of recursive coordination as such.
This gives domestic architecture its foundational temporality. Whatever arrangements exist around a sleeping body were made, in every case, by a waking person acting on behalf of a self who was about to become considerably less able to revise those arrangements. Locking a door, banking a fire, laying down bedding, and choosing who sleeps closest to the entrance are all decisions taken by a capacity that is about to be suspended for the benefit of the body that capacity temporarily inhabits.
Schoenauer's own material demonstrates this more clearly than his own framing acknowledges. His earliest, ephemeral shelters are tiny, rapidly built, barely subdivided, and made from whatever lies immediately to hand. The point worth extracting from them is not that they belong to some allegedly primitive stage of social development, a framing this article sets aside entirely, but that even when architecture is reduced to almost nothing, the nocturnal problem does not disappear. Among the BaMbuti, people may sleep on dry leaves or sticks laid directly on the ground, or on raised stick beds built for the purpose, with a fire normally kept burning immediately outside the hut through the night. This is nearly the whole article in miniature: the forest and the night's cold are dwelling, found and unmade; the sleeper is embodiment; co-sleepers are being-with; the leaves, the sticks, the hut, and the fire are multimateriality, made and specific. No bedroom is required. No permanent house is required. Not even a conventional bed is required. Sleep still has to be materially coordinated, every night, by some combination of these elements.
Schoenauer's broader catalogue of what he terms episodic, periodic, and seasonal dwellings extends the same lesson across an enormous range of climates and subsistence strategies, from windbreaks and snow shelters to the seasonal camps of pastoralist and hunting societies that dismantle and rebuild their sleeping arrangements as they move. In each case the architecture on offer is minimal by any modern standard, yet in each case it is precisely calibrated to the specific nocturnal vulnerability of the bodies it shelters, colder climates demanding more insulation and less ventilation, warmer and wetter ones demanding the reverse, every variant solving, in its own idiom, the same underlying problem of a body that must periodically stop watching its own surroundings.
The Arunta case adds a further, easily missed point. Cold nights there are managed partly through bodily proximity, family members and dogs huddling together for warmth, to the extent that Schoenauer reports temperature itself being described locally through the number of dogs needed for comfort on a given night. This licenses an argument worth stating explicitly: housing and bedding cannot be cleanly separated from being-with, because another warm, living body can do work that later architectures delegate to blankets, walls, or heating systems. A dog curled against a sleeper's back is, functionally, part of that night's multimaterial solution, even though no one built it and it will get up and leave in the morning.
The main conceptual result of looking at architecture's minimum case is that the history of sleep's coordination should not begin with an object called the house. It begins with a distributed nocturnal assemblage: a body, others nearby, the ground beneath it, whatever bedding covers it, whatever shelter surrounds it, and whatever fire keeps it warm. The relative burden carried by each of these elements can shift enormously from one arrangement to the next, and much of the history that follows is a history of exactly that shifting burden. The great history to be told, then, is not how humans eventually invented a bedroom. It is how the work of coordinating sleep migrated, unevenly and reversibly, among bodies, other bodies, furniture, walls, rooms, and infrastructures, none of these solutions any more fundamental than the others, and none of them a necessary stage on the way to any of the rest.
III. One Space, Many Times: The Multipurpose Dwelling
This section breaks one of the strongest assumptions built into modern housing, the assumption that functions naturally correspond to spaces. For most of housing history, the same built space changes function during the day, which means the right question to ask of a room is not simply what it is, but what it is doing at this particular hour. Sleeping can occupy, at night, a space used for eating, talking, working, or receiving visitors by day, without any sense of contradiction or compromise.
Schoenauer's African material makes the point immediately. The Luyia dwelling combines a large reception section with smaller living quarters in which occupants sleep, cook, and eat, while the neighbouring Luo arrangement can reserve a central room for both sleeping and social activity at different hours. Neither case maps onto the modern sequence in which sleep belongs to a bedroom, cooking to a kitchen, and sociability to a living room. There need be no such one-to-one relation between function and room at all.
The Mesakin Quisar case pushes the same point further. Young boys sleep on a raised platform. Older girls sleep in the hut where corn is ground. Some huts double as both bedrooms and granaries, while cooking, eating, and most social activity happen mainly in the courtyard rather than inside any of the sleeping structures. The point here is not cultural exoticism but an architectural principle worth stating in its own right: sleep demands provision, not dedicated space. Spaces can be shared across people, across functions, across generations, and across the hours of a single day, without any of that sharing registering as a deficiency.
Traditional Islamic urban houses supply a further variant worth noting briefly. Their courtyards, shaded and cooled by their own orientation and by fountains or plantings, could become the household's preferred sleeping location during the hottest months, with interior rooms reserved for sleeping again once the weather turned. The same dwelling therefore contained, in effect, two distinct nocturnal solutions, indoor and outdoor, activated seasonally rather than simultaneously, which extends nocturnal conversion from a daily rhythm into an annual one without changing the underlying logic at all.
It would be easy to treat all of this as evidence of an earlier, simpler stage that more developed societies eventually outgrow, and this is exactly the inference Schoenauer's own framing invites and that this article rejects. The traditional Japanese townhouse, the Kyoto machiya, is the case that makes the rejection unavoidable. Schoenauer describes rooms with little permanent furniture, in which futons are spread out for sleeping at night and rolled up and stored again in the morning, while sliding partitions allow the boundaries of the rooms themselves to change according to need. This is one of the most sophisticated domestic traditions Schoenauer documents anywhere in his survey, and it solves the problem of sleep through exactly the multifunctional logic the Mesakin and Luo material already displayed.
The contrast this makes available is worth stating precisely, because it supplies one of the article's central distinctions. A Western bedroom stabilizes sleep spatially: the room waits all day, largely unused, for the few hours in which it will be needed. The traditional Japanese room stabilizes sleep temporally instead: sleep arrives, transforms the room by means of the futon, and disappears again by morning, leaving the same square metres free for entirely different uses. A dedicated bedroom, on this reading, is not a necessary condition for stable, well-organized sleeping. It is one historically peculiar way of solving the underlying problem, achieved by permanently dedicating square metres to what remains, for most of the day, an unrealized bodily state.
The machiya extends this logic to the building's own boundary with the outside world. More solid exterior rain doors close the house at night, over and above whatever lighter partitions serve during the day, which means the building's enclosure itself has a day configuration and a night configuration, not a single fixed condition it maintains around the clock. It is worth coining a term for this recurrent transformation, since it will recur throughout what follows under different guises: nocturnal conversion, the regular reconfiguration of spaces, furnishings, and access arrangements to make sleep possible, and their reconfiguration again the following morning to make it stop being necessary. A futon unrolled, a mat laid down, shutters closed, bed curtains drawn, a fire moved, a door barred, a mosquito net lowered, co-sleepers regrouped for the night, all of these are instances of nocturnal conversion, and much of the architectural history that follows can be redescribed as a history of where nocturnal conversion's labour has been concentrated at different times and in different places.
Read together, the Luyia, Luo, Mesakin, and Kyoto cases dissolve a distinction that much Western domestic thinking treats as self-evident, the distinction between a room's true function and its merely temporary uses. None of these traditions treats sleep as an intrusion into space properly reserved for something else, and none of them treats the day's other activities as intrusions into space properly reserved for sleep. The room simply has more than one true function, distributed across the hours rather than across separate square metres, and a household organized this way loses nothing in coherence by comparison with a household that has assigned each activity its own permanent room. It has simply solved the coordination problem along a different axis, time rather than space, and the machiya shows that this solution is fully compatible with a great deal of material and social sophistication rather than being confined to the simplest dwellings Schoenauer catalogues.
IV. The City Outsources Domestic Life, But Not the Sleeping Place
This is the point at which the article's thesis becomes world-historical rather than narrowly anthropological. Urbanization is, among other things, the externalization of household functions that a smaller or more dispersed settlement would have had to perform internally. Cities make it possible to distribute what would otherwise be domestic activities across specialized infrastructures: markets and food sellers, bakeries and taverns, public water supplies, public baths, latrines, laundries, workshops, schools, temples, and places of entertainment. Domestic architecture can consequently become less functionally complete precisely as urban life around it becomes more infrastructurally complex, and the two developments are not in tension. They are the same development, viewed from two different scales.
Rome supplies the case that makes this precise rather than merely suggestive, and it is worth being careful here rather than overstating the evidence. Not every Roman apartment lacked a kitchen or private sanitation. Schoenauer distinguishes the elite domus clearly from the far more numerous tenement flats of the insulae, and the surviving Pompeian house includes its own kitchen alongside segregated family and public areas. The argument does not require universal absence of these functions from Roman housing. What it requires, and what Rome demonstrates unusually well, is that bathing, sanitation, eating, and other bodily functions can be distributed between dwelling and city in radically variable proportions, from the fully equipped domus to the insula resident who ate at a popina, bathed at the public baths, and used a public latrine. Yet whatever proportion a given Roman household settled on, its members still had to return somewhere to sleep. The insula matters conceptually not because it proves that kitchens disappear from housing, but because it exposes the difference between a service a city can plausibly provide elsewhere and a recurrent bodily state that requires the person themselves to become unavailable somewhere in particular.
Public bathing in particular deserves a further word, since the Roman baths were not merely functional substitutes for a private tub but major social institutions in their own right, complete with exercise grounds, libraries, and gardens, further evidence that a function can be almost entirely externalized from the dwelling without any loss of dignity or importance attaching to it. Nothing comparable happened, or could easily have happened, to sleep. No Roman city built a public dormitory in which citizens routinely spent their nights alongside strangers as a normal, high-status alternative to sleeping at home, and the absence of any such institution, in a civilization that externalized bathing, dining, and sanitation so thoroughly, is itself a kind of evidence for the asymmetry this article has been describing.
This difference is worth formulating carefully, because it gives the whole article its sharpest general proposition. Cooking can be outsourced because its product, food, can travel back to the eater from wherever it was prepared. Laundry can be outsourced because clothes can travel to and from a laundry without their owner going anywhere. Bathing can be relocated because the waking person travels to the bath and returns afterward under their own power. Sleep does not work this way. The sleeper cannot outsource the sleeping body while remaining somewhere else, because there is no product of sleep that could travel back to them independently of their own presence. To outsource sleep at all requires moving the person, bodily, to wherever the sleeping will occur. Domestic functions can therefore be divided into those whose product can be transported and those that require the body itself to occupy the site where the function happens, and sleep is the strongest possible case of the latter kind.
This asymmetry between transportable products and site-bound states has a further consequence worth drawing out before moving on. Every one of the outsourced functions listed above changes who performs a piece of household labour without changing anything about the underlying multimaterial requirement it satisfies. Bread still has to be baked somewhere, by someone, whether or not that someone is a household member. Sleep is different in kind, not merely in degree, because there is no equivalent division between the act and its product. The sleeping itself is the product, inseparable from the body performing it, which is precisely why urbanization, for all the domestic labour it manages to externalize, has never externalized sleep in the same sense. It has only ever relocated it, city by city, hotel by hotel, dormitory by dormitory, and each relocation still requires exactly the same nocturnal assemblage of body, ground, bedding, and shelter that the earliest ephemeral huts already required. This is also why hotels, hostels, barracks, and hospitals are not counterexamples to the article's thesis so much as confirmations of it. Each one, for as long as someone sleeps there, temporarily becomes that person's sleeping dwelling, regardless of what the building is called or what else happens inside it during the day.
A twentieth-century experiment makes the underlying logic unmistakable, and it is worth jumping forward in time briefly to see it demonstrated almost as a controlled trial. Sven Markelius's 1935 Stockholm Kollektivhus, as Schoenauer describes it, centralized household labour with striking thoroughness. The apartments themselves were kitchenless, with food supplied from a central kitchen and delivered by dumbwaiter, while childcare and various retail services were organized communally for the whole building rather than household by household. Yet the individual units remained one-bedroom dwellings or, in the smallest cases, dwellings built around a single bed-alcove. Modernist collective housing of this kind asked, quite deliberately, how much of the conventional house could be stripped out of the individual apartment and relocated to shared, centralized infrastructure. Cooking went. A great deal of childcare went. Substantial sociability went. The bed, or the bed-alcove that housed it, stayed. Of everything this article surveys, the Kollektivhus is about as close as built architecture has come to running the sleep test experimentally, under real conditions, with an explicit ideological commitment to removing as much domestic function as possible from the private unit, and the one function its architects did not attempt to remove was the one this article has been arguing all along could not be relocated without relocating the person.
V. The Bed Before the Bedroom: Concentrating Architecture Around the Sleeper
Before architecture dedicates an entire room to sleeping, a single object, the bed itself, can carry an extraordinary proportion of the relevant work. A large early-modern bed could supply elevation above cold floors, insulation, curtains forming a local thermal enclosure, a measure of visual privacy, a considerable display of material status, a sleeping surface proper, sometimes seating for receiving visitors during the day, and, across a household's lifetime, a site for birth, illness, and death as well as for sleep. The bedroom, on this reading, is not history's starting point so much as, in large part, the later architectural externalization of capacities that were once concentrated around the bed alone.
Jan van Eyck's 1434 double portrait of the Bruges merchant Giovanni Arnolfini and his wife supplies an unusually well-documented hinge case. The room the couple stand in has long been called a bedroom on the strength of the enormous curtained bed dominating its right-hand side, but the National Gallery's own technical catalogue identifies it instead as a reception room, the kind of chamber in which a prosperous household would receive visitors and conduct business, furnished, as such rooms commonly were among the fifteenth-century mercantile elite, with the household's single most expensive object. The room is therefore theoretically useful precisely because the bed is present in a space that cannot be reduced to a bedroom in the modern sense at all.
The mediational breakdown is worth stating with some precision, not least because it is easy to get wrong. The Bruges winter, its seasonal daylight, and its ordinary nocturnal cold are found dwelling, conditions the household did not make and could not alter. The bed, its curtains, and the walls around it are made ground, multimateriality proper, an answer erected against exactly that found instability. The body losing heat inside the curtains, and the small adjustments it makes to slow that loss, is embodiment. A spouse, a servant, or a child sharing the same enclosed space is being-with. The conventions governing who may share that particular bed, on what occasion, and under what recognized claim, are multisymbolization. Nothing about this breakdown should be described as the bed constituting a dwelling of its own, in the technical sense of the word used throughout this article. The bed does not generate found conditions but mediates a sleeping body's exposure to found conditions it did not create, using resources it very much did create.
The core proposition worth carrying forward is that a curtained bed of this kind is architecture nested inside architecture, a smaller, more intensively controlled material enclosure built from textile resources rather than architectural ones, occupying the same conceptual position inside the larger room that the larger room itself occupies inside the found world beyond its walls. It is best described as a micro-enclosure rather than, as an older and looser vocabulary might put it, a micro-dwelling, since dwelling in the strict sense belongs only to the found conditions the enclosure answers to, never to the enclosure itself.
It is worth being precise about what kind of necessity is actually at work in a case like this, since the size and expense of the Arnolfini bed can make its role look more fundamental than it really is. Sleep is metabolically indispensable. No household, however poor, can do without it. The bed is not indispensable in the same sense. Vast numbers of households, across history and in the present, meet exactly the same metabolic requirement on raised platforms shared by several sleepers at once, on mats, on bare floors, or within a single undifferentiated room, and none of these arrangements is a deficient version of the bed waiting to be corrected by affluence. They are alternative, equally adequate material solutions to a problem no particular piece of furniture created and none can claim to have solved uniquely. What made a bed of the Arnolfini kind unusual was not that it answered to sleep more successfully than a mat or a platform. It was that the economic and legal apparatus of its own society had made this kind of object into one of the most heavily instrumentalized a household could own, regularly among the first purchases a newly formed household made, among the first items specified in dowry inventories, and among the most valuable possessions named in wills. A bed of this kind was liquid, inheritable, negotiable value, made solid, a historically specific and highly contingent answer to an invariant requirement rather than a necessary one.
The broader historical proposition this licenses is not a simple path from bed, to a better bed, to an eventual bedroom. It is a more general and more useful claim: nocturnal coordination can be concentrated around bedding alone, distributed across a multipurpose room shared with several other functions, or delegated to specialized architecture built for sleep and little else, and a given household's position along this range depends on its wealth, its climate, its social organization, and its available materials rather than on any developmental necessity pulling it towards one endpoint. This prepares the ground for Europe's eventual invention of the dedicated bedroom without treating that invention as a universal culmination every household was always, in some sense, heading towards.
VI. The Expensive Invention of the Bedroom
This should be the largest historical section of the article, because it reverses one of the most familiar stories told about domestic privacy. Start with the oddity at its centre. A dedicated bedroom is materially extravagant. A room occupies land, walls, a roof, heating, windows, furniture, and ongoing maintenance, while standing substantially underused for most of the waking hours in every twenty-four. Put the question provocatively: why would any society become wealthy enough to leave an entire room waiting all day for the arrival of an unconscious person?
Answering this requires attending first to how recent corridor-separated functional zoning actually is. Schoenauer notes that early medieval stone houses commonly had only two or three rooms to a level, opening directly into one another rather than being served by any corridor, since corridors themselves had not yet come into use. The decisive historical transformation, on this evidence, is not simply the accumulation of more rooms over time. It is the growing ability to reach one room without having to pass through another. Doors, passages, corridors, staircases, and eventually locks reorganize who must physically pass through whose sleeping space in order to go about their own business, and as this reorganization proceeds, sleep becomes architecturally easier to isolate from the rest of a household's daily traffic, not because anyone has decided sleep deserves more privacy in the abstract, but because the physical means of routing people around a sleeping body, rather than through it, have only recently become available.
The corridor deserves one further comment before moving on, because its significance is easy to understate. A corridor is, among other things, a piece of architecture whose entire function is to let one person move past a sleeping room without entering it. Before corridors existed in ordinary domestic buildings, reaching the far side of a house after dark generally meant walking through whichever rooms lay in between, including, on many nights, rooms in which someone was already trying to sleep. The corridor's real achievement, considered from the perspective this article has been developing, is not primarily aesthetic or even principally about privacy in the abstract, but a piece of multimaterial infrastructure purpose-built to reduce the recursive burden a household's own waking members would otherwise impose on its sleeping ones, letting movement continue elsewhere in the building without that movement having to cross, quite literally, through anyone's unconscious body.
Heating technology deserves mention alongside corridors and locks, since a dedicated bedroom is not viable in a cold climate until a household can afford to heat a room that will stand empty and unheated-in-use for most of the day, only to be warmed just before it is needed. Centralized heating systems, and later still electric blankets and thermostatically controlled radiators, did for temperature what corridors did for access, letting a household concentrate warmth precisely where and when a sleeping body required it rather than needing an entire building, or an entire bed's worth of curtains and bodies, to achieve the same result. The dedicated bedroom, in other words, was not only an achievement of walls and doors. It was equally an achievement of fuel, and of the infrastructure needed to deliver that fuel exactly where sleep, rather than any other household activity, required it.
Nineteenth-century Paris makes the eventual endpoint of this process brilliantly visible, and Schoenauer's account of its apartment plans is worth following closely. By this period, fashionable apartments had become zoned into distinct social or reception areas, private or personal areas including bedrooms and bathrooms, and service areas including the kitchen and servants' rooms, three functionally distinct zones within a single dwelling, each increasingly reachable without passing through either of the others. The lady's bedchamber, in particular, could become what Schoenauer calls the home of the home, linked to its own dressing room and boudoir, and yet still partly social in character, a space where intimate visitors might be received even as it also served as the most private room in the apartment. This detail matters because it shows that specialization, even at its most advanced, is never total. The bourgeois bedroom of nineteenth-century Paris remained simultaneously a sleeping place, a dressing space, an intimate reception area, a gendered retreat, and a site of status display, all five functions coexisting inside one increasingly specialized room rather than being fully distributed to separate ones.
This specialization was also, from the outset, sharply classed, and Schoenauer's own material makes the point without needing much elaboration. The multi-zoned apartment with its separated social, private, and service areas was available only to households that could afford the additional square metres each zone required. Poorer households living in the same city, often in the same building, continued to sleep, eat, and receive whatever visitors they had within one or two rooms doing all of this work at once, not because they had failed to discover the value of privacy, but because privacy of this particular, architecturally expensive kind was simply not for sale at their income. A household with limited space cannot allocate an entire room to an activity that occupies only part of the day, and its rooms and its sleeping arrangements remain correspondingly multifunctional for exactly the same reason the Mesakin or Luyia arrangements described earlier were multifunctional, not through any developmental lag, but because multifunction is what a limited quantity of enclosed space is used for whenever more of it cannot be justified by the specific problem sleep poses on its own.
This evidence licenses a rewriting of the conventional history of privacy that architectural and social history have generally told. The conventional story runs from modernization, through rising individualism, to privacy as a value, and finally to the private bedroom as that value's material expression. The re-reading proposed here inverts the order of explanation without denying any of the historical facts the conventional story relies on. Increasing material resources allowed progressively more architectural work to be devoted to protecting and isolating the interval of a household's reduced waking availability. Privacy, on this account, is one outcome of that process rather than its primordial cause. The bedroom is better understood as a technical and social solution to several distinct sleep problems at once, solved together because a society had, for the first time, the resources to solve them together: interruption, noise, uncontrolled access, sexuality, status display, thermal control, and household hierarchy all found a single architectural answer in the specialized room, where earlier and poorer households had needed to solve each of these problems separately, if at all, using the bed, the platform, or the shared room as their only available instrument.
This is not, and should not become, a story of progress. The Kyoto machiya has already disproved any claim that a universal developmental sequence runs from multifunctional space towards the specialized bedroom, since it achieves a comparably sophisticated solution to the same underlying problem by an entirely different route, stabilizing sleep in time rather than in space. The anti-teleological point deserves stating without qualification: functional specialization is not an evolutionary endpoint that domestic architecture everywhere converges on given sufficient time and wealth. It is one way, among several genuinely viable ways, of allocating material abundance once a household has some abundance to allocate. Rooms that have become specialized can, and regularly do, become multifunctional again, whether through poverty, through changing household composition, through deliberate choice, or through the same forces of urban intensification that, elsewhere in this article, are shown pushing domestic functions back out of dedicated rooms and into shared, extra-domestic infrastructure.
VII. Sleeping-With: Architecture Does Not Merely House Individuals
Everything so far risks sounding more materialist than the evidence actually supports, and this section exists to correct that impression. Every architecture of sleep simultaneously organizes being-with, whether or not its builders ever thought of it in those terms. Spouses can be housed together or apart. Infants can sleep beside parents or elsewhere. Siblings can share a room or be separated by age and gender. Servants can sleep nearby or at a physical remove that marks their place in the household hierarchy. Guests can be placed close to the family's own sleeping quarters or kept well away from them. Unmarried adults can be housed with parents, with siblings, or independently. Animals can sleep inside the dwelling or be excluded from it. Elderly parents can remain with the household or be housed separately. Every one of these is, at bottom, an architectural question, because walls, beds, platforms, doors, and the allocation of rooms are exactly what materialize the answer a given household has settled on.
Schoenauer's own material already shows sleeping allocation marking age and gender directly. In the Mesakin cluster discussed earlier, young boys use one sleeping platform while older girls sleep in the hut where corn is ground, a spatial separation that tracks age and gender rather than any purely practical constraint. The Arunta case makes the opposite point equally well. Bodies themselves, human and canine, can form part of a night's thermal architecture, huddled together for warmth in a way that dissolves any sharp line between a household's living members and its material infrastructure.
A further body of comparative evidence, drawn from fieldwork in Bengal rather than from Schoenauer's own survey, sharpens the same argument considerably. In many Bengali households, the family bed is not reserved for a conjugal couple's exclusive and private use. It can remain reception furniture throughout the waking day, hosting visitors, conversation, and children's play, and it can just as easily host two adult male friends who simply lie down and sleep together after an afternoon of talking and eating, an arrangement that draws no comment and requires no explanation within the household in which it occurs, however unusual it might look from within a very different set of expectations about male friendship and physical proximity. The same comparative material shows families in which parents and an adult daughter, well into her twenties, continue to share a sleeping room as a matter of course, including when the family travels together, without any shortage of money or space to account for the arrangement. It is simply how that household prefers to sleep.
The important conclusion to draw from material of this kind is not that any particular society lacks privacy, a framing that quietly imports the very developmental assumption this article has been trying to dismantle. The conclusion is that different mesocosms make different relations between intimacy, privacy, hierarchy, and sleeping-with architecturally viable, none of them more advanced than any other. The dedicated solitary bedroom, so central to the nineteenth-century Parisian sequence traced above, is not simply a more developed solution to the problem sleep poses. It is one particular distribution of interrecursivity among the many distributions available, selected and sustained by a specific society's own priorities rather than dictated by sleep's own metabolic requirements, which remain fully satisfied by every arrangement considered in this section.
None of this material should be read as claiming that hierarchy disappears wherever sleeping-with is more widely practised. A household that sleeps five to a room can distribute comfort, warmth, and proximity to the safest or warmest spot exactly as unevenly as a household with five separate bedrooms distributes square footage and privacy. What changes across these cases is not whether hierarchy exists, but the material register in which it gets expressed, position on a shared platform in one case, allocation of an entire room in another.
The strong formulation worth carrying forward from this material is accordingly this: the history of domestic architecture is partly a history of how much being-with is permitted to survive the transition into sleep, and how much of it a given household, or a given society, has chosen to design away.
VIII. The House After the House: What Modern Domestic Architecture Keeps
Return now to the opening subtraction experiment, which the intervening history has demonstrated rather than merely asserted. Modern urban life has externalized or digitized an extraordinary amount of what used to be domestic activity. Food delivery and restaurants have absorbed much of what a kitchen once had to do alone. Laundries, gyms, offices, coworking spaces, and storage units have absorbed activities once conducted entirely inside the home. Childcare has moved substantially outside it. Streaming services have replaced a good deal of domestic entertainment hardware. Communal gardens and cafés increasingly function as extensions of, or substitutes for, a living room. At the same time, working from home and digital technology can flow back into the house in the opposite direction, so the movement of functions across the domestic threshold runs both ways rather than in a single, historically irreversible direction.
But the sleep problem does not move with any of this traffic. The modernist collective-housing experiment considered earlier already showed why. The Kollektivhus could centralize its kitchens and its childcare while retaining, in every unit, a bedroom or a bed-alcove that no comparable centralization ever touched, and nothing in the subsequent history of housing has reversed that particular asymmetry.
This licenses what might be called the residual dwelling hypothesis: as domestic functions become increasingly detachable from any particular building, sleep becomes more visible, rather than less, as the residual function through which a residential unit remains recognizably residential at all. A micro-apartment containing almost nothing beyond a sleeping platform, a shower and toilet, and somewhere to charge a phone can still count, without much argument, as housing. A large and well-appointed corporate office, with better kitchens, better lounges, and better bathrooms than most homes can afford, does not count as housing, however comfortable it is to spend a working day inside. The difference between the two is not a difference in comfort, or in the quality of their fittings, or even in how much time is spent in each. It is a difference in what is expected to happen there at night.
This asymmetry is worth dwelling on because it suggests something about where domestic architecture's future is actually heading, as against where contemporary design discourse often assumes it is heading. Much recent thinking about micro-living and shared amenity buildings treats the shrinking private unit as evidence that the home itself is dissolving into a wider urban infrastructure, following food, laundry, and entertainment out into the city. The evidence assembled in this article suggests a more precise claim. The home is not dissolving so much as being reduced, again and again, to the one function that has never in six thousand years of housing proved detachable from the individual body occupying it, and every wave of externalization, from the Roman insula to the Stockholm Kollektivhus to the contemporary micro-apartment, has stopped at exactly the same wall.
IX. Conclusion: Where Humans Put Themselves When They Cannot Keep Watch
Return, to close, to the earliest shelter and the contemporary apartment together. It would be too strong, and unnecessary, to conclude that sleep caused architecture. The claim this article has actually built is more modest and, for that reason, more defensible.
Sleep is among the most persistent architectural problems any human society has had to solve. It does not require a bed or a bedroom, as the BaMbuti hut and the Kyoto machiya both demonstrate in their very different ways. The problem can be solved through bodies, co-sleepers, bedding, fire, shelter, furniture, rooms, houses, and infrastructure, in combinations that shift constantly across history without any of those combinations representing a more advanced solution than any other. Architectural history has generally classified these arrangements by building type, style, climate, technology, culture, or the history of privacy. Reading them instead sub specie somni, under the aspect of sleep, reveals another continuity running beneath all of those classifications, one that connects a BaMbuti hut, a Bruges merchant's curtained bed, a Kyoto machiya, a Parisian apartment, and a Stockholm bed-alcove more directly than any shared style, material, or period ever could.
A house, on this reading, does not merely shelter waking activity. It makes possible the recurrent suspension of waking activity itself, and arranges, in advance, for what a body will need once that suspension has begun. A world history of the house may therefore be, at its deepest level, a history of where human beings put themselves when they can no longer keep watch.