The Unconditioned House: What If You Only Conditioned Half a House?

A speculative retrofit of a 1960s brick veneer. Half the house is sealed and insulated properly for the first time. The other half is deliberately left unconditioned — and that half is where most of the year happens.

This one is unbuilt. It is a finalist in the ArchiTeam Awards 2026, Unbuilt category, and it exists to argue with something: the assumption, built into the way we energy rate houses, that a good house is one held at the same temperature all year.

The question behind it

Ask an energy rating what a good house is and it will tell you: the one that holds every room at one temperature for the least energy. Conditioning isn't a question in that assessment — it's the premise. A good rating makes conditioning cheaper. It never asks whether the room needed conditioning at all.

That assumption is expensive to satisfy in a house like this one. Melbourne's middle suburbs are full of 1960s brick veneers: too numerous to replace, too expensive to seal properly. The two conventional answers are to insulate throughout, seal it, bolt a high-performance box onto the back and pay a great deal to reach current standards — or to demolish and waste a whole house of embodied carbon. Both end in the same place: a sealed envelope, conditioned all year.

The Unconditioned House asks a different question. Which parts of this house could go unconditioned, and what do you get back?

Half a house

Seventy-two square metres are sealed and insulated properly for the first time. Seventy-two square metres are purposefully unconditioned — not exposed, but moderated against the extremes. Shaded in February. Glazed and closable in July. Never heated, never cooled.

The unconditioned half is not leftover space. It is the loggia, the sleepout, the fire court, the breathing room and the front patio, and between them they hold most of the year.

The loggia in Buath Gurru (typically November). Doors folded away, the vine at full leaf, the core and the loggia reading as one room.

Most of the work is rearrangement

The existing plan ignores the sun. The bedrooms hold the north edge, sleeping through the best light of the day, while the lounge sits on the cold south side. So we swap them.

Kitchen, meals, lounge, the main bedroom and a bunk room move north and become the conditioned core. The south side becomes a sleepout behind a batten screen and a flexible breathing room with a window seat. The loggia is glazed north and west at the western end. Around the core sit a fire court, a front patio, a bathing court and a seasonal garden.

Drag to compare. The bedrooms come off the north edge; living moves onto it. The shaded zone on the proposed plan is the 72m² that gets conditioned.

The roof and the structure stay. Demolition is limited to openings and linings. At the unconditioned edge the linings come off entirely, back to raw brick and open frame — so you can tell the two halves of the house apart by touch, in the dark, without being told and returning the bricks' ability to moderate the temperature.

Warmth goes to people, not air

Heating a volume of air is the expensive habit. So in the parts of the house that get warmth, the warmth is aimed at bodies and stored in mass.

There is a reclaimed-brick seat under the north window of the lounge, a heated plinth in the fire court, and a heated fin that doubles as the bedhead in the main bedroom. All three are charged from the roof by day, held in masonry, and discharged slowly overnight. You sit on the heat rather than standing in it.

The fire court in Waring (typically April-July). Without a reason to go outside in winter, the whole argument collapses into "outdoor living is nice in summer" — which is not an argument.

A year in seven seasons

The hero drawing for this project isn't a plan of the house. It's a plan of the year.

A year in The Unconditioned House. The outer ring is the Kulin Nation season; each inner track is one space, shading from pale to dark as it comes into use. Nothing switches on a date. Open it full size to read the indicator species.

Melbourne has seven seasons, and always did. The Eastern Kulin calendar names them — Biderap, Luk, Waring, Guling, Poorneet, Buath Gurru, Gunyang — and defines them by events rather than dates. Luk is when the eels are fat, not a fortnight in March. The four-season calendar we run on is imported, and so is this house: a northern-hemisphere house form, on a northern-hemisphere calendar, in a place that has neither. Mechanical conditioning is the technology that made that mismatch survivable.

So each room in the plan is given three or four settings rather than one function. Closed. Solar. Open. Shaded. The season decides which one is running.

The core in Gunyang (typically December) — cooking early and late, closed through the heat of the afternoon.

The core inverts

Here is the neatest proof that this is a project about the year rather than about insulation.

The conditioned core is shut in July to keep the warmth in, and shut in February to keep the heat out. The single conditioned space in the house is conditioned in two opposite directions across the year. In Waring (typically April-July) it is the whole house. In Biderap (typically January-February) it is a cool store you retreat into at three in the afternoon and leave again by five.

Waring, in the core. Four months of the year the sealed half is the whole house, and it should be good enough to earn that.

Sleeping migrates

The household does not sleep in the same room in January that it sleeps in in July. In Waring, the cool months from around April-July, the main bedroom is the primary bedroom, its masonry heat bank charged by low winter sun and radiating overnight. By Buath Gurru, usually November, the household has moved out to the sleepout, and it stays there through Biderap in February.

Which gives the year two dated events: the first night out, and the last night out. They are marked on the drawing.

Guling (August). The heat bank beside the bed, wattle out the window, and a sunrise window doing the waking.

The breathing room in Poorneet (September-October). Bunks, a window seat, and somewhere to pass a wet spring day.

The sleepout in Biderap (January-February). Unconditioned, screened, and the coolest place in the house.

The hard part is Waring

Waring runs April to July. It is a third of the year, it is cold and wet and dark, and it is where a project like this normally falls over — six vivid seasons and one dead one, with everybody shut in the box.

So the winter got the same design attention as the summer, and it has four moves of its own: the fire court on dry cold nights, the loggia glazed and closed as a midday sun trap, the masonry heat bank charged by low sun, and the window seat as the light-seeking spot of a dark season. The seasonal garden does its most important work in Waring by not being there — bare and structural, letting the low sun reach the loggia and the heat bank it will shade out completely in February.

Why here

The inner west is the hardest place in Melbourne to make this argument, which is exactly why it is the right place to test it.

Melbourne's western suburbs average under 6% tree canopy against around 26% in the east, and a treeless street can run up to 10°C hotter than a shaded one. Unconditioned outdoor living needs shade, and the suburb doesn't have any. Which means this house cannot be fixed at the envelope. It has to be fixed at the block.

One house's canopy is private amenity. A thousand houses' canopy is a cooler suburb.

That is the part that scales. The house is one of tens of thousands of near-identical brick veneers across the middle suburbs, and the moves are not exotic: read the local climate, then adapt the rooms and the functions to its rhythms instead of insulating against them.

What it costs to be honest about

It asks something of you. This house works if you move with the day and the year — if you shift where you eat, where you sit, where you sleep. That is a real ask, and it is the reason the project is speculative rather than a product.

It also runs against the grain of how houses are assessed. An adaptive comfort model — the research says occupants of naturally ventilated buildings accept a much wider comfort band than sealed-envelope modelling assumes — supports the case. A standard NatHERS assessment, which models a static occupant in a sealed box, does not know how to score any of it.

And none of it is new. Lisa Heschong made the sensory half of this argument in Thermal Delight in Architecture in 1979: uniform conditioning as a kind of sensory deprivation, thermal variety as pleasure. What this project adds is the temporal half — variety isn't only delightful, it is how a house keeps time. A year held at 21°C has no events in it, which is a large part of why the days run together.

The migrant households of Melbourne's west were already doing the practical version of this in exactly this housing stock: summer kitchens in the garage, dinner under the grapevine, sleeping wherever it was coolest. Comfort for the body, not for the house.

Less an invention, then, than a recovery.

The Unconditioned House is a finalist in the ArchiTeam Awards 2026, Unbuilt category. If you have a 1960s or 70s brick veneer and this argument sounds like your house, we'd like to hear about it.