Skip to content

Architecture Myths #20: The Villa Savoye

Post date:

This is Sneferu Shining in the South Pyramid also known as The Bent Pyramid built circa 2600 BC for Pharoah Snefuru, Priest of Bastet, Guardian of Nekhen, eternal dude.


2,600 BC is a while back. Frankly, no-one has any idea why this pyramid was built the way it was but, people being people, they speculate.

  • Some archaeologists believe the Bent Pyramid is a transitional form between step-sided and more “perfectly shaped” pyramids. 
  • It has been suggested that the steepness of the original angle of inclination the structure caused the structure to become unstable during construction, forcing the builders to adopt a shallower angle to avert the structure’s collapse, such as had happened during the construction of the Meidum Pyramid.
  • For a while it was believed the shallower angle meant the construction could be completed in time for the Pharoah’s approaching death.

Nobody has ever suggested Snef.P_V1 was pushing the boundaries of pyramid aesthetics. Rather, these speculations all assume the intention was to aim for some sort of geometric perfection and that the as-built edifice is some sort of compromise. It’s what we want to believe. It’s our nature. Here’s another building for which we have incomplete information. Oddly, the opposite occurs.

This is an image that's as close to completion as I've ever seen.

We like to think this building was always meant to be what we see. Much of what’s been written about it assumes it was exquisitely inspired and designed to be precisely the way it is, and that nothing was left to chance or compromise. This is wrong.

In Modern Architecture Since 1900, William J.R. Curtis devotes Chapter 16 (pages 275–285) to the image and idea of le corbusier’s villa savoye at Poissy. No less than seven pages in, on p.282, he lets us know the design process was not straightforward.

apology copy

To him, this is evidence enough that LC knew what he was doing.


I’m not so sure. It’s true the history of architecture is, mainly, a record of things that got built but it’s also true we tend to ignore how susceptible to chance that record is. Not unlike Snefuru’s pyramid builders, huge edifices of words and analysis get built upon the most insubstantial of foundations.


  • 1928, September. A few sketches. The one below at top right is not unlike the as-built – from that angle. But look immediately below and see how what we today know as the rooftop was originally only what could be seen of a second floor. This seemed important to LC, perhaps because he’d said Architecture is the masterly, correct and magnificent play of masses brought together in light back in 1923 and the only evidence his buildings had offered were:
    • an external stair of the Ozenfant Atelier/house (1922),
    • a fireplace alcove in La Petit Maison (1923),
    • an enclosed spiral staircase and a curvy corner in Villas Lipchitz-Miestchaninoff (1923),
    • the gentle curved wall of Villas La Roche-Jeanneret (1923),
    • a fully curved wall in Maison Ternisien (1923),
    • the curved ends of the annex to Pavillon de l’Esprit Nouveau (1924),
    • many things at Villa Cook and Villa Stein–de Monzie (both 1926) but all in the shade,
    • a curved landing wall on an external stair at Villa Church No. 1 (1927),
    • grand external staircase at Villa Church No. 2 (1927),
    • another curved landing wall at a Weissenhofsiedlung Villa No 1 (1927) and
    • a master bedroom with a full-height semicircular wall on the uppermost floor at the Villa Baizeau (Tunisia, 1927). Significantly, this semicircular wall has no window openings.  It’s an obvious precursor but is unfortunately shaded by a roof slab.

So yes, getting that big curve out and under the sun seemed to be a driver.

  • 1928, November. Two months on, the whole thing is looking decidedly iffy. LC is not in the office much.
classic sketch
  • 1928, CIAM I, La Sarraz, Switzerland, Foundation of CIAM
  • 1929, CIAM II, Frankfurt am Main, Germany, on The Minimum Dwelling
  • 1929. Overseeing the production of Vol.I of his Oeuvre Complete 1910-1929.
  • 1929, April. Construction of the Villa Savoye begins.
  • 1929, September. Writing the introduction to Vol.I of his Oeuvre Complete 1910-1929.
Le_Corbusier_Vol_1_1910_1929 183
Le_Corbusier_Vol_1_1910_1929 184

Let’s pause it here. Construction of VS began in April 1929 and Vol.I of LC’s complete works was published in 1929. The next image shows Villa Savoye as it appears in Vol.I. If construction commenced in April 1929 and LC was still writing the introduction to Vol.I in September, then it’s safe to say this is what the builders were digging the basement and laying the drainage for. Note: The ground floor slab might not yet have yet been poured in September because the revised design has a couple of columns we’ve not seen before.

  • The main difference is that the master bedroom and bathroom are on what we now know as the roof.
  • The internal staircase is straight and, though it links all floors, is service stairs on the basement, ground and first floor levels, but bedroom stairs between first and second.
  • There are many curious storage spaces lining the ground floor service corridor.
  • The position of the chimney suggests the basement was much larger.
  • There’s a totally different feature bathtub above a feature w.c. below.
  • There’s not that column in the garage, or the one at the end of the maid’s bed.
  • The master bedroom has curved walls but no the bed has no direct view out.
  • Another curve contains a spiral stair that continues roofwards.
  • A third curve is presumably a wind-shield for a quasi-secluded sunbathing area.
  • There’s an external stair linking the terrace with the garden on the garage (east) side.
Le_Corbusier_Vol_1_1910_1929 185

LC was on a roll in 1929. Part of it was spent in South America, not to mention getting there and back. Some more of 1929 was spent sketching a naked Josephine Baker salacious reports salaciously report, but there are a few things like this floating around the internet that aren’t included in LC’s usual bios.

Sometime during 1929, LC also found the time to find someone to marry him. As if that’s not enough for any media star, there was still work to be done. [Thanks ncmodernist!]

For someone already publishing volume un of their oeuvre complète, 1929 was the year LC’s career took off. His thoughts however, and much of the time the man himself, were in Moscow. Since his first visit in 1928, LC saw himself as Moscow’s urban saviour and allied himself with the proponents of the Green City movement. However, by May 1930, he’d produced his own 60-page report and 22 drawings for the reconstruction of Moscow. I only mention this in this post because LC’s 1928–1932 infatuation with the Soviet Union perfectly overlaps the design and build timeline of Villa Savoye. It’s easy to imagine VS and its troublesome clients were not high up on his to-do list. I’ve no doubt the Savoyes sensed this, for the design of VS was changed during construction.

To change the design of a building once construction has begun is A BIG THING and only happens when clients are desperate to get an architect’s attention by threatening to turn off the cash. Occam’s Razor suggests the Savoyes were annoyed with LC being uncontactable and preoccupied. But get LC’s attention they did for, better or worse, VS was promptly redesigned and construction continued according to the VS–LITE design. The VS we know today is the consequence of clients wanting their project finished on time and on budget. Here’s how the plan appeared in Vol.II of LC’s complete works 1929–1934

Pages from Le_Corbusier_Vol_2_1929_1934

The sectional view hasn’t been updated – you just can’t get the staff! These days architects pay people to incompetently manage their social media pages. When I last had a facebook site, Zaha Hadid’s people once friended me. More recently, Patrik Schumacher’s people have reposted images from misfits [20092019: when it was on Instagram].


[Btw, misfits is now on Pinterest and Instagram and there’s also a Facebook page.] Anyway, let’s see how far construction progressed before the Savoyes sent LC their wake-up call. This next photo claims to be from the summer of 1929 and it may well be.


The only two other construction photos I can find show construction progressing according to the post-1929 design.


We need to dig deeper, and enter the realm of architectural forensics. If the design changed between five and eight months after construction began then it’s unlikely to have progressed farther than preliminary site works and perhaps the ground floor slab but, even so, that’s still major pain. Here’s the only drawing I’ve ever seen of the basement as-built.

villa savoye basement

Judging by the position of the furnace chimney and where the basement stairs were to have ended, the basement was shrunk from two structural bays to one.


Filling in an already-excavated basement is wasteful but is still preferable to having the position of the stairs multiply that waste over the levels above. Those straight stairs had to go! Creating some sort of lobby sculptural element à la Villa Stein was never the intention, but more interesting is what happened to the drainage. In the early 1929 plans above, there’s a curved wall concealing the washbasin for the “front-of-house” domestiques to wash their hands before touching anything belonging to the guests. 1920s Parisian outer suburbs being 1920s Parisian outer suburbs, that washbasin is on the main line to a septic tank that’s already been dug.

It would have been too expensive and time consuming to shift the drainage pipe. The redesign has two toilets placed immediately above where that washbasin was to have been. One constant in architecture is that the shit has to go somewhere. You can learn a lot about the art of architecture by studying drainage design.

Here’s some views of that waste pipe.

Proto High-Tech? I think not. I’m surprised no-one’s written a PhD about it. Perhaps, deep down, people know it’s crap. This hurried and careless redesign seems more and more like a botch job. That exposed furnace flue now seems more happenstance than artful contrivance. Let’s have a look at what happened to the master bedroom bathroom now it’s shifted down a level. The intended plan had two bathrooms on the outer wall but the quick fix plan now has bathroom in the middle. You know the one.

If all these people would get out the way, we’d see a black door for the wc that contributes to the exposed soil pipe we’re already familiar with.

The adjacent wc does as well. It’s the main wc for the salon level and thus all visitors. (Overnight guests in the guest room across the corridor have an en-suite bidet and washbasin but no wc.)  It’s that fancy relocated bath that’s the problem. It drains from the bed end.

villa savoye bathroom drain

In the next image, this column in the ground level has always been drawn egg-shaped. It’s not in the greatest of positions if you’re living in that room but, let’s not forget, you’re a laundry-maid and you should think yourself lucky to have your employers’ bathwater draining down a rendered attachment to the column at the end of your bed.

Remember how in the originally intended design, some serious bathroom drainage had been anticipated in that part of the house? Its groundwork wasn’t going to change. It’s responsible for the drain being in the domestique’s rejigged room and its off-grid column supporting the column artfully framing the relocated master bed above.


FWIW, the guest bedroom’s bidet and washbasin drain through the wc provided off the lobby for guests caught short.


It’s all a bit messy. It stinks of compromise and of decisions made hastily and the simplest explanation I can think of is that Les Savoyes had threatened to turn off the money. What this all means for us is that the huge architectural cultural construct that is the Villa Savoye, rests on a building that was never intended to happen. We’re led to believe LC cared about VS when, given what else was on the cards for him careerwise, it’s more likely he wanted VS done and forgotten. Au contraire! you may say but, as a conjectural history of VS’s design, these conjectures are at least based on physical evidence.

The VS we know today and endlessly analyze and ponder would not exist if the Savoyes hadn’t been so short on patience and money. True, given LC’s formidable media footprint at the time, the Villa Savoye would probably still have become an architectural cultural phenomena of similar magnitude, but the same things would have been written about a totally different building – reminding us once again of how the history of architecture is built upon foundations that aren’t as solid as we think.






  • Thanks Graham, for expanding on the VS design process , and for the Guggenheim model pictures. That FLLW project was a bit off your original topic, but fascinating (to me) because the design process involved such a huge variety of plan-shapes, elevation-forms, colors (e.g. circular or polygonal in plan, sloping up or down or straight sided, red or white or blue, etc.);
    all those changes seem to have been in response to “outside forces” (death of the client, negative input from the museum director, the NY building department, artists, etc.).
    But in most of those revisions, including the as-built one, the architect was apparently able to retain the “seed idea” of a single, continuous path from entry to exit (like the in-out path of a labyrinth as opposed to a meandering, room-to-room path like a maze). BTW, those designs included at least two versions in red (one inverted and the other not), and evolved from FLLW’s Gordon Strong Automobile Objective project (he came up with some catchy project names, no?), which in turn (in my opinion) was “inspired” by the Spiral Minaret in Samarra. I haven’t seen all of those unbuilt Guggenheim projects presented in one place, but if you’re interested you can see some of them by Googling:
    “frank lloyd wright guggenheim drawings” , and then go to the IMAGES page.

    • David, here’s another one for you. FLW is on record somewhere as saying Fallingwater is really his 1940 Pew House pimped. It does seem to have been on his mind. The corner windows are there, and there’s an idea for a fireplace. Importantly, there’s a general massing idea that could be adapted to work on any site – but particularly one with a directional view as he was soon to have. I’m sure the design of Fallingwater wasn’t “inspired” as we like to believe, even though the drawings may have been a bit of a rush. The real genius was knowing when to reapply something one had already thought of and done before. Unlike LC and VS, FLW really needed this job to re-start his career.

      You can see more photos of the Pew House here, from when it was recently for sale.

  • Isn’t it common practice for designs to change (trivially or radically) — before, during and even after construction — in response to all kinds of issues, not just those involving the patience and money of clients? What might be written about a building after it has been built is a different issue, but I wonder if the same things would have been written — for example — about any of the other designs (had they been built) that Wright made for his Guggenheim: that design started out upsidedown (and bright red) — see

    • Thanks David, That’s true, it’s not unusual for a design to change during construction but, at least these days, it involves a whole world of pain with change orders having to be approved by client/project manger, contractor, architect and so on. Contracts were probably simpler then but Pierre Savoye was, after all, a director of a successful insurance company so he could probably spot a financial risk and wasn’t in the habit of writing blank cheques. The things in the Villa Savoye that didn’t change are the idea of the glassy lobby, the internal ramps and the whole processional thing to the salon. The position of every internal wall did change so that says a lot for the flexibility of the free plan. Basically, the master bedroom pushed the kitchen into the living room, shifting everything else inbetween and on two levels because stairs were involved. It’s amazing the house appears as coherent as it does. The only giveaways are the drainage compromises, that internal bathroom that now had to be skylit, and that lonesome column downstairs.

      But thanks also for pointing me towards Guggenheim images. I’d seen a pink one and a reversed one but never a red upside-down one before. I used to think that this next image was reversed as there was no lettering and I can’t identify the painting.

      However, in this next image, the breast pocket of FLW’s suit is on the side it should be.

      There was clearly a bit of work still to do.