Skip to content

The Placement of Windows

Post date:

As a follow-on from past posts on the subject of architectural reductions and the gradual lessening of the number of avenues open to architects to create architecture, the placement of windows may well be one of the last to go. I say this because walls and openings in them are a natural and essential pairing – they go together. Sure it’s possible to have an external wall without a window, but the presence of windows indicate the presence of people on the other side and this a fundamental reason why buildings exist in the first place. Today however, I don’t want to contemplate an architecture without windows.

In his wordy 1896 essay The Tall Office Building, Artistically Considered, Louis Sullivan distances himself from the idea of tall building as classical columns with base, shaft and capital, and then proceeds to arrive at the same divisions from a different direction. Below, from left, are his 1891 Wainwright Building, its section, the 1891 Monadnock South Building by John Wellborn Root, along with section justification for its divisions. Its 1893 North building extension by Shepherd Books reverted to an uppermost floor that so wasn’t a capital. [Thanks DG!]

What Sullivan was redefine those accepted divisions in terms of functions while forgetting to remember that classical columns had bases, shafts and capitals because of different functional requirements in the first place. (Did nobody point this out to him, I wonder?) Whether Sullivan was being obtuse or simply reframing the problem for newly modern times, his functional justification for a three-part tall building is less dependent upon time and fashion. It’s true that mechanical and services floors have no need for conventional windows but then why do the external walls of the uppermost floors have fluted cornices looking suspiciously like capitals? If we’re going to talk about function suddenly, then it’s not as if these tall buildings are actually supporting anything. Maybe Sullivan saw the absurdity of this in 1896. In 1922, Adolf Loos definitely did.

Anyway, the 20th century wore on, the division stuck and the windows of the floors above the bottom and below the top remained resolutely stacked despite superficial stylistic differences. Architects didn’t give much thought to what else windows could do until Josef Polasek’s 1930 Municipal Housing in Brno. [c.f. The Persistence of Beauty] The material of the basement base is a darker color but its openings are the same width as those of the typical floors. The uppermost floor has laundry rooms that are service rooms in the Sullivan sense, but have windows to make doing the laundry more cheery. So far so different-function-different-windows but, if you squint, you can see these laundry room windows as capitals above five levels of paired windows forming a shaft. These aren’t heavy-handed post-modern columns but nor are they applied decoration. There’s nothing about this facade that can’t be rationalized but, once you see those abstracted columns you can’t unsee them. It’s a shame all Modernist buildings weren’t as well thought through as this, with a gratifying meaning staring us in the face if that’s what we are looking for, there to be appreciated.

The placement of windows can do other things. The image on the left below is of Ignzaio Gardella’s first proposal for his 1956 Casa del Zattere, the photo on the right is it as built, and the photo below from some time in-between. From the start, it was important to Gardella to have some lining through with the base of the pediment of the adjacent church but, on the other side of the building, the four floors of typical floor windows are also working very hard. Offsetting the lowest floor windows creates a virtual line that Iines through with the balustrade of the building across the alley and, at the other end, lines through with the bases of the pediments above the church’s lower windows. On the building, this virtual line is also the actual line of the differently coloured floor slabs, and that the slightly overhanging lowest slab lines through with the cills of the church’s lower windows. In the as-built building, the number and placement of windows on the two middle floors is identical, and on the felt side is as if one window has been omitted from a line of four. This creates a two-storey high rectangle of blank masonry not unlike that of the building on the other side of the white one. On the church side of the building, our eyes are mainly led to the base of the church’s pediment by the offset balconies on the fourth floor, but also by its windows offset by one bay. All this works to knit the building into its context to make the two inseparable. It’s an example of that peculiarly mid-century Italian view of history as what’s already there.

Milanese architects Asnago & Vender are of the same era and also advocates of this view of history as what’s already there. [c.f. Architcture Misfits #26: Asnago Vender] This windows in their 1935 (via Col Moschin 3, Milan) residential building (left, below) are evenly spaced until the ends where it is as if this infill building was squeezed into its site. The same could be said about their 1952 Apartment Building at via Plutarco 15, Milan, or their 1948 Apartment Building on piazza S. Ambrogio 14, Milan (middle and right, below). With the 1952 building, see how that large window on the fourth floor creates equal spacing between windows of different widths, drawing attention to the entrance?

The spacing of windows in this next building of theirs increases towards the corner and the doubled windows on the second and third floors turn the corner and lead into the adjacent building that they also designed. The height of the windows increases towards the top, making their vertical spacing smaller. I’m only guessing, but I think this might be to make the parapet appear thicker and more like a cornice and termination to the building and, by the same logic, for the corner to appear more of a corner. Both these progressions are confounded in the third bay where the window sizes aren’t the expected ones. Again, they indicate the location of the entrance.

At the far corner of the same block is this building is their 1958 Isolato tra via Albricci e piazza Velasca on Milan’s via Albricci. Its windows do the same things with their height increasing with height, their spacing increasing towards the corner, and four unexpectedly sized windows turning the corner.

The image on the left above shows that this building is also an infill building and the building on its left is their 1950–1952 Mixed-use Building at Piazza Velasca 4, Milan, below. For a mixed use building, we surmise that the upper floors with the larger windows are residential but only the position of the entrance door is determined by the column positions. The three square windows immediately above the entrance appear axial but then … I don’t know. The spacing increases towards the sides, but only in general. I’m lost, and can’t follow the architects’ logic. All I know is that I’ve looked at this facade for far longer than anyone would reasonably be expected to.

Gio Ponti was also a master of the offset window but I doubt he even thought of them as offset windows. With Asnago Vender the offset variations make sense because they set up patterns and then create the variations that are, I believe, aesthetically functional. With Ponti, offset windows are wall decorations and an end in themselves. [c.f. Career Case Study #8: Gio Ponti] Below we have, left to right, his 1963-1967 Building for the Istituto Nazionale Assicurazioni, his 1965 Building 14, Politecnico Milan and two views of his 1964-1970 Montedoria Building, all in Milan.

These windows have nothing to do with what’s already there but then, the Politechnico di Milano crowd had nothing much to do with any architect who wasn’t. Ponti was to freestyle the placement of windows on his later churches even more, but I forgive him his off-piste offsetting simply because the results are so gorgeous. If it was done as an indictor of design intent then, as with most of Ponti’s output, that design intent was worth it. The same can’t be said for the fashion for “offset” windows it prompted later in the century. Offsetting windows was a device favored by the developers of multi-story residential accommodation, presumably because architects’ justifications of “breaking down the scale” seemed to make a certain kind of sense by returning the position of a window to something the size of a person rather than the size of a whole side of a building. I’m most familiar with examples in the UK and Australia but it was a worldwide phenomenon. For a little while, offset windows were good contenders for what this blog one called The Misfits’ Challenge in which persons wishing to “add beauty” to a buidling should be made to prove that what they are adding (1) Does not compromise the performance of the building, (2) Can be achieved without the use of additional resources, and (3) Actually is beautiful. However, after a few decades, offset windows became tropey indicators of non-design, of something that architects just did anyway, and so began to fail the third condition.

When offset windows are neither as gorgeous as Ponti’s or as thoughtful as Gardella’s or Asnago-Vender’s, the question is whether offset windows such as these above are better than not making a design decision at all, or whether they’re merely indicators of no design decision having actually been made? This next image is a recent example but from where I’ve no idea and it doesn’t matter. The placement of windows has no purpose other than to stop their positioning being identical and creating a “boring” building even though this is setting the bar very low. With this building, every third or fourth window is offset left or right by exactly its width and is the bare minimum to indicate the presence of a design decision. But is it beautiful? Someone must have thought it was at least an improvement. I don’t mind it. For what looks like a re-clad and the addition of an extra storey, I like how the spacing of the windows on the mansard has been left regular. The lower floors and mansard have visual differences yet conceptual similarities call attention to the mansard and how it completes the building despite having an identical function. There also seems to be longer vertical runs of windows adjacent to the downpipes. This happens three times and is probably a device do horizontally divide the side of the building into the smaller divisions determined by the downpipes and the inclination of the roof.

Be that as it may, it is still following the path set by Ponti. Diener & Diener architects seem to be the only contemporary practice following the path set by Asnago Vender. Just flipping through my excellent Diener & Diener Housing book [2020 Park Books AG, Zurich], the windows of their 1992 Rue de La Roquette apartments in Paris seem to be determined by column positions while the position of the end windows is determined by the end party walls. The windows are spread as widely and as evenly as possible given these two constraints but the spacing is not uniform. Moreover, a decision has been made to place the windows either side of the alley not as close to the alley walls as they could be and this creation of a third spacing for windows that are neither end nor middle helps bridge the gap.

Their 1995 KNSM and Java Island in Amsterdam has been planned so that there is scope to offset the living room windows, giving rise to facades like these next.

Or how about their 2002 Renaissance (Mombimo Tower) building in Zurich that combines hotel and residential? According to the architects, [the windows] “are wider in the apartments above than they are in the hotel rooms, so that the facades appear more permeable above. This interplay of window sizes makes it more difficult to estimate the height of this otherwise homogenous tower.” I’m not sure about that last bit,

Their 2005 Westkaai apartment towers in Antwerp do a similar thing but with groupings of apartment types into what looks like three different typical floors. On each elevation it’s possible to trace some structural and/or party walls running the full height of the building and creating vertical lines that windows don’t cross, providing the elevations with a somewhat wonky verticality. This, together with reduced number of openings on the lowest level and the generally larger windows on the uppermost level, takes us back to Polacek, if not Sullivan.

• • • 


  • Hi Graham,

    Really great piece as usual.

    Just a small comment regarding the Monadnock Building, Chicago.

    Burnham & Root were the architects for the (famous; load-bearing masonry; non-Classical, non-differentiated facade; no ornament; 1891) massive “pile-of-brick” north half of the Building (designer: John Wellborn Root) which was completed after Roots early death at age 41.

    Subsequently, Holabird & Roche were the architects for the (not-as-famous; steel-framed; Classical “base/shalft/capital” differentiated facade; a lot of ornament; 1893) south half of the Building.

    Interestingly, the north half was subject to cost overruns and limited window area (the old-fashioned piling up so much brick was very bulky and expensive) whereas the south half (featuring modern steel framing) was able to achieve large expanses of glass, was built faster, cost 15% less, was 15% lighter and had 15% more rentable area.

    Peter and Shepard Brooks were the developers. My understanding is that Shepard went with Holabird & Roche for the south half since Root had died and Burnham was pre-occupied with his huge World’s Columbia Exposition commission by that time.



    P.S. BTW, I seem to recall a Spanish proverb (can’t seem to locate the source and the exact quotation, though — sorry) defining an Architect as “one who knows where to put the doors and windows”.

    • Hello Mark!

      Thanks for pointing out my bit of brain fog there. I’d already corrected it though as another astute reader had already pointed it out. I should have known better, having written about the Monadnock before in a post called The Core. Monadnock North is interesting for not having one but then, it was still early days for high-rise office buildings. I’ll try to hunt down the source of that Spanish proverb.

      Thanks again Mark,