If you’ve heard about handshake apartments, they’ll probably be in Shenzhen, China, and in buildings built so close that occupants in neighboring buildings can shake hands from their windows. You can get the idea from these images. They’re intense, but it’s what happens if you want to solve density, access, daylighting and ventilation without recourse to elevators, mechanical ventilation or artificial illumination. The gaps between the buildings are lightwells, airshafts and access paths combined. It’s an intelligent vernacular response to getting things done with a minimum of resources of which land is one.
Many people think it’s a bit messy and third world because people can look into each other’s windows and, not surprisingly, architects eager to establish their visionary credentials in such a lucrative market have been on the case, either adding colorful floors and bridges to the uppermost levels, or cutting into intermediate levels to provide interconnected shared space.
The project below is by Dutch outfit doffice who have completed many projects in Shenzen. This one’s a conversion of 29 out of 35 existing buildings into accommodation more attractive to young tech workers. Seven elevators were added with “sky bridges” on the third and fifth floors linking them to building stairwells and each other. It’s fairly decent if the only problems with handshake apartments are a lack of shared facilities, the difficulty of getting to and moving between buildings, and the absence of elevators.
The problem with adding elevators is that they block what unbuilt space there is between buildings and compromise movement at ground level. The problem with sky bridges above paths is that they further compromise what little light and airflow there is. It also means that making apartments more attractive places to live means people passing 50cm from your window on the 3rd or 5th floors. Forget shaking hands, this is kissing distance. These elevated access corridors also reduce the area of one in four apartments on two of the six floors. There’s little reason to connect buildings at intermediate level other than to reduce the number of elevators and it means nobody will have to walk more than two flights of stairs to get to their apartment. For comparison, connecting the buildings at roof level would mean occupants walk a maximum of three floors down or three floors up from ground level, and that no apartments need be sacrificed.
This second project with similar buildings connected at roof level makes more sense despite its apparent frivolity. If elevators were placed on the periphery then those gaps between the buildings would remain fully functional.
It’s fine to compensate for space subtracted but a lack of space on the roof or a lack of horizontal routes through the district isn’t the main problem – an excess of proximity is. The exploded axonometric below shows four one-bedroom apartments, each accessed from the same landing. Entry into the stairwell is from beneath the stair and is as spatially minimal as stairwell access gets. You can see how the building overhangs the access corridor outside, further reducing the distance between adjacent upper floors. [I hope there’s a mistake in this axonometric below for if the building were mirrored about the access corridor, there’d be no space at all between adjacent upper walls and, moreover, the stairwell half-landings would overlap.] These apartments have gone as far as they can. Architects have nothing to offer that won’t make them less.
Here’s my approximation of the layout of the building above, assuming 2 metres as the length of a bed. There are two types of apartments, mirrored around the stairwell.
Building dimensions are 9m x 12m and the approximately 25sq.m apartments contain a kitchen, bathroom and two habitable rooms. Balconies aren’t a good use of space if the entire footprint is built on. I expect it will be almost impossible to make these apartments unhandshakeable without sacrificing floor area. It won’t even be easy to redesign them so windows in opposing buildings don’t look directly into each other. This next layout gets two out of four bedrooms onto the corners where there’s the best light and ventilation.
Mirroring the layouts around the ground level access path (so the lowered ground level is shared, as is likely), and mirroring them across the access path in the other direction simply creates direct overlooking problems for every window,, while repeating horizontally and mirroring vertically is pointless as the apartment layouts are mirrored already.
However, the blue apartments each have bedrooms with both opposing windows and opposing blank walls, suggesting a pinwheel arrangement in which all bedroom windows of blue or yellow apartments face a blank wall (below). This counts as an improvement, minor as it is.
All living room windows still look directly into each other. This could be solved by shifting the living rooms to the corners and giving them a pinwheel arrangement of windows but this would need to be done without creating problems for the bedroom windows and this is the problem. The bedroom window problem could be solved by having them on the same side as the living room windows and it would still count as an improvement if those bedroom windows still opposed a bathroom window and/or a kitchen window. However, it’s going to be difficult to achieve this for the red and green apartments as they’re only 4m wide.
It worked, although the red apartments suffer and the green apartments still have the bathroom in the best position.
The last thing to try is to see if it makes any difference not having all the sunken access pathways in a straight line. This means that each intersection will have one of each type of apartment.
It didn’t make that much difference, although the layout of the yellow apartment had to change and whether for better or worse I don’t know. This is as good as it gets for the main parameters of 1) no directly overlooking habitable room windows and 2) no reduction in floor area.
- I would prefer to plan an apartment of this area as a studio but I’ve preserved the separate bedroom because a family of four might be living in this apartment.
- My personal preference would be for an open kitchen but that is not how the Chinese prefer it and I appreciate that.
- The pinwheeled layouts preclude mirroring and mean each apartment drains to a separate riser.
So what’s next, now there’s one acceptable way for the windows to be arranged?
- It would be good if the layout had drainage and service risers shared by at least two apartments. Pairing kitchens and bathrooms around building indents will reduce internal floor area and just might make for more workable layouts.
- It would also be good if those kitchens and bathrooms were identically sized and shaped.
- Moreover, although one of the initial parameters was to have no reduction in the internal floor area, it’s not as if the floor area there is, is always being used the most effectively. This is largely a consequence of the entrance doors being in one corner of the layout, but there’s also a width constraint due to the unstated requirement for all kitchens and bathrooms to be naturally ventilated. This is most apparent in the red apartment.
- Ideally, I would like to solve similar problems in similar ways. Because there is only one staircase on one side of the building, there’s nothing that can be done about the different proportions of the apartment without changing the dimensions of the block. Here’s my preliminary sketch for an improved handshake apartment. The overlooked construction gaps still serve as light wells and ventilation shafts but construction is more rational and there is increased visual privacy. I don’t know if this is going to work.
It didn’t happen exactly how I’d thought. The largest problem, as ever, with corner access to an apartment, was how to get from the front door to the rooms using a minimum of space along the way. Securing a block of space for the bedrooms also wasn’t easy.
- The side indents mean the outer wall doesn’t have to be used to ventilate the kitchens, although it does of course increase the surface area of the building. This can’t be avoided.
- I had planned to have those indents ventilate the bathrooms as well as the kitchens and, although there was sufficient apartment length to do this in the red and green apartments, there wasn’t for the blue and yellow and I gave priority to solving similar problems in a similar way, and to have identical kitchens and bathrooms.
- The re-entrant corner indents allow bedroom and living room windows to pinwheel around the intersection. It is possible for certain green and red living rooms to have an additional high-level window as overlooking is a problem for eye-level windows only. There’s nothing wrong with seeing a neighbours’ lights on.
- Both types of indent represent 2.5m2 of floor area “lost” but compensated for by increased daylighting and better ventilation for the habitable rooms, whether or not the trees happen.
- The stairwell is responsible for the difference in apartment sizes but the red and green are wider with their living rooms large by the width of half the stair, and that area is compensated for in the blue and yellow apartments by the longer living rooms with more wall space.
- Areas of green/red apartments and yellow/blue apartments differ by 1.2 sq.m but could easily be made equal.
- The building still needs some construction logic imposed, but not today.
- The trees would be nice to look at but they’d only reduce available light. Plus, there’s almost certainly important pipes beneath those paths. It’d be better to let people grow their own plants in a planter outside the window of the balcony shared with the reverse-cycle A/C compressor.
I don’t think anything more can be done without rethinking the building type itself, or the dimensions of the plots. It might be possible to use less corridor to link four stairwells. It’d also be good if there were a more equal distribution of daylighting across all apartments instead of half being on the dark side. The challenge now is to achieve all and a modicum of privacy when the only feature these buildings have is a hardworking two-meter gap between it and its neighbours. I’ve been interested in polyfunctional access corridors for about a decade now but I never thought light and air passing through a 50cm gap between a window and an access corridor would ever be as good as it gets. We should think through some options just in case.