How to design a house to last 1000 years (part II)

(For part I of this series, click here)

Designing for decay and drift

Designing our house to survive disasters (natural or otherwise) is important, and is something building codes and design standards are well-equipped to address - we can get significant survivability simply by designing our house to the standard that things like power plants and tornado safe rooms must meet. But these sorts of events account for a relatively small fraction of homes that get removed from the building stock. Looking at this graph from O’Conner 2004, which recorded statistics about several hundred buildings that were demolished, we see only 7% were demolished due to anything like a disaster (in this case, fire damage).

Reason for demolition of a sample of 227 buildings

We see this even in years with extremely destructive natural disasters. For instance, 2018 was an extremely destructive wildfire season in the US. It was severe enough that it doubled the total amount of fire damage in the entire country that year, and nearly 25,000 structures were destroyed (most of which we can assume were single family homes). But if houses leave the building stock at a rate of 1% a year (which is what we see),  and there are ~100 million single family homes in the US, that means there were another 975,000 homes torn down for other reasons.

US annual fire loss, via NFPA

Looking at the graph above, we see that the majority of building demolition falls into roughly two categories:

• Maintenance costs are too high, or building hasn’t been maintained and is in poor condition

• The owner would prefer a different building on the parcel of land, such as a newer or larger house, or a more intensive use of the space

We can think of these as the result of something like a decay process - over time, the various systems of a home experience wear and tear and will need to be repaired or replaced. If people stop performing basic maintenance (so that damage accumulates), or large, expensive systems reach the end of their life and need replacing, or if the decay process follows something like a Gompertz curve and gets exponentially worse over time, the cost of repair may be high enough that people will prefer to just tear the house down rather than try to repair it.

In addition to physical decay, there’s also a process of “cultural decay” or “cultural drift” that occurs. Over time, the assumptions that went into the design of the house become less and less true, and the house will become less and less attractive compared to other things that might occupy that parcel of land. A house built in the 1950s might remain perfectly serviceable, but if it has just a single bathroom and no easy way to add a new one, that limits its appeal to today’s buyers. Or if the surrounding area has massively urbanized, the land might have become so valuable that the owners are incentivized to make better use of it. My home, for instance, is built on a plot of land that used to be occupied by a farmhouse. The farmhouse got knocked down and replaced by ~30 closely spaced single family homes, because the immediate area is seeing massive population growth and there was profit to be made in doing so.

Decay will be both a function of the physical characteristics of the house (how durable the various building systems are, or how easily it can be changed to suit an owners’ needs), and how willing people are to invest in maintenance and upkeep. With sufficient motivation and investment, buildings can be maintained virtually indefinitely regardless of the technology used to construct them (assuming you don’t adopt a very hardline stance on the Ship of Theseus problem). The Ise Jingu temple complex in Japan has survived for well over 1000 years despite using a relatively simple timber frame structure. It remains because every 20 years, the main structures of the temple are completely rebuilt from scratch. But most buildings are not sacred shrines, and we can increase the odds of our home surviving by reducing the amount of maintenance work required to keep it in service.

I’ve previously mentioned the “pace layers” concept popularized by Stewart Brand, which provides a useful way of thinking about this. In this formulation, a building consists of several different layers, which age and get replaced at different rates. Some layers (such as the foundation) last for the entire life of the building, but others (such as the mechanical and electrical systems) will get replaced much earlier. 

In this framework, we don’t want to stop decay-based processes completely - there would be little utility in designing an electrical system to have an incredibly long lifespan, for instance, as it will almost certainly be replaced by newer, more capable technology at some point. Instead, we want to a) reduce the decay on the most permanent parts of the building, b) reduce the overall maintenance burden as much as we can, and c) make it as easy as possible to replace the faster pace layers and keep the house updated.

Physical decay

Our number one priority for avoiding physical decay (and the associated costs of repair and maintenance) is to keep the water out. Our number two priority is to prevent water from damaging the building when we inevitably fail to do this.

A huge number of decay processes can be traced back to water finding its way into the building. Rust, corrosion, rot, mold, settlement, salt crystallization, sulfate attack, and freeze/thaw cycles are all due to water being somewhere it shouldn’t. To fight this, we’ll want to construct our home with well-vetted design details that will successfully keep water from intruding. Building Science Corporation is an excellent source for these, including a series of assemblies (supposedly) designed for a 500-year lifespan. Things like a properly installed water resistant barrier, eaves that direct water away from the building, and simple roofs with steep pitches that quickly and effectively shed water are all design strategies for preventing water damage and intrusion.

But for the time spans we’re talking about, this probably isn’t sufficient - it’s likely water will make its way into the building at some point over the course of 1000 years. So we’ll also want to use corrosion resistant materials, especially for the more permanent pace layers (the structural framing, the foundation, and the exterior ‘skin’). Stainless steel, galvalume, unreinforced concrete, unreinforced masonry, slate, and stone are all materials that have demonstrably long lifespans - if we look at a list of surviving old buildings, their construction is dominated by corrosion-resistant materials such as these. 

A track-record of long-term survival is important - for many modern materials (such as OSB), we don’t have good data on extremely long-term survival, and they’re often manufactured using glues, resins, or other components where it’s hard to predict how they’ll perform over long periods (and are also likely to be damaged by water exposure). If we can avoid having to replace the foundation (because the rebar has corroded) or the framing (because the resins in the OSB have degraded), that obviously gives our house a much better chance of survival.

Most other physical decay problems are a distant second after water exposure (and they will in general also be reduced by picking materials from a corrosion-resistant palette). UV exposure will damage polymers like plastics, glues and resins over time, so we should limit their exterior use (though things like vinyl siding will have a UV-protectant to avoid this problem). Large temperature swings can cause gradual damage even in the absence of freeze/thaw action, which suggests locating our house somewhere in a moderate climate. Termites can potentially cause severe damage, but are unlikely to be a problem for the materials we’ve chosen.

Gradual settlement and ground movement is another potential source of long-term damage, one that will be difficult to fix if it occurs. Long-lasting buildings tend to be built on large, substantial foundations to prevent this. Our best bet here is avoidance, by using some sort of deep foundation system.

None of this prevents damage from occurring - our building is subject to entropy like anything else. But it should dramatically reduce the burden of it.

Cultural decay and drift

Cultural decay (or cultural drift) is what we get when a building’s design stays fixed, but society gradually changes around it. This isn't something that’s especially easy to accommodate with the design of the house itself. But we can try to guess at which trends are least likely to change, and use that to inform our house design.

Consider, for instance, the basic architectural design of the house. A “trendy” house design has a good chance of looking dated and unappealing after a relatively short period of time (80s modern houses are often offenders here). We’re probably better served by picking a ‘classic’ house style, one that has been popular more or less continuously. This will obviously vary depending on the region - in the US, this might be something like a Cape Cod. We’ll also want to be strategic in how big we make it - small enough that maintenance and upkeep doesn’t become burdensome, but large enough to be appealing to a wide range of buyers.

Choosing a single family home at all for our building can be thought of in terms of picking stable trends. Most building types have changed significantly depending on the needs of the economy and the technology available - the office building is a relatively modern invention, and factories changed from tall, thin buildings to short, flat ones as technology changed. But single family homes are much more constant throughout history.

This strategy can also inform where we build the house itself - we want to pick an area that’s likely to still be somewhere people want to live over the next 1000 years. It doesn’t matter how robust your design or beautiful your architecture is if no one is interested in living where the house is built, as plenty of abandoned homes in Detroit or Baltimore can attest to. 

This isn’t trivial - looking at a list of largest cities in 100 AD, for instance, shows that many of them would later be abandoned, razed, or destroyed and later rebuilt (not ideal for our house’s survival chances). Our best bet here might be to bank on some sort of Lindy Effect, and pick a place that has a long period of being inhabited (and not just inhabited, but a popular urban area) - somewhere like London or Rome.

And we can also extend this principle to the systems we use to build the house - we’ll want to pick systems for the permanent elements that are as likely as possible to still be in use in the future, or at least able to be understood and modified by any future construction workers. This pushes us away from anything non-standard, proprietary, or that requires a complex production process or supply chain. This sort of failure is what will ultimately doom the Lustron Houses, prefabricated houses that turned out to be surprisingly durable (many of them survive today in excellent condition despite being 70 years old) but are essentially impossible to obtain replacement parts for. Things like concrete, masonry, or timber on the other hand can often use locally available materials, and can draw from a large pool of commonly available skills.

Of course, it’s hard to predict skill trajectories for the next 1000 years. Things like stone masonry or roof thatching were staple skills of civilization for centuries, until they weren’t. When building technology changed, the pool of skilled workers dried up, and many of these craft skills are becoming endangered. But sticking to something resembling current standard practices should at least tilt the deck in our favor.

One successful strategy for ensuring a building stays maintained has been to have it stewarded by a long-lasting organization, like a church or a government. But while it’s easy in retrospect to see which organizations ended up surviving, picking future winners is harder. The modal organization has a shorter lifespan than the modal building, and it’s not obvious to me which modern organizations are like the Catholic Church (which successfully maintained its buildings for centuries), and which are like the worship of Zeus or Marduk (which faded away and took their buildings with them).

Repair and replacement

No matter how careful we are with our design, we’ll inevitably accumulate damage that will need to be repaired. And no matter how much we design the building to be timeless, future owners will inevitably want to make changes. The easier it is to make repairs and renovations, the better the chances for survival.

So, for instance, we should make it as easy as possible to swap out parts, favoring things like bolted connections over welded ones, batt insulation over spray-foam, and mechanical connections instead of epoxies or glues. We should also aim for components that are small and light enough that workers can manipulate them without needing a lot of heavy equipment.

We should also make the design of the building as legible as possible - drawings and other documentation are almost certain to be lost over 1000 years, so it should be possible to understand how the building works via visual examination. This pushes us away from things like reinforced concrete, where much of the capability is a function of the steel reinforcing, which is hidden from view once the concrete is cast. One option for ensuring our house stays legible might be to emboss material properties and design information onto the structural frame itself.

To keep the building as adaptable as possible, we want to make sure that the more permanent, slower pace layers are decoupled from the faster pace layers. It should be as easy as possible to replace the mechanical and electrical systems, rearrange the interior partitions, add or remove bathrooms, etc. So we’ll want to avoid, for instance, load bearing walls, which would make it harder to rearrange the interior room layout. This also pushes us away from slab-on-grade construction, which requires running building services such as plumbing and electrical through difficult-to-remove concrete.

Low road to high road

Our design strategy so far has been based around ensuring our building stays economically viable, by minimizing the costs of keeping it useful. But if we want to guarantee our houses’ survival for the full millenia, it’s unlikely we can get there by relying on pure economic calculus. Most of civilization’s longest surviving buildings tend to be culturally important ones that people have devoted time, money, and effort into preserving in the absence of any sort of commercial gain. Historic England lists no houses built before 1000 AD, for instance, but lists over 50 churches. At some point our house will no longer be economically viable, and we’ll need to rely on cultural value to justify keeping it around.

Relying on cultural value will also allow us to escape the trap of urban land becoming too valuable to justify a lone single family house. If the building becomes culturally valuable enough, it can justify huge investments or otherwise foregone economic benefits - the restoration of the Notre Dame Cathedral, for instance, will cost in the neighborhood of $15,000 per square foot, which is as much as the most expensive urban real estate costs.

So our goal will be to reach “cultural value escape velocity” - ensure the building survives to the point where people want to keep it around based on its cultural value alone. Fallingwater is an example of a house that has done this. Once a building accumulates enough cultural value, social infrastructure tends to spring up around it to ensure that it continues to survive. But the longer it can survive purely on economic grounds, the greater the chance that this process will occur - a house 100 years old may or may not be worth preserving, but one 500 years old almost definitely will be.

Other than simply surviving long enough for age-based value to kick in, we can help this process along by designing our house to be the sort of thing that people want to preserve. This means making it, on the margin, architecturally impressive and beautiful - injecting some of the numinous that makes cathedrals such attractive targets of preservation efforts.

There’s some drawbacks to this approach. For one, cultural value alone isn’t necessarily enough to ensure our house’s survival. Plenty of culturally or architecturally important buildings haven’t survived. And this process probably only works for a relatively small number of homes - if every house had a 1000 year lifespan, it wouldn’t be considered important to preserve every one.

And if we’re not careful, this process might actually hamstring our house’s survival, if it’s considered too culturally important to change or modify (hamstringing its economic value) but not culturally important enough to attract preservation efforts.

England’s listed buildings program gives a glimpse of what this looks like. Buildings deemed of cultural importance are marked as “listed”, which places restrictions on what can be done with them. England has over 400,000 listed buildings, including nearly every one built prior to 1700. But listing is no guarantee that a building will be kept in good repair. It seems clear that even for culturally important buildings, keeping it usable and adaptable remains important for all but the largest, most impressive buildings.

(we’ll conclude next week with Part III, where we take what we’ve learned to lay out the basic design and systems of the house).