Christopher Alexander's architecture for learning

Link post

On the 17th of March 2022, Christopher Alexander, the architect and mathematician, passed away. Alexander, whose intellectual influence extended far beyond architecture and urban planning, gave the impetus to several central ideas in modern software development, such as wikis, agile, and object-oriented programming. But his main contribution was in vernacular architecture – that is architecture without architects – where houses are built gradually by the people who live in them.

To honor his memory, I would like to take this moment to reflect on some of his ideas in the area where he has influenced me the most: on how architecture can unlock learning in society. How can we best make sure that the knowledge we need is sustained over generations?

For knowledge to pass from one generation to the next we fundamentally need three things. Firstly, we need ways for the young to access the environments they are to master so they can learn through imitation. Secondly: for skills that are hard to learn through imitation, we need deliberate instruction. And finally, for emotional support, we need houses where children can seek refuge from their families or get support when their parents are busy.

These points are not original on their own; schools try to do at least the last two. What is more interesting is how thoroughly Alexander went about solving it. He proposed not a new type of school but a series of architectural patterns that would weave the functions into the very fabric of society. Access to environments, deliberate instruction, and safe homes would be a seamless part of everything – an “educational system so radically decentralized” that it “becomes congruent with the urban structure itself.”

At the heart of this was a pattern he called network of learning.

Network of learning

The book where Alexander most clearly lays out his ideas about learning is in his 1977 cult classic A Pattern Language: Towns, Buildings, Construction, which he co-wrote with Sara Ishikawa and Murray Silverstein.

It is a peculiar tome: a choose your own adventure for how to build a modern medieval city-state.

It is quite dizzying in scope and detail. Together with its companion pieces, The Timeless Way of Building and The Oregon Experiment, the book spans 1912 pages and contains 253 architectural ”patterns”. Each pattern describes “a problem that occurs over and over again in our environment, and then describes the core of the solution to that problem, in a way that you can use this solution a million times, without ever doing it the same way twice”. These patterns range from the world defining to the mundane – from how to beak up the countries in the world into city-states and incorporate them in a world federation to how to keep weeds from stone walls and what colors to use to make a room homely (red, yellow, and brown!).

The patterns are mostly derived from close observation of places from the past that not only function well but are full of life – Alexander and his students were practicing a sort of ethnography of vernacular architecture.

The most fundamental educational pattern in the book (18: NETWORK OF LEARNING) is derived from Ivan Illich, the catholic priest and historian famous for his critique of institutionalized care. Illich had purposed turning schools – the epitome of institutionalized care and its ills – inside out and bringing forth instead learning webs. Alexander took this idea and turned it into a design pattern. The way he did this is akin to how an object is created from a class in object-oriented programming: he created an instance of the Illichian idea, by pasting several page long quotes, and then he set about modifying it. I will return to how he modified it. But first, let’s look at what Alexander and his collaborators inherited from Illich.

The basic impulse behind the network of learning pattern is an observation. Most of what we know we have learned not through formal educational structures – like schools – but simply by living. We soak up the culture that surrounds us; we pick up things from blogs we read; we watch how-to videos on YouTube; we figure stuff out by talking with our friends. Most knowledge, in this manner, reaches us not through a curriculum but through a decentralized network of connections to other humans and their artifacts.

By consciously designing infrastructure and services to facilitate the growth of these connections we can create a society where “living and learning are the same”, writes Alexander. By crafting the right infrastructure we can enable and even unleash learning.

Instead of “social control through the schools”, he quotes Illich, what we want is to enable “voluntary participation in society through networks which provide access to all its resources for learning.”

The main difference then between this pattern and the educational pattern used in modern societies is that learning is voluntary in the network model but mandatory in the school model. This change from mandatory to voluntary learning changes the entire design space. Instead of trying to steer learning, you have to find ways to unblock it. You have to, as I have argued elsewhere, enable the learning system.

Now, what is the difference between Illich and Alexander? It lies in how they purpose to facilitate voluntary learning. Illich focuses on reference libraries and services that allow people to access tools, information, mentors, and peers interested in similar things. Alexander is more interested in designs that encourage people to work in public, making their knowledge visible to others.

Working in public

The need for children to have access to the world of adults is so obvious that it goes without saying. The adults transmit their ethos and their way of life to children through their actions, not through statements. Children learn by doing and by copying. If the child’s education is limited to school and home, and all the vast undertakings of a modern city are mysterious and inaccessible, it is impossible for the child to find out what it really means to be an adult and impossible, certainly, for him to copy it by doing.
– Christopher Alexander, PATTERN 57: CHILDREN IN THE CITY

If children cannot navigate their community on their own, they are severely limited in their capacity to pursue their interests and form the connections that facilitate learning. Enabling this is largely a question of culture. In post-war Germany, children were allowed to roam the ruins of bombed-out cities – an experience Werner Herzog still raves about. In Glasgow, in the mid-twentieth century, children let loose in the city would coordinate in giant playgroups (sometimes numbering in the thousands!) going on vampire hunts, trekking Springheeled Jack and the Grey Lady, forming militias, stalking graveyards – and other such things that are entirely out of bounds today and which most likely taught them plenty about organization and collective action.

Not unlikely, quite a few of these children got traumatized or hurt or died in the ruins and the churchyards. So wanting to reduce the risks for young people navigating the community is not unreasonable – all things equal.

One hazard in modern cities is cars. To circumvent this problem, Alexander purposes the construction of special bike lanes suitable for children. These bike lanes would cross the roads in tunnels or on bridges when necessary, but generally keep away from cars altogether, instead passing “along and even through those functions and parts of a town which are normally out of reach: the place where newspapers are printed, the place where milk arrives from the countryside and is bottled, the pier, the garage where people make doors and windows, the alley behind restaurant row, the cemetery.”

In this way, the city can be, yet again, opened up to children.

In every domain, we want to make sure that activities are as mixed as possible so knowledge can spill across. Talking about universities, for example, Alexander argues, we need to keep them from forming secluded enclaves. We want the university, like all work, to spread out through the city and operate openly (PATTERN 43: UNIVERSITY AS MARKETPLACE) – scholars congregating in cafés and public libraries, perhaps, or in labs in residential areas, where children playing in the park can observe the researchers breeding mice or doing brain scans or splitting atoms, just like they can observe garbage collectors and policemen.

There are encouraging signs of this ethos of working in public spreading online. Live streaming is growing across all knowledge domains. Communities of practice are forming on Discords – where people share learning resources, work on projects together in public, and do community calls with experts. There is also an increasing number of researchers that keep open notebooks where you can glimpse into their ongoing work (see here, here, here, and here). Some independent researchers even use hanging out in public forums as their main research strategy.

A few weeks ago, talking to José Ricón, an independent researcher mainly focused on longevity studies at the moment, I asked him how a random Hungarian teenager can get involved in anti-aging research. “Well”, said José, “they can just google it.” (Ricón’s Longevity FAQ is also a good place to start.) After having read the popular sources, you go to the research papers they cite (which you can unlock with Sci-Hub). You read those and look up the researchers that are active on Twitter – where you can see how they debate, search through their past dialogues, and ask questions.

Since many important seminars are held online, a random Hungarian teenager can easily, by lurking in the community, sign up and join the seminars to listen to the latest findings and the gossip and locate the open questions. This was how Ricón himself got involved in the area. There are of course still obstacles to be overcome—how to get hands-on experience breeding mice? how to access a PCR machine? – but we are seeing teenagers move into the space. There is a pathway.

How can we bring this openness into the offline world?

Alexander purposes a provision for people that build street-facing home workshops (PATTERN 157: HOME WORKSHOP):

[This pattern] brings the workshop out of the realm of backyard hobbies and into the public domain. The people working there have a view of the street; they are exposed to the people passing by. And the people passing learn something about the nature of the community. The children especially are enlivened by this contact. And according to the nature of the work, the public connection takes the form of a shopfront, a driveway for loading and unloading materials, a workbench in the open, a small meeting room . . .

University as marketplace

However, not everything can be learned through immersion. Sometimes we can accelerate our growth by submitting to a master, or teacher, who disciplines us to practice – a violin teacher demanding we run through scales for hours; a combat simulation during military training; a math professor doing a problem on the blackboard to tease out the mental operations we need to internalize.

(This is best seen as playing second fiddle to what we discussed above: it is the doings in the real world that gives structure and motivation to our learning. But deliberate instruction can unblock us in our work.)

To accelerate learning in society, we, therefore, need to make it easy, and affordable, for people to get access to classes, mentors, and teachers that can give them support. And we need to provide a large range of different support structures – since people differ in their needs and aims.

One metaphor that Christopher Alexander uses when describing the infrastructure needed for this is a university as marketplace (PATTERN 43). By this, he does not mean, primarily, a market in the economic sense. He means the actual, physical marketplace with its stalls and its small vendors, its smells and shouts, the alleys that meander from shop to shop, and the bustle of it that bleeds out into the surrounding city. The learner should navigate this market, looking for the support structures that will help her accelerate toward her goals.

This was how the early universities were organized. Writes Alexander:

The original universities in the middle ages were simply collections of teachers who attracted students because they had something to offer. They were marketplaces of ideas, located all over the town, where people could shop around for the kinds of ideas and learning which made sense to them.

If we want to help learners access environments where they can work deliberately on their skills, we need to nurture these types of decentralized exchanges – offline, as well as online. We should have teachers walking the city (PATTERN 85: SHOPFRONT SCHOOLS) and apprenticeships (PATTERN 83: MASTERS AND APPRENTICES). We need to shoulder the work of connecting people with each other and the resources they need. To use the Illichian terminology: we need to take on the roles of network administrators, pedagogies, and skill models:

While network administrators would concentrate primarily on the building and maintenance of roads providing access to resources, the pedagogue would help the student to find the path which for him could lead fastest to his goal. If a student wants to learn spoken Cantonese from a Chinese neighbor [a skill model], the pedagogue would be available to judge their proficiency, and to help them select the textbook and methods most suitable to their talents, character, and the time available for study. He can counsel the would-be airplane mechanic on finding the best places for apprenticeship. He can recommend books to somebody who wants to find challenging peers to discuss African history. [Italics added.]

When connected, people can meet in each other’s homes, in libraries with special equipment, or in simple shopfront schools, which can be kept much cheaper than big mass institutions since there is no administrative overhead and special building requirements. In this way, we can create a rich and decentralized ecosystem of learning services, that help people in their diverse aims.

And now we almost have an education – the only thing lacking is emotional support and logistics.

Second homes for children

...in a society where most children are in the care of single adults or couples, the mothers and fathers must be able to have their children looked after while they work or when they want to meet their friends. This is … the adult’s view of the situation. But the fact is that the children themselves have unsatisfied needs which are equally pressing. They need access to other adults beyond their parents, and access to other children; and the situations in which they meet these other adults and other children need to be highly complex, subtle, full of the same complexities and intensities as family life—not merely “schools” and “kindergarten” and “playgrounds.”
- Christopher Alexander, PATTERN 86: CHILDREN’S HOME

To meet these needs, Alexander purposes we set off large, rambling homes in each neighborhood where children can congregate. There should be a core staff of two or three adults; but more than an institution, it should be a home, where at least one of the adults live. It does not open or close. It simply is.

And unlike schools, it is not to be fenced off from the surrounding world. It is to be built so that people walking by pass through it, rather than around it, to tie it naturally into the city (PATTERN 101: BUILDING THOROUGHFARE). Ideally, it is co-run with a café or a small store, or some other community facility, that brings people into its orbit.

Using this home (as well as their family home) as a base, the children can make exertions into the city to observe the goings-on – perhaps by biking in a glass tunnel through a research lab! – or to join a class or an apprenticeship. And then they can return home – to be safe, to play, to talk, and be heard.

Like all of Alexander’s visions, there is something deeply human about this image. He wanted to shape our cities after us, rather than discipline us to serve our cities. There is also something utopian about this. But it is a utopia that is within reach – each individual pattern is simple, and can be implemented from the bottom up. We can, and do, shape the world we live in, at least locally.

22 march 2022. It is the evening of the fifth day. As I’m writing these final words, A Pattern Language flung open on the floor, my two daughters are sleeping in the summer house. Not knowing how to end things, I close the laptop and walk up the terrace to look in on them. The stars are clear tonight. The four-year-old has dropped an arm from the bed, the fingertips precisely reaching the floor. Alexander is no longer among us, I scribble on a piece of paper I hold against the window. He will never learn again, never grow.

But we live on in his pattern – I can see them shape my daughters.

In the distance, a tail light shines red across the valley. A low horn. I crumble the paper and go in to lay down beside them.

Christopher Alexander’s architecture for learning