Archive‎ > ‎2014‎ > ‎

Reflections on PUARL 2013: hope & surprises

January 16, 2014

If you haven't read any book by the writer and architect Christopher Alexander (CA), these reflections won't make much sense to you. 

I won't provide a sketch of his influence over the last 50 years. Let's just say that he's done respected research into what we mean by good cities, good design, and humane buildings. It's easiest to start by thumbing through the sections of his 1977 book A Pattern Language. To anyone interested in helping the world, and in understanding what a healthy community looks like, it covers extremely important territory.

So, it's not surprising that there's an international conference of planners, activists and architects focussed on continuing his research. It's hosted by the University of Oregon's Portland Urban Architecture Research Laboratory, or PUARL. But it's strange that this biennial conference just started in 2009. Especially since there have been countless conferences in other fields that were inspired by Alexander's ideas. Especially in computing.

Still, no complaints -- this conference made me quite happy, and I think that's true for all its participants. It provided a tangible sense of community among people scattered across the world.

I've been writing about CA's projects in RAIN for a couple of decades, and have done a few projects with him. We had unusually long discussions about experimental and theoretical issues. Many former colleagues, co-authors and students of his attended this conference. Although the majority at PUARL was not so tightly associated with CA, all had terrific practical experience, and were obviously very sensitive observers of the human condition. They'd been inspired by the books, or by people in his circle. A dizzying range of projects was presented -- this was not a group of armchair researchers. The conference was rich with fascinating detail.

But, I was surprised about what was not discussed.

Innate Judgement

One was the criteria by which we form judgments about what is 'good'. Students rightly asked "why is this approach to architecture the right one?" I wasn't on the panel tasked to answer that question, but I'd like to respond now.

Alexander spent great energy convincing his colleagues that the standard for 'making something alive' is not subjective. To see that, one needs to dig beneath torrents of training that everyone receives these days. We're increasingly trained to respond to stuff that's flashy, or trendy, or familiar, or symbolic. But once researchers get underneath this training, they find that everyone shares an important faculty of judgement -- they agree on which things 'have more life', or 'reflect their true self'.

This approach to the study of instinct, using an informant for judgment, and trying to present judging 'problems' or experiments that can provide insight into internal mental structure, is fundamentally the same kind of research that ethologists like Tinbergen and Lorenz did to uncover the structure of animal instinct, that Chomsky uses to uncover ever more about the structure of syntax, and which researchers like Marr and Ullman use to determine the structure of human visual perception. These techniques have a long history in philosophy, but were only accepted into modern experimental science after the trend of behaviorism lost credibility in the early 60's. 

Now, these kinds of informant-based scientific inquiries are integrated into the study of human biology. These are terribly complex research subjects. One of the hardest parts of the study, is teasing out factors for any human faculty or feature, which fall into essentially three categories:

1) genetic or inherited or instinctive
2) stimulated or developmental or learned
3) physical laws of nature

So, for example in human language (I'll user the example in Michel Gondry's excellent and accessible Is the man who is tall happy?) a given sentence can be found to have an instinctive factor (hierarchical structure), a developmental factor (the actual selection of phonemes that shape the sound of a word), and a biophysical factor ('minimizing', say, during the traversal of the hierarchical structure to form a question from a declaration).

Christopher Alexander's work in The Nature of Order showed, I think quite convincingly, that our judgment of life is also innate. It also argued extensively that this innate faculty is deeply entangled with the natural laws of physics, chemistry, biophysics, etc. Now, I had extra help from CA in understanding these ideas while reading drafts of his book ... so I'm not sure that the books by themselves convey this understanding that there is a phenomenon of life among our faculties of judgement.

But that is the basis for CA's work, at least in his opinion (and mine). So I was rather surprised to see no one mention this during the discussions of criteria and qualities. There might have been a fear of seeming a bit too 'woo-woo' ... but it's straightforward (through hard work) to address these issues within the natural sciences, so I'm hoping more people will come to terms with this at the next conference.


The concomitant surprise was to hear no mention of a major tool that CA believes best evokes the experience of unfolding a living structure, the sequence. I was happy to be "the sequence advocate" at this conference, because that was the heart of the work CA and I were trying to convey with gatemaker, among computer people -- and that was the subject of one of my talks. 

I must admit, not only were CA and I not successful at convincing computer people about the importance of this kind of sequence, I haven't even been able to convince them to take a serious look at gatemaker's evident psychological effect -- its ability to engender the smooth production of high-quality living structure -- as an objective phenomenon. That said, I must point out that computer science is unfortunately not a natural science, so it's not surprising that computer people weren't sure what they should do with prima facie evidence of an 'unfolding sequence effect'.

Living Structure

So, that was a third surprise: sequences help produce living structure. But where was that structure mentioned? A smattering discussed some of Alexander's 15 properties. There was certainly passionate discussion about places that were good for people and the environment. But little about the actual geometry of good places.


Instead, I saw patterns, pattern languages, local typologies of patterns, new categories of patterns, et cetera. That was the fourth surprise.

Now, the more senior presenters did not really mention patterns. Researchers need to describe interesting things that they find, techniques they've discovered, properties they've noticed, and principles they've uncovered. They could choose to call all these things 'patterns', but that's silly: if everything is a pattern, then the idea really has no useful meaning. So, there was an overuse of patterns, something that is, it must be said, far more common in the computer industry.

But, even putting that kind of terminological enthusiasm aside, many people weren't familiar with the deeper discussion over whether or not patterns are important.

Christopher Alexander gave a lecture, almost 20 years ago, where he said, quite explicitly, that the 'standard format' patterns, found in A Pattern Language, might occasionally be useful for communication, but they don't really address the critical issues. He explained why.

It's important to understand this. The hidden innate judgment, described above, provides the criteria used to find good patterns, even if you're not aware of it. This judgment allows good structure to emerge, makes diagnostics possible, identifies important generalities and important details ... and of course it allows us to explore the nature of this inner faculty, and the things we produce with it

In a real project in the outside world, even after all the important exploration and analysis is done, it's this inner gift, which we all possess, that enables us to unfold a project with a wonderful functioning structure. Patterns, pattern languages, specific project pattern languages, generic core pattern languages -- these were only research projects for uncovering the nature of our critical judgement, enabling us to build a thing that is harmonious with its surroundings and its purpose. In this sense, patterns distract us from using the better tool.


I'm very hopeful about the future of this diverse PUARL group, and look forward to its meeting in Austria next year. It's the sign of a very good conference when the activist-attendees don't want to leave, so they can work together to make positive change.

I think before the next conference, we'll see online interaction (and maybe I'll initiate it), towards a large-group collaboration around common goals. We have all these tools for addressing difficult urban problems -- over half-a-century of work, and millions of hours of experience -- and we need to find ways for the world-at-large to take advantage of them. 

I'll post any progress here at RAIN. 


[Thanks to Olga Volchkova for letting me clip from her painting, and for her photo of the candles.]