In an unfortunately fascinating 1963 collection of essays on computer-simulated personality, Silvan Tomkins, the founder of affect theory, wrote about:
…the tendency of jobs to be adapted to tools, rather than adapting tools to jobs. If one has a hammer one tends to look for nails, and if one has a computer with a storage capacity, but no feelings, one is more likely to concern oneself with remembering and with problem solving than with loving and hating.
Tomkins’ observation, which comes in the middle of an extended meditation on computer personality that is a wild ride for lay readers interested in AI, has concentrated my thinking about the tools that made our current networks: An ever-accelerating explosion of technical possibilities. Silicon Valley relationship networks. Venture capital’s it’s a fountain of money or a failure mandate. Open source’s programming-first culture. A lot of defensive libertarianism, a little charisma. The emotional range of a wellness beverage.
We made what we made because of what we carried in with us.
Consider the difference between two places: The first is a courtyard ringed in verandas or porches that provide shelter from rain and bright sunlight—a half-outdoor experience; it also offers views onto a street or open square, and is criss-crossed by pathways between multiple entrances and exits. The second is a fully enclosed courtyard with a single way in and no views into other open spaces; it has an abrupt transition from inside to out, with no arcade or partial cover at the edges.
In The Timeless Way of Building (last discussed here), Christopher Alexander writes about the first courtyard as alive and the second as dead, for several reasons, all of which I find illuminating.
First, Alexander says, the closed-off courtyard is dead because it produces—and prevents the resolution of—conflicting desires in the people meant to use it. People want to go out, but the stark contrast between out and in is too abrupt to be inviting, and the total enclosure and lack of views produces a claustrophobic feeling and quickly sends those who do venture out back indoors. Without functional pathways that bring people into a custom of crossing through the courtyard, the building’s inhabitants also spend less time there, and the courtyard is outside patterns of daily habit. In these ways, the courtyard fails to strengthen the life and wholeness of the people for whom it was made.
Second, the courtyard is dead because it fails to be self-sustaining. Because it’s uninviting, unpleasant to stay in, and removed from the normal walking patterns of the building, the courtyard becomes neglected. (I would add that commercially maintained but dead-feeling spaces, which are very common in many institutions of modern life, reflect intense upkeep and still project a sense of standoffishness or gloomy abandonment.)
Finally, the dead courtyard is dead because it pushes conflictedness out into the surrounding spaces. Alexander makes a precise argument about this, so I want to quote it at length in all its 1970s-gender-norms glory:
We try to go out, but are frustrated, because the courtyard itself pushes us away. We still need, somehow, to go out; the forces remain within us, but can find no resolution here. We have no way of resolving the situation for ourselves. The unresolved conflict remains underground; it contributes the stress which is building up. First, it reduces our capacity to resolve other conflicts for ourselves, and makes it even more likely that unresolved forces will spill over in another situation. Second, if the force does spill over, it may create even greater tension, in another situation, where there is no proper outlet for it.
Suppose, for example, the people who want to be outside go out instead and sit on the road, where trucks are going by. It is OK. But then perhaps a child gets hurt. Or, even if a child does not actually get hurt, the mother fears for it, and shouts, and conveys a continuous sense of unease to the child, so that his play is spoiled. … In one fashion or another, the effects always ripple out.
You may say—well, people can adapt. But in the process of adapting, they destroy some other part of themselves. We are very adaptive, it is true. But we can also adapt to such an extent that we do ourselves harm. The process of adaptation has its costs. It may be, for example, that the child adapts, by turning to books. The desire to play in the street conforms now to the dangers, and the mother’s cries. But now the person has lost some of the exuberant desire to run about. He has adapted, but he has made his own life less rich, less whole, by being forced to do so.
The “bad” patterns are unable to contain the forces which occur in them.
As a result, these forces spill over into other nearby systems […] the courtyard which fails makes children want to play outside and causes stress and danger in the street.
But these forces make other nearby patterns fail as well. The pattern of the street may not be conceived as a place for children to play. So, suddenly, a pattern of the street, which might be in balance without this force, itself becomes unstable and inadequate. […]
In the end, the whole system must collapse.
I think I blew past all of the parent-child stuff the first few times I read Timeless Way, but now it snags me every time. The parent here is not presented as a neurotic mother or its update, a helicopter parent: The danger of the street is not imaginary. The mother responds to the shape of the built system in which she and her child exist—and then her shouts of anxiety become part of the system through which the child’s personality is formed.
When the subject of social media problems arises, commenters tend to divide into one of two very low-level orientations toward attributing problems (or successes) to systems vs. attributing them to personalities.
It’s not always obvious that this happening, because sometimes the surface discussion is more about whether the root problem is, say, interface design or capitalism, both of which are systems. But it’s rare to get through an online conversation without someone proposing that systems talk is largely a distraction, because some people are just assholes. And indeed, many people are assholes. Most of us, even, given the right contexts and deficits, though it’s clearly overrepresented among the leaders of social media companies. But lean hard enough on personality as explanation for mass phenomena and we get arguments that should be all too familiar, in which poor people are individually lazy, Black men individually scary, mass gun-murderers individually mentally ill, and homeless people individually recalcitrant.
I think it’s obvious that it’s always both, systems and people: People and systems can synchronize and strengthen each other (for good or evil); good personalities can patch bad systems, or sabotage them in service of good; good systems can reduce the harm destructive personalities inflict. But unless you work in therapy or social services, I think systems-level work offers the only practical place to put your lever if you want to move the world.
When I write about our networks as haunted machines, I am writing toward the systems sense of the mess we’re in. When I write about Christopher Alexander, it’s because his work has been one of my touchstones for a systems sense of something better—and better in a genuinely different way.
To be alive in the sense found in A Timeless Way of Building, a system would have to:
- avoid piling unresolveable stresses onto the people inside it,
- maintain its own aliveness through self-sustaining evolution and repair, and
- avoid worsening the life of the systems around it, ranging from peer-level technical systems to things like “civil society.”
This framework is so heartening to me because it torches the constructions of a system’s inhabitants’ well-being and the spillover of generated stresses into the wider world as externalities we need not consider.
The Alexandrian orientation puts these factors at the center of the work, in terms sufficiently specific to make them hard to evade—unlike, say, the “triple bottom line” meant to indicate that corporations should consider worrying about “People” and “Planet” in addition to “Profits.” In practice, this has aways been too woozy and vague to do much besides stretching the humanish skin of corporate social responsibility over the usual forms of extraction.
In the same way, tech slogans like “We put users first,” could mean nearly anything, so they usually mean nothing. The first criteria for Alexanderian aliveness, though, translates into something like, “Create only features, interfaces, and systems that resolve conflicting human desires.” I think that’s something we could use.
Moving to the second requirement in the aliveness criteria—being self-sustaining—clarifies problems of organizational character and pulls me back into adrienne maree brown’s sense of fractal trouble. In my own experience, most organizations that genuinely do a good job serving humans and avoid making the systems around them worse exist in a state of permanent precarity. I think this happens both because those orgs burn out the people who hold the place together, and especially because under modern capitalism, it’s nearly impossible to achieve financial stability when funding essential social support work is constructed as charity.
The third requirement of aliveness—to build systems that avoid worsening the world by dumping outward the stresses they create—cuts to moral vacuum at the tech industry’s core, which is the neutrality that sets the burning world at a professional distance behind UV-filtering glass.
Fine, but what does it mean in our actual work? Let’s begin with a person. Since I’m right here, we can start with me. A handful of the conflicting desires that arise on the networks I’ve been using for a couple of decades:
- I want to have interesting conversations online without participating in modes of interaction I would never tolerate offline, ranging from brigading to individual abuse to soul-erodingly tedious explanation of my own words back to me.
- I want to be visible enough to make interesting friends and attract work that I’m good at without getting stalked, doxxed, or successfully targeted for mass harassment. Relatedly, I want people who are structurally least likely to be heard offline to be granted space and access without being made into sacrificial bait for hate campaigns.
- I want to be able to have semi-secluded conversations with variously durable and ephemeral sets of people without giving up my overall ability to participate in a wider, more public conversation. And I want to be able to be a good host, who can invite people into conversations without exposing them to the common brutalities.
A ~product design~ that resolved these kinds of conflicting desires would be vastly more useful than the notions of “healthy networks” that zoom around every time a tech millionaire’s kids turn thirteen. There’s nothing alive about networks that enforce civility over care, and there’s much more life in honest squabbling than in the weird rictus of LinkedIn. But an emotionally and culturally literate form of social design would work toward spaces that let us be our petty, banged-up selves while structurally lessening the damage our bad days do to the people around us, instead of fanning each minor flare of ill temper into a housefire.
Stopping here feels depressing, so I won’t. There’s so much more life to be had if we start with human cultural patterns and try to strengthen what is alive and good in them.
- When I’m invited into semi-sheltered spaces or moments, I want it to feel like I’m wandering through a particularly nice party—outdoors, day’s heat just subsiding, one cold drink, maybe some lightning bugs?—catching edges of conversations—and then I want to be able to trace unthreaded clusters of people and meaning back through time so I can perceive deeper relationships and shared interests.
- I want to be able to browse through my online acquaintances’ varied interests like I was nosing through their bookshelves while they make coffee without sifting through stilted hashtags or stalking them across sites. I want to skip political slogans in favor of shared pathways into what we actually do in support of the ideas we care about.
- I want systems of trust, recommendation, and vouching for people that work at least as well as the ones I use when I need a local dentist or a good lunch without the bandwagon problems that arise when trust gets conflated with celebrity.
- I want to be able to summon a cloud of expert discussion on any topic, with each analysis clearly situated within the commenter’s background and history—and I want to do it without becoming a human filter for the corrosive exhaust of misinformation campaigns.
- On the purely selfish side, I want social tools to support my cognitive capacity during wide-ranging reading, not to weaken it. I want an easy way to find half-remembered fragments that doesn’t involve me carefully labeling everything I might someday wish to remember. I want a stereotypical little demon to coalesce, sift through what I told people I would email them or promised I’d check out, and send me a list—and then I want it to break down harmlessly into water and salt, retaining exactly nothing, selling out exactly no one.
Tomkins, in the essay I quote at the top of this post, is paraphasing a central idea from George Kingsley Zipf’s Human Behavior and the Principle of Least Effort, a psycho-linguistic treatise that explosively decompresses a single insight about the inter-relation of tools that seek jobs and jobs that seek tools into an argument that includes passages like this one:
As to the incentives, first, of the greedy and venturesome outsider who wants to supplant some y lord of the system, we can see that as the y status of a lord increases, his Ay, income increases proportionally, and therewith his attractiveness to the bold outsider. In short, the y lord’s attractiveness is proportional to Ay. On the other hand, as the Ay, income increases, the number of y lords who have that income decreases according to c/Ay2, with the result that the N, opportunities of supplanting a y lord decrease more rapidly than the Ay, income increases.
…which is the kind of thing you don’t see as much now outside the twenty-five-cent stack at an estate sale, and also extremely funny to me. I’m afraid it’s also reflective of some of the communication problems I have when I try to talk about social forms with people trained to think algorithmically. It’s not the encoding of life’s weird pageant that I’m after; I just want tools better suited to human hands.
As I type this post, Reddit-the-company is steadfastly clinging to their decision to demolish third-party tools critical to Reddit-the-community’s ability to self-moderate, and therefore to produce anything of value to anyone but the edgelords and grifters who form the sticky ring around the bathtub when the water drains away.
Given that we seem to be stuck, as a techno-culture, on something as basic as “don’t actively harm the people whose gift of free labor is your company’s only value,” maybe it’s unhelpful to be thinking out loud about how much better our social worlds could be! I don’t know. But unless we’re going to excise the influence of our networks from our societies entirely, and especially given the whole Gibsonian jackpot situation (spoilers for The Peripheral there), it feels impossible not to try.