«

two concepts of legibility

2020

In Seeing Like a State, James C. Scott defines “legibility” as something a state imposes on its people and resources. It is a coercive abstraction, not only treating different people, places, and ways of life as if they were the same1, but creating an environment which encourages people to forget those differences ever existed.

One of Scott’s first examples is land ownership in pre-modern rural villages, each of which had its own peculiar practices built over generations, describing the obligations between neighbors, the obligations between relatives, and the use of common pastures and forests. These practices were perfectly legible within a given village, but a cacophonous mess to state officials trying to understand all of them at once.

And so the states sent in surveyors, who drew sharp boundaries on their maps. A year later came the tax collectors, maps in hand. If a family paid taxes on what was once common property, they had little motivation to let their neighbors continue to use it. Instead, the villagers reshaped their lives to fit the maps.

In The Image of the City, Kevin Lynch defines “legibility” as something that a complex environment offers to its inhabitants, allowing them to easily navigate it. It is a clarifying abstraction, making the world more than an endless deluge of minute details.

Lynch interviewed long-time residents of various cities, and asked them to describe how they’d navigate from one part of the city to another. He created reference maps of shared elements from these journeys, representing how each city was understood by its people.

Some cities, such as Boston, were highly legible; its map was full of reference points, and its residents were confident in each step of their imagined journey. Other cities were less so:

[A]lmost anyone can, if attentive, learn to navigate in Jersey City, but only at the cost of some effort and uncertainty. Moreover, the positive values of legible surroundings are missing: the emotional satisfaction, the framework for communication or conceptual organization, the new depths that it may bring to everyday experience.

The moral hazard described in Seeing Like a State exists in any field of design. In software, for instance, anywhere one implementation cannot be easily exchanged for another, users are forced to bend themselves to the needs of their software. This is trivially true of enterprise software, since the users are not the buyers, but given the increasingly monopolistic bent of all software companies, it can be also true in the consumer space.

This is worrisome, not least because of the tech industry’s continuous attempts to resurrect logical positivism. Someone who claims their design was conceived from first principles has almost always injected their own biases and incuriosity into those principles. As coercive abstractions, these designs will preserve whatever is familiar to a small, homogeneous group, and allow everything else to wither away.

Scott offers prescriptions for avoiding the problems he describes, but they’re mostly process-oriented (“take small steps”, “favor reversibility”) and entirely focused on harm reduction. For Scott, abstraction is always something done to people, never for people.

But this is just one side of the coin; to apply the insights in either book we must understand both kinds of legibility, and what separates them. To get there, though, we must first explore two works of fiction.


The Gnostics were a collection of early Christian sects which fused the gospels with Platonic dualism.2 A true Creator was responsible for the world of ideal forms, while a lesser being, a demiurge, was responsible for the pale shadow of our physical world. For the Gnostics, the work of the demiurge was a veil over our eyes, created out of malice or ineptitude, and their goal was to pierce that veil, to achieve gnosis.

Philip K. Dick had a lifelong fascination with the Gnostics, and their ideas appear constantly throughout his novels. Nowhere is this more evident than in The Three Stigmata of Palmer Eldritch.

In the novel, the Earth has become uninhabitably hot, forcing everyone to take shelter in buildings or subterranean structures during the day. People are being forcibly migrated to even more hardscrabble lives on nearby planets, which can only be escaped via Perky Pat layouts, effectively Barbie Dream Houses they can temporarily inhabit using Can-D, a drug illicitly distributed by the creators of Perky Pat.

Using the drug, users are transported to a single, perfect day in San Francisco, before the Earth became too hot, in the bodies of a wealthy, beautiful couple. They can accessorize their escapism by purchasing “minned” items, which will appear full-size and functional within Perky Pat’s world.

This is a false gnosis, layering something simple and idealized atop the messiness of the physical world, rather than peeling it back. False gnosis is a central theme in all of Dick’s writing, which makes him perhaps the only science-fiction author who truly anticipated our present day. Like everyone else, he failed to predict the smartphone, but he alone seemed to understand how completely technology could intermediate our understanding of the world.3

Three Stigmata begins with the return of Palmer Eldritch, an industrialist who had travelled to Proxima Centauri in search of new business opportunities. He has brought with him a new drug, Chew-Z, which promises the same escape as Can-D without the need to buy any accessories.

Anyone taking Chew-Z can create their own world to inhabit, but every person in that world takes on aspects of Eldritch himself, sharing his prosthetic hand, eyes, and mouth. Even as the drug seems to subside, Eldritch remains, making it impossible to know what’s real.

In time, it’s revealed that the returned Eldritch is not the businessman, but rather a dying demiurge who has assumed his form. He toys with anyone who takes Chew-Z, creating escapist fantasies that prove to be every bit as flawed as their real lives. By the end, Eldritch has begun to merge with humanity, even infecting people who never touched the drug.

Eldritch doesn’t offer gnosis, but neither does he offer pure escapism like Perky Pat. He gives humanity an understanding of himself, of the nature of the veil over our eyes. Even if it tells us nothing of what lays beyond, Dick seems to suggest it’s the most we can hope for.4

A different sort of demiurge is portrayed in Italo Calvino’s Invisible Cities, in which Marco Polo tells Kublai Khan of all the cities he’s seen in his travels. They sit in the Khan’s palace, at the center of an empire which has exceeded his grasp:

It is the desperate moment when we discover that this empire, which had seemed to us the sum of all wonders, is an endless, formless ruin, that corruption’s gangrene has spread too far to be healed by our scepter, that the triumph over enemy sovereigns has made us the heirs of their long undoing. Only in Marco Polo’s accounts was Kublai Khan able to discern, through the walls and towers destined to crumble, the tracery of a pattern so subtle it could escape the termites’ gnawing.

His descriptions of the cities, however, are narrow. In one, he only describes a memory it evokes. In another, he only describes the relationship between the city and its reflection in the surrounding water. Sometimes the cities themselves are narrow; one such city has been reduced to freestanding pipes, with attached bathtubs and fountains, populated by nymphs and naiads.

These are pieces of cities, mixed and distilled from everything Polo has seen in his travels. But at their core, they are facets of a city he never describes, the city he knows best, his birthplace:

And Polo said: “Every time I describe a city I am saying something about Venice.”

“When I ask you about other cities, I want to hear about them. And about Venice, when I ask you about Venice.”

“To distinguish the other cities’ qualities, I must speak of a first city that remains implicit. For me it is Venice.”

Marco Polo is a worldly man, but he is not of the world; he is Venetian, and everything is understood and remembered through that lens. What he has discovered, in the course of his travels, is that certain pieces of Venice are timeless. These are what will survive the collapse of the Khan’s empire, what will constantly reoccur throughout time; always different in detail, but unmistakably the same.

And yet Polo, just like Eldritch, does not offer true gnosis. Each city hides all but a tiny piece of Venice through omission and fantastic obfuscations like the nymphs and naiads. Unlike Eldritch, however, he offers choice. The set of cities he describes is neither minimal nor exhaustive; each can be considered or ignored in isolation.

These cities may obscure their underlying reality, but they don’t attempt to replace it. Instead, they simply reveal a resonance, a commonality of experience we might otherwise have missed.


The legibility described in Seeing Like a State, and imposed by Palmer Eldritch, is singular; It overrides our own experience, forcing us to live within the mind of its creator. The legibility described in Image of the City, and offered by Marco Polo, is faceted; it complements our own experience, allowing us to apply it where we see fit.

Creating singular legibility is simple: find something that’s easy for you to understand, and force people to use it. What’s hard is using it responsibly; to predict its full effect, you’d need to first have an exhaustive understanding of the environments in which it will be used.

Faceted legibility is more forgiving, allowing us to understand the world incrementally and collaboratively. It’s not, however, something we can simply create on our own. Instead, we can only lay the groundwork and hope it flourishes.

How do we lay the proper groundwork? It depends on what you’re creating. If you’re interested in urban design, you might read Image of the City and follow it with Jane Jacob’s The Death and Life of Great American Cities, which expands on many of the same themes. If you’re interested in literature, you might read Calvino’s Invisible Cities and Mr. Palomar.

But if you, like me, are interested in software, there are no readymade references.5 I believe, however, that the answer lies in articulation points:

As we compose our abstractions, we must ensure we create, at regular intervals, interfaces which fully separate our upstream and downstream components. These interfaces not only delineate subgraphs which can be understood in isolation, they represent decision points in our design process. From any given articulation point, we can branch off in a dozen different directions, perhaps even all at once. These are separate interpretations of what the interface represents, each of which can be considered or ignored in isolation.

This is more easily said than done. Countless decisions go into every bit of software, but not all of them deserve to be enshrined as an interface. Once used, interfaces ossify. They allow the code on either side to drift apart, to lose consistency in idiom and purpose. If we don’t expect an interface to require multiple interpretations, it shouldn’t exist. To make the right decision, we must predict the future.

Even worse, we must ensure that our understanding of the physical world is also faceted. Any given application provides a singular view of the user. Any given dataset provides a singular view of whatever it represents. We must poke multiple holes in the boundary between our software and the world it reflects.

Sometimes this may be impossible, in which case we must heed Seeing Like a State, and adopt its methods for harm reduction wherever possible. I suspect, however, that in most cases we give up far too easily. Through habit and incuriosity, we treat understanding of our software as a proxy for understanding the world. Software design, as a discipline, is much broader than we like to think, almost impossibly so.

If we live in our own heads, we force everyone to live there with us. In the end, there may be nothing else left.


  1. The definition of abstraction as “treating things which are different as if they were the same” comes from Liskov’s Abstraction and Specification in Program Development, and it remains the only definition I’m aware of which encompasses all of the term’s common usages. 

  2. Fully half of the study of Gnosticism is debating what is and isn’t Gnostic, so while this characterization is broadly accurate, it elides a lot of nuance. Anyone looking for a more complete description should read Lewis’ Introduction to “Gnosticism”, scare quotes and all. 

  3. William Gibson’s cyberspace, conversely, was meant to be something akin to true gnosis; it was the world laid bare, rendered as Platonic solids, revealing the underlying forces that shaped it. This is echoed in the recurring ability of Gibson’s characters (Colin Laney in the Bridge trilogy, Cayce Pollard in Pattern Recognition) to perceive deep patterns in the world around them. 

  4. Ten years after writing The Three Stigmata of Palmer Eldritch, Dick would experience something like gnosis, and the stories that followed were more optimistic about our ability to truly understand the nature of the universe. 

  5. My own book, Elements of Clojure, explores some of these questions, but it’s far from a definitive reference.