paper · 19 October 2026
Reasons, Causes, and the Preemptive Society
On the replacement of moral agency with physical control.
§I
Two ways to stop a car
Consider two ways a city might keep drivers from speeding down a residential street.
In the first, the city posts a sign that reads 25 MPH. Nothing in the road prevents a driver from going 45. The asphalt is smooth and the lane is clear. For the sign to work, a sequence of mental events has to happen. The driver has to see the sign, understand what it asks, recognize the reasons behind it (children or neighbors might step into the street; living together takes some mutual restraint), and choose to ease off the accelerator. Even if the driver is partly moved by the chance of a ticket, slowing down is an act of self-government. The sign addresses the driver as a person who can respond to reasons.
In the second, the city pours a mound of asphalt across the lane. Taken at speed, it jolts the car hard, so drivers slow to 15 or 20 miles per hour to cross it comfortably. Here the driver's grasp of community norms is irrelevant. It makes no difference whether it is three in the afternoon as school lets out or three in the morning on an empty, well-lit street with no pedestrian within half a mile. The hump gives the driver no reason to slow down. It changes the shape of the road so that failing to slow down is punished on the spot. It addresses the car as a mass and the driver as an organism that avoids jolts and repair bills.
At first the shift from the first case to the second looks like a minor practical upgrade, from an unreliable method of speed control to a reliable one. American cities have been retrofitting residential streets with humps, tables, and cushions for years.
But the speed hump is only a small, visible example of a much wider change in how institutions relate to the public. The same shift shows up when pharmacies lock household goods behind shatterproof plastic, when software platforms replace community standards and human appeals with automated filters, when employers replace professional trust with keystroke and activity monitoring, and when parks install benches built so that nobody can lie down on them.
In each case an institution stops trying to govern behavior through rules and starts preventing it through constraints. Underneath is a distinction between governing people through reasons and controlling them through causes. A rule leaves an action possible and marks it as wrong. A barrier makes the action costly or impossible, and with it removes the occasion for choosing at all. This paper argues that the shift is neither trivial nor benign. When a society systematically replaces rules that appeal to judgment with mechanisms that bypass judgment, it changes what kind of beings it takes its citizens to be. And it sets off a dynamic that erodes the capacity for self-regulation it claims to be compensating for.
§II
The inside and outside views
Many people feel a flare of resentment crossing a run of speed humps on an empty street, or waiting for an employee with a key to unlock a three-dollar tube of toothpaste, even though they fully support safe streets and oppose shoplifting.
One could put this down to selfishness, the impatience of someone who wants to save thirty seconds whatever it costs others. That misses what is going on. The resentment comes from a collision between two ways of seeing human action: from the inside, as a deliberating agent, and from the outside, as part of a statistical system.
From the inside, each person is a center of judgment. Driving down a street, you can see whether the sidewalks are crowded or empty, whether visibility is clear or blocked, whether your situation is ordinary or urgent. When you respect a law from the inside, you weigh the reason behind it against the facts in front of you. Even when you obey a rule strictly with no one watching, the obedience is your own choice, an expression of your commitment to living on fair terms with others.
From the outside, the view of a traffic engineer, an insurance actuary, or a corporate risk manager, individual judgment disappears. Looking at a city from above, the planner does not see distinct agents making decisions in context. The planner sees ten thousand vehicles moving through a corridor, a distribution of speeds, and a probability of collision. From there, human discretion is not a moral faculty. It is a source of variance. Some fraction of drivers will be distracted, reckless, or indifferent, so the surest way to reduce aggregate harm is to take discretion out of the system.
A functioning society has to hold both views. If the inside view goes unchecked, people excuse their own selfishness by deciding their case is special, and public order breaks down. If the outside view goes unchecked, institutions treat every person as a hazard whose freedom has to be curtailed before they make any choice at all.
For much of the postwar era, American civic life tried, imperfectly, to bridge the two through shared norms. Institutions stated the rule from the outside and expected citizens, shaped by upbringing, community expectation, and occasional human enforcement, to take it on from the inside.
What has changed is that institutions are giving up on the bridge. Instead of asking how to help people bring public responsibilities into their own choices, institutional design removes the choice. You crawl over a hump at midnight on an empty road because the system has no way to register that you can see the road is empty. To the system you are a data point in a risk model. The insult is less the ten lost seconds than the discovery that the public environment has been rebuilt on the premise that your judgment can't be trusted.
§III
The utilitarian defense and the atrophy problem
The strongest objection to this critique is consequentialist. A defender of preemptive design would put it something like this:
Perhaps it is mildly insulting to careful citizens to have their judgment bypassed by a speed hump or a locked cabinet. But weigh the interests. On one side is a small daily annoyance and a bruised sense of autonomy. On the other is a child who is not hit by a speeding car, or a store that does not close because of theft. If physical preemption reliably prevents severe, irreversible harm at the cost of trivial, reversible inconvenience, isn't it morally required?
In any single case this argument is close to irresistible, and the evidence is on its side: speed humps do reduce child pedestrian injuries. 1Tester, J. M., Rutherford, G. W., Wald, Z., & Rutherford, M. W. (2004). A matched case-control study evaluating the effectiveness of speed humps in reducing child pedestrian injuries. American Journal of Public Health, 94(4), 646–650. Children living within a block of a hump had roughly half the odds of injury near home. If the only question is whether to put a barrier on one street this year, the reduction in risk will nearly always outweigh the abstract value of preserving driver discretion on that block.
The defense runs into trouble only when the time horizon widens and preemption becomes the default operating principle of a whole society. That wider view shows two traps.
1. The defection cascade
High-trust norms are fragile equilibria. Most people will give up some convenience for the common good, drive reasonably, wait their turn, follow civic rules, on the condition that they believe almost everyone else is doing the same and that open defectors will face consequences.
Both conditions have been weakening for decades. The share of Americans who say most people can be trusted fell from about 47% in 1972 to about 25% in 2022. 2General Social Survey, variable TRUST, NORC at the University of Chicago: 46.8% in 1972, 25.3% in 2022, as tabulated in secondary analyses of the cumulative file. Exact values move a point or two with weighting. The pandemic years then hit both conditions at once. Isolation weakened the face-to-face ties that make strangers feel accountable to one another, and formal enforcement pulled back in many cities through staffing shortages and political conflict. On emptier roads, a minority found they could defect without consequence. In 2020 Americans drove less and more of them died, with speeding-related deaths up 17%. 3NHTSA. (2022). Overview of motor vehicle traffic crashes in 2020 (DOT HS 813 266). Deaths rose to 38,824 as miles driven fell; the fatality rate per 100 million vehicle miles went from 1.11 to 1.34. Speeding-related deaths rose from 9,592 to 11,258. Retail theft became a national story as well, though the industry's headline numbers turned out to be overstated. 4Digital Commerce 360. (2023, December 4). NRF retracts statement on financial impact of organized retail crime. The National Retail Federation withdrew its claim that organized crime caused nearly half of $94.5 billion in losses. Shoplifting data are noisy; the Council on Criminal Justice’s year-end 2024 update put it about 1% above 2019 across 40 cities.
Once a visible minority defects without consequence, keeping the unwritten norm starts to feel foolish to everyone else. Facing that, and lacking the staff or political consensus to restore enforcement, municipalities and companies reached for the cheapest, fastest substitute available: physical and automated barriers.
2. The atrophy loop
The trouble is that preemption treats the symptom of defection while deepening its cause.
The capacity for self-government is not a fixed part of human nature that stays intact when it goes unused. It is a disposition kept up by practice. People learn to weigh others' interests, read context, and restrain their impulses by living in environments where they could act badly, are expected not to, and experience what it means to choose well.
Consider a society that systematically removes the possibility of choosing badly through barriers, locked doors, and rigid software. Where the barrier stands, compliance is complete, not because anyone is exercising restraint but because defection is blocked. What do people learn there? Plausibly, that whatever is physically possible is permitted, since anything truly off-limits would have been made impossible.
If so, the same people will have weaker internal brakes the moment they step into a space where no barrier has been built yet, and they will behave worse wherever friction is absent. The institution sees the worse behavior, concludes that people are even less trustworthy than it thought, and extends constraints into the remaining open spaces. Preemption confirms itself. It produces the kind of citizen whose behavior justifies more of it.
This loop is the paper's most important claim and its least proven one. There is evidence that the form of enforcement can change why people comply. When Israeli day-care centers began fining parents for late pickups, lateness rose, and it stayed up after the fine was dropped. 5Gneezy, U., & Rustichini, A. (2000). A fine is a price. Journal of Legal Studies, 29(1), 1–17. A later replication found the effect was temporary (International Review of Law and Economics, 2020). But no study I know of shows that living among barriers makes people behave worse where barriers are absent. The loop is offered as a hypothesis worth taking seriously, not as a finding.
§IV
Two models of a low-trust future
A society that loses the shared fabric that makes voluntary norm-following work still needs basic order. Two models are visible today. Call them order by total visibility and order by universal friction.
The first is most associated with the contemporary Chinese state. There the constraint moves off the pavement and into digital architecture: dense camera networks, facial recognition, linked payment and identity systems, and administrative blacklists that let a defection be attributed to a specific person and penalized quickly. In this model the citizen keeps freedom of movement because detection is close to certain. The price is privacy and political autonomy, since the apparatus that catches a reckless driver can also make dissent impossible.
This is an idealized type, not a portrait of a country. China's “social credit system” is a patchwork of blacklists and local schemes rather than one national score, and Chinese streets have their share of speed bumps. 6Creemers, R. (2018). China’s social credit system: An evolving practice of control. SSRN 3175792. Automated camera enforcement of traffic rules, on the other hand, is dense and well documented in Chinese cities. The model is useful for what it shows about the options.
The United States is drifting toward the second model. Constitutional design, cultural temperament, and institutional fragmentation make a unified state surveillance system hard to build here. Meanwhile American society has spent half a century drawing down the interpersonal trust, local associations, and shared expectations that once let a decentralized republic function without heavy-handed control.
Unable to watch everyone, and unable or unwilling to rebuild the conditions of trust, American institutions fall back on degrading the physical and procedural environment for everyone alike. Because the city can't reliably identify and penalize the few drivers doing 50, it pours mounds that make everyone else crawl at 12. Because a retailer can't stop a small fraction of shoplifters, it treats every customer as a possible thief. Because a bureaucracy can't trust an employee's judgment, it routes every worker through a maze of compliance checkboxes.
Where the first model buys smooth surfaces with submission to an all-seeing state, the American drift keeps formal political liberty while encasing everyday life in mechanical hindrance and suspicion. Surveillance arrives anyway, piecemeal, through private cameras, plate readers, and data brokers. The likely destination is less a dramatic tyranny than a slow civic sclerosis, where little works smoothly because no system will let a person make a decision.
§V
Why the trajectory persists, and how to reverse it
If universal friction is alienating and self-defeating in the long run, why do institutions keep building it? A call to restore civic trust is useless unless it deals with the incentives that make mechanical control so attractive to the people making decisions.
1. The structural blockers
First, institutional risk is asymmetric. Imagine a city official deciding whether to put speed tables along a whole corridor, or an executive deciding whether to lock down a workflow. If the official declines and a single bad driver causes a serious crash six months later, the official faces public outrage, career damage, and liability. That harm is concentrated, visible, and traceable to a decision. If the official installs fifty speed tables, the costs are spread thin: thousands of hours of lost time, slower emergency response, 7Portland’s measurements found fire engines losing up to about 9 seconds per hump or table: City of Portland. Offset speed tables for reduced emergency response delay. wear on thousands of cars, and a steady erosion of civic trust across tens of thousands of residents. Nobody can sue a city over the slow loss of moral agency. Institutions will keep choosing a large, invisible, dispersed cost over a small, visible, concentrated risk.
Second, there is the political economy of procurement. Nobody can award a contract for “neighborhood mutual accountability.” Contracts go to civil engineering, license-plate cameras, security hardware, and compliance software. In a political system heavily shaped by private money and vendor lobbying, problems of social cohesion get reframed as technical problems that call for purchased hardware. Friction is profitable for the people who sell the barriers.
Third, shared local life has thinned. Rules work when people feel that the strangers around them belong to the same moral community. When citizens consume polarizing national media and barely know the people three doors down, the street stops feeling like a shared commons and starts to feel like a contested transit zone between strangers.
2. Reconstructing a society of agents
Nobody can go back to 1955, and nobody should idealize an era whose internal cohesion often rested on conformity and exclusion. But a free society can't last indefinitely by treating its citizens as objects to be deflected from one another. Reversing the trend takes deliberate choices on three fronts.
- Targeted accountability over indiscriminate friction. The core defect of physical preemption is that it punishes the careful majority to control the reckless minority. Restoring agency means enforcing rules through human judgment aimed at actual violators: leave the road smooth for the person driving safely, and stop the one doing 55. That enforcement has to be accountable, because human discretion has its own record of abuse; the largest study of US traffic stops found Black drivers stopped and searched on weaker evidence than white drivers. 8Pierson, E., Simoiu, C., Overgoor, J., Corbett-Davies, S., Jenson, D., Shoemaker, A., Ramachandran, V., Barghouty, P., Phillips, C., Shroff, R., & Goel, S. (2020). A large-scale analysis of racial disparities in police stops across the United States. Nature Human Behaviour, 4(7), 736–745. Paying for accountable human judgment is part of the price of living among free agents.
- Rebalancing liability. The legal and administrative structures that punish institutions for treating adults like adults need reform. Design standards and tort rules should weigh the cumulative public cost of friction and lost autonomy against marginal reductions in risk, and should protect communities that choose open, trust-based environments over defensive ones.
- Returning design power to the local scale. City-wide scoring rubrics apply uniform formulas to very different neighborhoods. When residents of a block or neighborhood have real authority, and have to sit across a table from one another to decide how their shared streets and spaces work, rules stop looking like impositions and start working again as agreements. Local control isn't enough by itself, since a meeting that only hears about the worst case will vote for the barrier. It works when the dispersed costs are on the table too.
§VI
Conclusion
A speed hump is a small thing, a few tons of asphalt laid across a strip of road. Yet an archaeologist a thousand years from now, looking for the moment late-modern democratic society lost faith in itself, could do worse than excavate our streets.
They would find a civilization of great technical ability that could no longer persuade its members to slow down for a neighbor's child, and so broke the smoothness of its own roads instead. The question in front of us is whether we are content to live in a padded room, where harm is prevented by making sudden movement impossible, or whether we will do the harder work of rebuilding a society whose members can be trusted to steer.
§VII
Rigor layer
What the argument rests on, what it doesn't claim, and what would count against it.
How well each claim is supported
| Claim | Status |
|---|---|
| Physical barriers reduce harm case by case | Supported for speed humps (Tester et al., 2004). Granted in §III. |
| Interpersonal trust in the US has fallen for decades | Supported (GSS). The decline long predates 2020. |
| Visible defection rose in 2020–2021 | Supported for road deaths and speeding (NHTSA). Mixed for retail theft, where industry figures were overstated. |
| Institutions increasingly replace rules with constraints | Observational. The examples are real and familiar; I know of no systematic measure of the overall trend. |
| Preemption weakens self-restraint and feeds on itself | Hypothesis. Indirect support from studies of how enforcement changes motives; no direct study. Labeled as such in §III. |
What this paper doesn't claim
It doesn't claim any particular barrier is a mistake. The argument is about preemption as a default, not about any one street or store.
It doesn't claim the past was better across the board. Mid-century cohesion was high partly because many people were kept out of it.
It doesn't offer a portrait of China or the United States. Visibility and friction are idealized types, and real places mix them.
What would count against it
If people who live among heavy preemption turned out to behave as well as, or better than, similar people elsewhere when the barriers are absent, the atrophy loop in §III would be wrong, and the paper would shrink to an argument about dignity and inconvenience.
If careful accounting of dispersed costs kept showing that the barriers pay for themselves, the claim in §V that institutions overbuild them would be wrong too.