In the Milgram post, I argued that what he uncovered was not an innate human instinct, but the end-state of a long cultural installation. Obedience, as we encounter it today, is not a raw reflex wired into our species. It is architecture: hierarchy, bureaucracy, institutional conditioning, all layered over generations until there is no control group left.
In the obedience-architecture piece, I pushed that further: once obedience is built into the scaffolding of schools, offices, armies, and states, you don’t need explicit commands. The structure itself issues them. People simply grow up inside an environment where deferring to distant authority feels like “how the world works.”
Once you see obedience as installed rather than innate, a harder question appears:
If obedience was added, what did it replace—and what did it reshape?
It didn’t just cover over aggression or selfishness. It also displaced a different default:
- distributed authority
- coalition-based checks on power
- voluntary association and exit
- refusal of coercion
- the freedom to challenge
- the freedom to leave
For most of human history, those weren’t fringe behaviors. They were the baseline template of social life.
Which leads to the uncomfortable next step:
If obedience is a cultural construction, then our modern concept of autonomy is one as well.
Not autonomy as our ancestors lived it, but a domesticated variant that evolved inside the very systems that constrain it.
The question stops being, “Why do people obey?” and becomes:
What does autonomy mean when every available concept of freedom has been shaped inside obedience-based architecture?
Autonomy Inside an Obedience Framework
Modern societies talk about autonomy as if it were a universal human constant: a timeless property of the individual mind. You either “have autonomy” or you don’t.
But the version of autonomy we inherit today didn’t emerge in a vacuum. It was born inside:
- territorial states and citizenship regimes
- legal personhood and contractual systems
- markets and wage labor
- mass schooling and bureaucratic adulthood
- moral philosophies of duty and responsibility
- religious traditions of submission and discipline
Autonomy, in this context, is not a return to some original human freedom. It is a negotiated permission structure—freedom that is allowed, recognized, and recorded by the very institutions that limit it.
You are autonomous as long as:
- your choices fit within legal codes,
- your survival is routed through markets and employers,
- your dissent stays inside predefined channels,
- your refusal does not disrupt core functions of the system.
That is not a glitch. It is the design.
Derivative, not original
For forager societies, the things we now theorize as “autonomy” did not need a name. They were simply the way life worked:
- leadership was fluid and situational
- association was voluntary
- mobility and exit were real options, not metaphors
- resource sharing blunted extreme dependency
- coalitions could form to check would-be dominants
- coercion carried high social costs
You didn’t need a philosophy seminar to argue that people had a right to leave. You just…left.
Real constraints existed—ecology, conflict, scarcity—but the baseline was self-direction within a small band, not obedience to distant institutions.
Once sedentary states appeared, that baseline changed. Now there were walls, taxes, armies, landlords, priests, and later, managers and HR. Under those conditions, “autonomy” had to be reintroduced as a concept precisely because the lived version had already been curtailed.
Philosophers, theologians, and jurists spent centuries constructing the sovereign individual—but always inside:
- kingdoms and empires,
- churches and bureaucracies,
- markets and property regimes.
The individual they defended was never an unbounded human being within an egalitarian band. It was a subject, citizen, or employee whose freedom was being negotiated with higher structures.
In that sense, modern autonomy is derivative:
- a reaction to obedience-based societies,
- defined in opposition to them,
- constrained by the very architectures it tries to resist.
Autonomy defined by its opposite
We measure autonomy using terms like:
- freedom of speech
- freedom of conscience
- freedom of movement
- economic independence
- privacy and property
- rights and protections
These are real achievements. Losing them makes life worse.
But notice what they imply. Each of these freedoms is:
- granted or at least recognized by a higher authority,
- recorded somewhere—in law, policy, databases,
- enforceable only through the same machinery that limits you.
This is permissioned autonomy:
freedom that exists as a corridor carved into a larger structure you don’t control.
You are free to move, until borders close. You are free to speak, until speech is reclassified as something actionable. You are free to earn, until the platforms you rely on change their terms.
The situation isn’t binary obedience vs pure autonomy. It is conditional autonomy inside an obedience-shaped reality.
The Control Group We Don’t Have
In the Milgram experiments, there was no true control group—no adults raised entirely outside obedience architecture. Everyone in the lab had already spent years inside schools, workplaces, militaries, churches, or their cultural equivalents. They arrived pre-configured.
From that perspective, Milgram wasn’t testing “obedience vs autonomy.” He was testing:
obedience vs the narrow range of culturally permitted behaviors that we’ve agreed to call autonomy.
The participants believed they were autonomous moral agents:
- they could consent or withdraw,
- they could question the setup,
- they could refuse.
In practice, their internal maps of freedom had already been formatted by institutions. The experiment only nudged them along channels that were built long before they entered the lab.
Milgram did not reveal some timeless human flaw. He revealed the limits of an autonomy that had been domesticated to fit obedience-based structures.
The real control group—the humans whose autonomy evolved without states, markets, and bureaucracies—was already gone.
The autonomy we can’t remember
If humans spent roughly 95% of our evolutionary time in small, relatively non-hierarchical bands, then our brains were tuned for a very different social world than the one we inhabit now.
Under that ancestral regime, “autonomy” would have meant something closer to:
- choosing which band to associate with,
- walking away from intolerable situations,
- reshaping norms face-to-face,
- sharing risk and resource in ways that limited extreme dependency,
- forming coalitions to block would-be tyrants early.
We don’t have direct access to that mental model. What we do have are:
- mythic stories of golden ages and lost commons,
- ethnographic snapshots of societies that resisted full incorporation,
- developmental windows (like adolescence) where refusal, testing, and boundary-pushing resurface despite institutional pressure.
It is possible that what we call “teen rebellion” is not a developmental bug, but one of the last remaining channels where an older autonomy template still tries to assert itself before being fully folded into obedience architecture.
If so, our theories of autonomy are being built after the fact, over a landscape that has already been terraformed.
Why This Matters Beyond Milgram
If our concept of autonomy is itself a product of obedience-based systems, several things follow.
-
Experiments like Milgram are not testing free agents. They are testing humans whose choice architecture has already been narrowed and labeled “normal adulthood.”
-
Political and moral theories overestimate what autonomy can do. They imagine individuals as independent evaluators of systems, while those individuals are in fact produced by those systems.
-
Our sense of responsibility is entangled with our domestication. We judge people for failing to resist, while the tools that would make resistance thinkable were never allowed to develop.
Seen through this lens, autonomy is not a clean counterweight to obedience. It is a managed opposition—a set of freedoms calibrated so that the system can function, adapt, and claim legitimacy, without ever giving up the capacity to command.
That doesn’t make autonomy meaningless. It makes it partial.
What Would Untamed Autonomy Look Like?
If the only autonomy we know is the version compatible with states, markets, and bureaucracies, then we are like people trying to imagine wild ecosystems from inside a manicured park.
You can walk, sit, even go off the main path a bit. But the layout, the fences, the opening hours—all of that was decided without you.
So the hardest questions become:
- What would human autonomy look like if it had been allowed to evolve outside obedience-based systems?
- What kinds of relationships, institutions, or non-institutions would it produce?
- How would responsibility, care, and coordination work if they weren’t routed through obedience?
- Can we design anything today that approximates that space, without pretending we can roll history back?
I don’t have final answers. What I do have is a working suspicion:
Our current language of freedom is built from the vocabulary of obedience.
If that’s true, then part of the work ahead is not just resisting specific commands or institutions, but slowly recovering a different conceptual toolkit—one that can talk about:
- refusal without treating it as pathology,
- exit without framing it as betrayal,
- shared power without assuming a central command,
- responsibility without defaulting to hierarchy.
Milgram’s lab shock machine is long gone. The obedience architecture it revealed is not. Any serious conversation about autonomy has to start there—not with the myth of the fully independent individual, but with the reality of humans whose understanding of freedom has been shaped, from the beginning, inside systems that expected obedience.