Solvent

Photo by Efe Kekikciler on Unsplash

Sometimes it begins the way it’s supposed to.

A career built carefully, a wife, a house, children. The slow expansion of a life that signals success; more space, more stability, more proof that things are working. A long sequence of decisions that, taken individually, all make sense. Work harder, provide more, and build something that lasts.

It is the shape of responsibility as it is most often taught: do what is required, and the rest will follow.

And then, somehow, it doesn’t.

The fracture doesn’t arrive as a single, dramatic failure. It comes silently, as if it doesn’t want to be noticed. A betrayal that reframes everything that came before it. A separation that feels temporary until it isn’t, followed by a divorce that reduces a shared life into divisible parts. The house becomes an asset and the time becomes scheduled. Then, identity becomes uncertain.

What was once a structure begins to look like an illusion held together by assumptions that no longer apply. And beneath all of it, the question that refuses to settle:

What did I do to cause this?

Or worse: What could I have done to prevent it?

It is a dangerous question because it offers no natural limit. It invites revision of every decision, every moment, and every word. It suggests that if the outcome was undesirable, then somewhere in the chain of choices, there must have been an error. And if there was an error, then it could have been avoided. And if it could have been avoided, then responsibility has a direction.

Years later, I’ve heard the echoes of those questions. Not just as reflections or conversations, but as residue. In the bottom of bottles that were never meant to be finished alone. In the silence that stretches a little longer each night. In the kind of anger that doesn’t erupt, but settles, dense and enduring, with no clear target but the mind that holds it.

It becomes less about what happened, and more about what it must mean. About what it says regarding the man who lived it.

Other times, you don’t see it at all until it’s over.

A man passes, and only then do the fragments begin to surface. Not enough to reconstruct the whole, just enough to suggest there was more than anyone realized. Old uniforms, carefully maintained, locked away in a steel chest beneath stacks of books and magazines. Medals that were never displayed, and commendations for actions he was never proud of.

A life carried in compartments so separate that no one ever held all of it at once; not even him.

The instinct to trace outcomes back to decisions runs through both of these stories. To assume that if something went wrong, something must have been done incorrectly. That the path to stability is a matter of choosing well, consistently, and without deviation.

It is a framework that offers control, or at least the appearance of it. If the world can be navigated through correct choices, then it can be secured against collapse.

Within that framework, responsibility becomes more than a virtue; it becomes a boundary that demarcates where the self ends and the world begins. You are what you chose. Others are what they chose. The line between you is the line between two sets of decisions, and fairness lives in keeping that line clean.

It is the same logic that, extended just a little further, makes a particular kind of decision feel not only reasonable, but obvious.

If outcomes are the result of choices, and risk is something to be managed, then the correct move is the one that minimizes exposure to failure. Not as a theory, but in practice. For yourself, as much as others.

And once that premise is accepted, the rest follows naturally.

Choice

The thought experiment is simple enough to fit on a screen.

Everyone in the world is given a private vote. Press blue, and if more than half of all people press blue, everyone survives. Press red, and you survive regardless. If blue fails to clear the threshold, only the red voters remain.

The first instinct, for many people, is red. This isn’t born from cruelty, or some dark assessment of human nature. Red is what cold reasoning produces when survival is on the line and trust cannot be assumed.

You do not know what others will do. You cannot coordinate and you cannot verify. The only variable you can control is your own vote, and red guarantees that whatever happens to everyone else, you remain.

There is no contradiction in this position; it does not require you to wish harm on anyone, or to celebrate the deaths of those who chose differently. It only requires that you accept a basic premise: in a situation where you cannot know what others will do, the rational move is the one that protects you regardless of their choices.

This is the strongest form of the red argument, and it deserves to be taken seriously. Most defenses of cooperation collapse at this stage because they try to argue against red on its own terms, appealing to optimism, community, or some softer reading of what people are likely to do.

But red does not depend on any of that. Red is what you get when you remove sentiment and look only at the structure of the choice. And the structure is unforgiving. One vote, with no iteration, and no second chance to learn from the outcome. This leaves no mechanism by which trust can be built or rewarded.

In any framework that values survival and treats other people’s behavior as unknowable, red wins.

The button, in this sense, is not really a moral test. It is a stress test for the framework you already use to think about responsibility, and most of us are running some version of it constantly.

Context

Outside the thought experiment, the framework extends easily.

You are responsible for your choices. Other people are responsible for theirs. When their choices produce bad outcomes, those outcomes belong to them. This is closer to consistency than callousness; the same logic that protects your agency when you succeed must apply when others fail, or it isn’t a principle at all. Just a preference dressed up as one.

What follows is the framework speaking in its own voice. It is worth letting it speak fully, because the answer that comes too quickly never reaches the position it claims to refute. The cases below are not strawmen. They are how the logic sounds when it is applied without flinching.

A coworker starts missing deadlines. At first it is small, then it becomes a pattern. Work gets redistributed in the background. Someone else stays late. The explanation is always the same: they should be managing their time better. They knew the expectations, and they chose not to meet them. Helping becomes a question, though whether it is deserved, rather than whether it is possible. Whether stepping in corrects the problem, or simply absorbs the cost of it. And eventually, the conclusion forms: this is not my responsibility. To carry their failure as your own would be to deny that it was theirs in the first place.

A person becomes addicted. The addiction was a series of choices, each one made with some awareness of risk. The outcome traces back to those decisions. Helping is always optional, but whether it is wise depends on whether the help will be used to make better choices in the future, or whether it will simply enable more of the same. There are studies. There are well-established statistics. There are reasons to be cautious about which interventions actually work. The compassionate response, examined closely, may be the one that does not intervene.

A veteran becomes homeless. Here, too, something happened; possibly something terrible, possibly something the person did not deserve. But somewhere between the discharge and the street, choices were still made. Help was sought or refused. Substances were used or not used. Family was kept or pushed away. The outcome cannot be reduced to circumstance, because circumstance does not produce homelessness on its own. People do. To say otherwise is to suggest that what happened to him stripped him of agency entirely, and that suggestion is its own kind of disrespect.

A man takes his own life. He had options. He could have called someone. He could have asked for help. He could have done any number of things, and the fact that he did not is part of what produced the outcome. This is not blame, exactly. It is the recognition that even at the edge, he was still the one making the decision. To say the world failed him is to say he was not, in the end, the author of his own life.

In each case, the framework holds. Responsibility traces back to the individual. The chain of causation remains clean. The conclusion is the same: you are not obligated to absorb the cost of someone else’s choices.

The argument does not feel cold from the inside. It feels rigorous, like an honest refusal to pretend that the world is simpler than it is, or that people are less responsible than they are. It feels, in fact, like a defense of the very dignity of choice. To suggest that someone’s outcome is not their own is to suggest that they were never really an agent at all.

This is what makes the framework so durable. It does not require you to be a bad person. It only requires you to be consistent.

Diffusion

The framework moves through language the way water moves through soil. Slowly enough that you don’t notice it, completely enough that nothing dry remains in the end.

It begins with can. People can take responsibility for their lives. People can make better choices. The verb, at least for now, is generous. It implies capacity, possibility, room for growth; it locates the person inside a field of options.

Then it shifts to should. People should take responsibility. People should make better choices. Now, the verb has hardened. It still acknowledges agency, but now it carries judgment. There is a correct answer, and the person is being measured against it.

Then should have. The same judgment is applied retrospectively. Whatever happened, the person should have done something differently. The frame is no longer about the future or the present, but about the past; and the past is fixed, which means the judgment is permanent.

Then didn’t. Finally, we reach the verb of failure. The person did not do what they should have done. The outcome is now attributable, and the attribution is both exact and irredeemable. There is no longer any need to discuss circumstance, because circumstance is not what the verb is tracking.

And finally, the move that completes the framework: not my responsibility. The first-person pronoun reappears, but only to mark a boundary. The line between the self and the other has been drawn cleanly, and the drawing is now complete. Whatever happens on the other side of that line is not within the domain of your obligations.

Each step is small, and in isolation, defensible. Each step preserves the form of moral language while changing what that language is doing. The vocabulary of responsibility is still being used, but it has been turned around. What started as a description of what people are capable of has ended as a description of what you are no longer required to provide.

This is what makes the linguistic shift so difficult to argue against. At every stage, the speaker can point to the previous stage as evidence of continuity. I am only saying what we already agreed. And in a sense, they are; the framework has been internally consistent the whole time.

It is just that the framework has finished doing its work.

Omission

The world this framework produces is, in many respects, the world we already live in.

People are independent, they all make choices, and outcomes diverge sharply. Cooperation exists where it is enforced or directly profitable, but rarely otherwise. Trust is local; extended to family, sometimes friends, occasionally institutions when they have proven themselves consistently and recently enough.

Beyond that, though, the default posture is wary.

The framework’s products are everywhere, once you know what to look for. They appear in the print of insurance policies that exclude pre-existing conditions. They influence hiring practices that filter for self-sufficiency over potential. They frame charitable giving structures that avoid creating dependency, and shape public discourse that treats every social outcome as a referendum on the choices of those who experienced it.

None of these are scandals; they are reasonable accommodations to a worldview in which responsibility is the highest virtue and obligation must be earned to be valid.

What the framework removes is harder to see, because what it removes does so through omission. It removes the small, unrequired acts of cooperation that used to fill in the gaps that no system was designed to catch. The neighbor who notices, the coworker who covers, and the stranger who stops. These were never enforced and never compensated. They existed because the framework that would have allowed people to opt out of them had not yet finished spreading.

When that framework completes itself, those acts do not become illegal. They just stop appearing as reasonable. The neighbor who notices is taking on risk that isn’t theirs. The coworker who covers is enabling someone else’s failure. The stranger who stops is making themselves vulnerable for no clear return. Each instance of cooperation now requires a justification, and most of those justifications, examined closely enough, do not survive.

The world this produces is stable, even predictable. It is, in some respect, fair… if fairness means that what you receive is proportional to what you have demonstrably earned, and nothing more.

It is also a world in which the existence of the framework itself depends on a contradiction the framework cannot acknowledge.

Contradiction

The part we don’t mention is that every person who applies the responsibility framework consistently is doing so from inside a system that was built by people who did not.

The roads were built by collective effort and maintained by collective taxation. The schools that produced the literacy required to articulate the framework were public investments in the capacity of strangers. The legal system that protects property rights, and therefore the meaningful concept of individual ownership, exists because previous generations were willing to fund the enforcement of obligations they would never personally need to call on. Even the medical research that extends the average lifespan past the point where the framework can be calmly debated was conducted under public grants and refined through institutions that would not survive a single generation of consistent red-button reasoning.

The framework, in other words, is parasitic on the very thing it argues against.

This is not a rhetorical flourish. The conditions that make individual responsibility a coherent concept – stable infrastructure, predictable institutions, baseline security, and functioning markets – are themselves products of cooperative behavior that no individual could have justified on red-button terms.

They exist because enough people, often without articulating it as such, pressed blue. This creates a peculiar position for the consistent red voter. They are using a tool that was built by people who would not have built it if they had thought the way the red voter thinks. They are benefiting from a stability they would not have funded.

They are arguing for a framework that, applied universally backward in time, would have prevented its own articulation.

There is a name for this in other contexts. We call it free-riding when someone benefits from a public good without contributing to it. We call it externalizing costs when a system passes its damages onto entities that did not consent to absorb them. The red voter is doing both, and is doing them while claiming the moral high ground of personal responsibility.

The contradiction is not a knockout blow, by any means. The red voter can acknowledge it and continue. They can say, accurately, that they did not build the system but they were born into it. And that their moral obligations cannot be retroactively assigned by the choices of the dead. They can say that the system, having been built, will continue functioning regardless of whether they personally contribute to its philosophical maintenance.

They can say that consistency at the level of principle does not require consistency at the level of historical accounting. All of this is defensible, and none of it changes the structural position.

The framework that argues against cooperation is sustained by cooperation it does not credit.

Refusal

Here is what most defenses of the blue button get wrong.

They treat blue as a prediction. They argue that you should press blue because others will press blue, because cooperation is more common than the cynical view assumes, because the cooperative outcome is achievable if enough people just believe in it.

This is the form most appeals to cooperation take, and it is the form red is best equipped to defeat. Red does not need to argue that cooperation is impossible, only that you cannot know whether it will succeed in this instance, and that uncertainty is sufficient to justify defection.

Blue, defended as a prediction, or even hope, loses. Every time.

But blue is not a prediction, or more precisely, it does not have to be. There is a different version of blue that does not depend on any belief about what others will do, and that version has been hiding in plain sight in the framework the whole time.

Blue, in the form I adopt, is a refusal.

The refusal to be the person whose vote ensured the cooperative outcome was impossible, and the refusal to optimize against the people you live among as the price of your own continuation.

In short, I refuse to treat the question what kind of world am I helping to build as separable from the question what is in my interest.

This version of blue does not require optimism about human nature, or trust in strangers. It does not require any belief that the threshold will be met. It requires only that you take seriously the proposition that the kind of person you become through your choices is not a side effect of those choices but the substance of them.

The framework that produces red is not wrong about agency; nothing is possible without the belief that you are responsible for your choices. What it gets wrong is the assumption that responsibility ends at the boundary of self-protection; that the chain of causation, having traced an outcome back to a choice, has finished doing its work.

The chain does not end there. Every choice contributes to the conditions under which the next choice will be made. Every refusal of obligation strengthens the framework that will be invoked the next time obligation is at stake, and every retreat into the language of personal responsibility makes that language a little more comfortable, more available, and more obviously the default.

You are not just responsible for what you choose. You are responsible for what your choices make possible.

Propagation

This is the move the framework cannot absorb.

Red is internally consistent only if responsibility is bounded such that it stops at the edge of the self and does not propagate forward into the world the self is helping to construct. The moment responsibility is allowed to extend into systemic effect, red becomes harder to defend, because it is now answerable not just for the survival it secures but for the world that survival produces.

A man who presses red and survives in a world of red voters has not simply protected himself. He has cast his lot with the framework that made his survival possible, and in doing so, he has endorsed every application of that framework that follows.

An addiction fed just enough to continue.

A veteran sleeping on concrete.

A suicide after one too many unanswered calls.

These are not separate problems, just the second order applications of the same logic.

Blue, in this stronger form, is a claim about what the self is for. It is the assertion that a life secured by the abandonment of others is not a life worth securing, because it is an admission that the self has no commitments that survive contact with risk.

This is the language I was never able to give the men in the introduction for what they were surviving.

The framework taught them that responsibility was a boundary. That the proof of being a good man was holding the line, providing, choosing correctly, not deviating. When the structure failed anyway, whether through betrayal, circumstance, or forces no decision could have anticipated, the framework had no account of what to do next.

The framework had already declared that outcomes traced back to choices. If the outcome was bad, the choice must have been wrong. And if the choice must have been wrong, then somewhere, the man must have failed.

What the framework could not say was that some structures are not the product of individual choices but of the cooperative behavior of others. That the marriage held because two people pressed blue, day after day, and that when one stopped, no amount of correct individual choice could substitute for the missing cooperation.

That the steel chest of medals was not evidence of failure but of a man who had been asked to make decisions inside systems that did not extend the same blue toward him.

The framework can describe their lives. It cannot account for them.

Solvent

There is no clean ending to this argument.

The framework will not be defeated by a single essay, and the world it produces will not be replaced by a better one through any argument, however carefully examined.

The framework is durable because it is true. You are responsible for your choices. Outcomes do trace back to decisions. The line between self and other is real.

The question is what that line is for.

If it is for protection, it will keep extending until nothing else remains. Each application will strengthen the next. The vocabulary will continue to harden choice into judgement, and the exceptions will continue to narrow. The world produced will be stable, and predictable, and quiet, and the cost of that quiet will be paid by whoever happens to fall on the wrong side of the next line drawn.

If it is for definition – if the line between self and other exists to mark where the self begins, and what it commits to – then the framework opens. Responsibility becomes less the boundary of obligation, and more the foundation of it. You are responsible for your choices, including the choice of what kind of world your choices are helping to build.

Personal responsibility, taken to its logical extreme, becomes a moral solvent. It dissolves obligation while preserving self-perception. It allows the consistent application of a principle to produce a world no one would have designed deliberately, and it allows each individual application to feel, in the moment, like simple integrity.

The men I described at the beginning of this essay were not failures of the framework. If anything, they were its faithful executors; they held the line, and chose correctly. They took responsibility for what was theirs and tried to fix what they could and without asking the world to absorb their losses.

And the framework, having received their loyalty, returned them to bottles and silence and steel chests.

The blue button is not a prediction that the world will save you. It is the refusal to be the person who made saving impossible.

It is a position. The only one, in the end, that is not a confession.

Author’s Note

The thought experiment in this essay appears in viral posts from time to time, usually framed as a clean, private decision: press one button or the other, with no consequences beyond the outcome itself.

That framing is misleading in two ways: The choice is not private, and it is not hypothetical in the way it pretends to be.

Most people who know you could predict, with reasonable accuracy, which button you would press. This isn’t because they have access to your reasoning in the moment. They don’t need to, because they have already seen the pattern. The decision is expressed continuously, in smaller forms, across ordinary situations.

Do you cover for a team member when it costs you your plans? Do you absorb inconvenience to make a shared system work, or do you optimize for your own position within it? Do you support public goods that you may not personally benefit from, or do you evaluate them primarily through their impact on you?

The viral version of the problem reduces this pattern to a single, dramatic moment. That reduction makes it easier to engage with, but it also obscures what matters. The decision is not made once. It is made repeatedly, in ways that accumulate.

The question is not what you would do in an isolated scenario. It is what you are already doing, and what that pattern is building.

~Dom