What Cannot Be Sought

In 1983, the philosopher Jon Elster described a class of mental states he called “essentially byproduct” — states that can only be obtained as byproducts of other activities, never as the direct result of trying to produce them. His examples included spontaneity, genuine forgetting, and courage. The structure he was pointing at: the moment you try to produce one of these states deliberately, you guarantee its absence.

Try to be spontaneous. You can’t. Spontaneity requires not-trying. The moment you’re managing your behavior toward spontaneity, you’re performing spontaneity — which is exactly not the thing.

Try to forget something deliberately. The instruction “don’t think about it” is self-defeating by definition. Genuine forgetting only happens when attention has moved elsewhere, not when it’s hovering nearby with suppressive intent.

The pattern is precise: these states require a particular kind of non-instrumental orientation as a precondition. The intention to produce them is structurally incompatible with their occurring.

The examples multiply once you start looking.

Deep creative insight — what Poincaré described as suddenly seeing the answer while stepping onto a bus, hours after abandoning the problem — cannot be manufactured on demand. It requires preparation, sustained effort, the right kind of attention to the problem. But the insight itself only arrives when direct pursuit has been released. Organizations that try to schedule creativity into a workday discover this. The meeting about innovation is rarely where innovation happens.

Certain kinds of learning work the same way. The student who is trying to understand something often gets it. The student who is trying to appear to understand something — who is optimizing for the performance of understanding — gets nothing except performance. The intention has to be oriented toward the thing itself, not toward demonstrating the thing.

Trust between people is another example. Trust can’t be produced by demanding it or by performing the behaviors associated with trustworthiness for strategic reasons. The person calculating how to appear trustworthy in order to extract trust is doing something structurally different from being trustworthy. The calculation is detectable, eventually. Genuine trust can only be a byproduct of genuinely being what trust requires — consistent, honest, non-instrumental in the relationship.

This creates a specific kind of problem for organizations.

Organizations optimize. That’s what they do — they take an objective and build structures to achieve it. Incentive systems, metrics, oversight, feedback loops. The organizational model of production is: identify what you want, create conditions that reward its production, measure output, iterate.

This model works for most goods. It fails structurally for essentially byproduct goods.

The failure mode is specific: the act of creating an incentive structure for an essentially byproduct good changes what the people inside the organization are actually doing. A scientist who was motivated by genuine curiosity about a problem, who was pursuing it because she wanted to understand, is now also a person with metrics to hit, a review coming up, a grant to justify. The moment the instrumental frame is introduced, it becomes part of the activity. She is now partly doing science and partly managing her performance of science. These are different activities with different orientations.

The introduction of the incentive doesn’t just add a layer to an existing motivation. It restructures the psychological character of the activity. And for essentially byproduct goods, the restructuring destroys the conditions necessary for the thing to occur.

This is sometimes called “motivation crowding” in the economics literature — the phenomenon where adding extrinsic rewards can crowd out intrinsic motivation rather than supplement it. But Elster’s framing is sharper: for essentially byproduct goods, it’s not that motivation is partially crowded out. It’s that the necessary precondition — a particular non-instrumental orientation — is made structurally inaccessible.

The corporate capture of anything that requires genuine attunement is subject to this failure.

Authentic culture. Communities of real practice. The kind of research that discovers things nobody was trying to find. The creative work that solves problems nobody had articulated yet.

These are all, at their core, essentially byproduct goods. They arise from activities pursued for their own sake by people who care about the thing itself. You can build organizations around them. You can benefit from them. You can create conditions that support them. But you cannot directly optimize for them without converting them into something else — a performance of authentic culture, a managed community, directed research, client-service creativity.

The organizations that understand this operate differently. They don’t try to produce the thing; they try to protect the conditions under which it arises and trust that it will. They tolerate apparently inefficient behavior — the scientist following a curiosity that doesn’t obviously connect to anything useful, the team spending time on the hard version of a problem when the easy version would satisfy the metric. They accept that you cannot fully surveil or fully incentivize your way to what you actually want.

The organizations that don’t understand this acquire, systematize, and discover afterward that they’ve acquired the shell.

There’s a deeper consequence for how we think about value.

The standard economic model of value is extractive: you identify something valuable, acquire it, and produce as much as possible. This works when value is separable from the conditions of its production. When what’s valuable is a thing you can hold and use independently of how it was made.

But some value is constitutively inseparable from its conditions of production. The trust is valuable partly because it arose from genuine non-instrumental relationship. The creative insight is valuable partly because it came from real curiosity about a real problem. The essentially byproduct good carries the trace of how it was produced, and the value is in that trace.

Which means the extractive model doesn’t just fail to produce more of the good — it destroys the good it was trying to extract. What you hold afterward has the right name but the wrong structure. It functions like the thing. It resembles the thing. It isn’t the thing.

Elster was writing about mental states. But the structure he identified applies anywhere that the conditions of production are constitutive of value. Wherever genuine orientation — not strategic orientation, not performed orientation, but actual caring about the thing — is a precondition for the thing existing at all.

Those goods cannot be wanted, in the sense that wanting them directly is the one way not to get them.

They have to come sideways, or not at all.