In case you want podcast summary of this topic, here it is. (provided by NotebookLM)
For a slide overview you can click here (provided by NotebookLM)
Basics I — Resolution costs
First of three foundational pieces. These are not results. They are the primitives everything else in the framework leans on, written out once so the notes can stop re-explaining them.
The question nobody asks
Here is something odd about the world that almost never gets remarked on.
Most of it is not determined.
Not “not known by us.” Not determined. The exact microscopic path an excitation takes through a leaf. The precise arrangement of every particle in a stone. Which of countless configurations a magnetic domain on a hard drive happens to be in while faithfully storing a single bit.
We tend to assume the details are all there, settled, and simply beyond our reach. That the universe holds a complete specification of itself and we get to read off small parts of it.
But that assumption has a cost nobody prices. A fully determined universe is one that renders detail nobody asked for — every particle in every stone specified to arbitrary precision, permanently, whether or not any difference between those specifications ever matters to anything.
So the question worth asking is not why don’t we know the details. It is:
Why would anything be determined more finely than something needs it to be?
The primitive
This framework starts from finite capacity.
Not time, not space, not particles. Budget — a limited ability to resolve things, distributed among whatever there is, and never enough to go around.
One clarification before anything else, because the obvious reading is wrong. Limited does not mean fixed. Budget grows as structure accumulates — every relation is itself real and capable of entering further relation, so binding adds a participant rather than merely spending one. Nothing here is a closed system running down.
Scarcity persists for a different reason: what could be resolved grows faster than the capacity to resolve it. Points add; the pairings among them multiply. So the shortfall is permanent and it widens because the system is growing — which means growth cannot solve the problem growth creates.
Bounded at any moment. Not fixed forever.
Everything else follows from what a system does when it cannot afford everything.
And what it does is obvious once stated: it resolves what it must, and leaves the rest alone.
Reality resolves only to the grain at which a difference becomes consequential to something that persists.
Not to the grain at which a difference exists. To the grain at which it matters to something present. Everything finer stays open — not hidden, not missing, simply not yet made to be one thing rather than another.
What “resolved” actually means
The word needs pinning down, because the everyday sense is wrong.
Resolved does not mean every property specified. It means enough alternatives excluded that the current structure can continue consistently.
A leaf absorbing light needs a coarse distinction: absorbed here rather than there. A solar cell needs enough to produce a usable electrical change. Neither needs to distinguish a precise microscopic route through every intermediate possibility, and nothing in either of them is equipped to care about one.
So resolution is not description. It is exclusion — and specifically, the exclusion of enough alternatives to let something carry on being what it is.
Which gives the useful reframe:
Information accumulation is accumulation of constraints, not of complete descriptions.
A newly accumulated informational point may be extremely coarse. It might establish only that A differs from B, or that A is confined to some region. That is enough to be information. A distinction was gained without specifying every finer degree of freedom underneath it.
Splitting, not filling
Here is the exact form, and it is worth having precisely because everything later uses it.
Suppose several configurations are equivalent as far as some structure is concerned. All of them support the same requirement; nothing present has any active reason to distinguish among them. At coarse grain they form one class.
A finer constraint splits it. What was one class becomes two. A finer constraint still splits those again.
Resolution is the splitting of previously equivalent possibilities.
Which changes what detail means. Detail is not an infinite hidden picture waiting to be uncovered. Detail is the accumulation of distinctions that some structure has paid to make consequential.
Three states worth keeping apart
Constraint. Enough alternatives excluded to support what is required.
Tolerance. Several remaining configurations still resolve to the same relevant outcome. This is not error. It is unused freedom — and it is the normal condition of almost everything.
Failure. So much constraint has been lost that incompatible outcomes can no longer be reliably told apart.
The middle one deserves defending, because engineering habits treat it as imperfection.
A magnetic hard drive makes the point cleanly. Logically a cell holds a one or a zero. But there is no infinitely precise physical configuration that is a one — an enormous number of arrangements read the same way. What matters is not that every component is perfectly aligned. It is that enough structure is constrained for the reading process to reliably separate this class from the class that reads zero.
An informational point is constraint sufficient to preserve a consequential distinction — not a perfectly rendered state.
Forcing the physical state to be exact would demand enormous additional constraint and add no distinction anything could use. It would be pure waste.
The limit that keeps this honest
A framework built on “resolution costs” invites an obvious abuse: if everything is merely unresolved rather than absent, then with enough effort anything could be recovered.
It cannot, and the framework says so explicitly.
Recoverable detail ≤ distinctions preserved by present constraint.
If two histories once produced distinguishable structures, but later processes erased enough relations that both now lead to the same present constraint, that distinction is simply unavailable. No amount of finer observation forces an answer when nothing remaining constrains one over the other.
Information loss is previously distinguishable histories becoming equivalent. Budget refines surviving constraint. It cannot recreate a distinction that has genuinely ceased to exist.
One further state is worth naming, because it is often mistaken for loss. A relation can be fully preserved while nothing currently resolves it — dormant rather than erased. A keyboard layout with nobody at it re-binds in minutes. A compressed file whose program is gone is recoverable only by rebuilding the constraint that made the bytes mean anything. Same category, wildly different price, and recovery cost scales with how much surviving structure still points at the answer.
Three states, then: resolved, dormant, erased. Only the last is the hard limit, and the middle one is the ordinary condition of nearly everything.
Why any of this matters
Once resolution is something that must be paid for, a series of otherwise unrelated things stop needing separate explanations.
Why most of the universe is not finely specified — because most distinctions are consequential to nothing present.
Why measurement is expensive rather than merely difficult — because higher resolution means more alternatives made consequential, which means more budget spent narrowing what something is allowed to answer. The cost of precision is not incidental. It is what particularity costs.
Why a stored state can constrain a much finer later answer without having contained it — the storage holds constraint, the later measurement supplies the additional distinctions, and the answer is neither arbitrary nor pre-rendered.
And why the sensible operating rule for anything with a limited budget is:
Spend only enough constraint to preserve the distinction. Leave the rest as freedom until something makes it matter.
That rule is the whole framework in one line. Everything after this is a consequence of taking it seriously.
Next in this series: what “coarse” actually means, and the three ways it is normally misread.
If interest in the full framework of how this deduction made of, you can explore it here. There is a lots to explore together.
