What this paper is arguing

Some problems become hard because we are solving them inside a frame we no longer notice. We inherit concepts such as motion, identity, distance, agency, intelligence, value, ownership, and time, then treat those concepts as if they are simply there.

This paper describes a thinking exercise I will call Null-Origin Reasoning. The exercise asks:

What is the smallest distinction that has to exist before this problem can even be stated?

It is not a theory of physics. It is not a shortcut to impossible technology. It is not a replacement for mathematics, experiment, or engineering. Most of the time it will produce nothing except a clearer understanding of why the original problem is hard.

That is still useful.

Introduction

A good model becomes invisible.

Use it long enough and its abstractions stop feeling like abstractions. They start to feel like reality itself. Space, time, object, motion, cause, identity, intelligence, value — we talk through these ideas so often that we forget they are doing work.

Most of the time, that is fine. It would be absurd to redesign your ontology every time you build a bridge, debug a program, or plan a railway timetable. Mature models are useful because they save us from having to begin again. They give us a furnished world and let us get on with arranging the furniture.

But occasionally a problem resists every rearrangement.

At that point, it can be useful to ask whether the furniture is the wrong unit of analysis.

This essay is about a deliberately uncomfortable thinking exercise for those moments. I call it Null-Origin Reasoning, though the name sounds grander than the method deserves. Its purpose is simple: strip a problem back until you can see the assumptions required to state it.

It starts not with “how do we solve this?” but with:

What has to be true before this problem exists?

That question can be clarifying. It can also be useless. The difference is whether the reframing returns to the world with a testable consequence, a better engineering decision, or at least a sharper statement of the real constraint.

If it does not, then it is only philosophy in a dramatic coat.

The method

The ordinary way to solve a problem begins inside the inherited model.

For propulsion, we ask how to generate more thrust, reduce mass, improve efficiency, or use a better energy source.

For AI, we ask how to improve inference, training, memory, reasoning, or alignment.

For software architecture, we ask how to scale the database, split the service, reduce latency, or redesign the API.

Those are usually the right questions.

Null-Origin Reasoning is for the rarer case where you suspect the problem is being shaped by assumptions hidden beneath the ordinary questions.

The method is:

1. State the problem in ordinary language.
2. List the concepts the statement assumes.
3. Ask which of those concepts might be constructed rather than fundamental.
4. Ask what minimum distinction must exist before the problem becomes meaningful.
5. Rebuild the problem from that minimum distinction.
6. Ask what changed.
7. Most importantly: ask whether the change produces a test, a prediction, a design decision, or a clearer constraint.

The last step is the safeguard. Without it, this method can generate endless profound-sounding nonsense.

A reframing is not valuable because it sounds deep. It is valuable only if it changes what you can explain, measure, build, or rule out.

The propulsion temptation

Propulsion is the most tempting example because it is where the method almost becomes exciting.

The normal question is:

How do we move an object from here to there?

That question already assumes a lot. It assumes an object with a stable identity. It assumes a here and a there. It assumes distance as something to be crossed. It assumes motion through an external space. It assumes inertia. It assumes energy cost. It assumes that changing location means traversing intermediate locations.

Null-Origin Reasoning asks a different question:

What has to exist before “moving from here to there” is meaningful?

At once, the problem looks less obvious. If space, distance, and location are not fundamental but emergent, perhaps propulsion is not the deepest description of what is happening. Perhaps “move the object” is only a surface-level way of describing a change in relational state.

That sounds seductive.

It is also exactly where the method needs to disappoint us.

Even if space were emergent from some deeper structure, changing an object’s relation to that structure would still require changing something real. If the deeper substrate carries information, configuration, constraint, or energy, then editing the relation is not free. The cost has not vanished. It has moved.

This pattern is familiar from thermodynamics.

Temperature is emergent. Entropy is emergent. They arise from the statistical behaviour of many microscopic states. But knowing that temperature and entropy are emergent does not let us bypass the second law. Maxwell’s demon does not get free order merely by working at a lower level. The cost returns through information, measurement, and erasure.

The same caution applies to propulsion.

Calling motion “relational editing” does not make the energy bill disappear. It only changes the language in which the bill is written.

This is why warp-drive ideas are such a useful warning. The fantasy version says: if we can manipulate spacetime itself, perhaps we can avoid ordinary propulsion. But serious work in the warp-drive literature repeatedly shows how stubborn the underlying costs are. Exotic geometry does not automatically remove the need to accelerate mass, provide energy, satisfy constraints, or account for what the spacetime configuration physically consists of.

So the useful conclusion is not:

Null-Origin Reasoning points toward a warp drive.

It is:

Null-Origin Reasoning shows exactly what a warp-drive claim would have to prove.

It would need to show not merely that motion can be redescribed at a deeper layer, but that the deeper-layer operation has a lower physical cost than ordinary propulsion. Without that, the method has produced no escape — only a clearer burden of proof.

That is a good failure.

Good failures matter

A bad thinking method produces excitement where it should produce caution.

A better method sometimes produces disappointment, but disappointment with sharper edges.

In the propulsion example, Null-Origin Reasoning begins with the intuition that movement may be a surface phenomenon. It asks whether the need for engines comes from assumptions about objects moving through space. It then runs into a familiar wall: deeper descriptions do not abolish higher-level constraints for free.

That failure is instructive because it blocks a false shortcut.

It tells us that “emergent” does not mean “optional.” A thing can be emergent and still be binding. A higher-level law can arise from lower-level behaviour and still be impossible to cheat from inside that behaviour.

This matters beyond physics. People often assume that finding a deeper layer will make a difficult problem easier. Sometimes it does. Often it only moves the difficulty to a place where it is harder to see.

In software, replacing a monolith with microservices does not remove coupling. It relocates coupling into contracts, networks, deployment, ownership, observability, and failure modes.

In organisations, replacing hierarchy with autonomy does not remove power. It relocates power into norms, access, incentives, and informal influence.

In AI, replacing explicit rules with learned behaviour does not remove design responsibility. It relocates design into training data, feedback loops, evaluation, system prompts, and deployment context.

Null-Origin Reasoning is useful when it helps us see that relocation.

It is dangerous when it mistakes relocation for escape.

When “add another thing” is science, and when it is a patch

Many models survive by adding something new.

An unseen particle. A hidden variable. A new field. A new layer. A new service. A new team. A new policy. A new abstraction.

That is not automatically suspicious. Some of the best scientific ideas began as additions introduced to explain anomalies. The neutrino was proposed because beta decay appeared to violate conservation laws. It was not directly observed at the time, but it later became part of a wider, testable structure.

The problem is not adding a thing. The problem is adding a thing that explains only the discomfort that required it.

A useful addition tends to have several qualities:

• it predicts more than the anomaly it was invented to explain;
• it connects to independent evidence;
• it could have failed a test;
• it reduces confusion elsewhere;
• it does not merely add adjustable knobs until the model fits.

A patch has the opposite shape:

• it exists only to rescue the model at one point;
• it is difficult to test independently;
• it adds parameters without explaining why they take their values;
• it makes the model harder to challenge;
• it turns a question into a label.

Null-Origin Reasoning is useful in the patch case. It asks:

What assumption made this extra thing necessary?

Sometimes the answer will be boring: the extra thing is necessary because reality is like that.

Sometimes the answer will be revealing: the extra thing is compensating for a frame that should be reconsidered.

The method does not tell you which answer is true. It only gives you a way to ask the question cleanly.

The Standard Model, carefully

It is tempting to point at modern physics and say: here, this is where the method should matter.

That temptation should be handled carefully.

The Standard Model is one of the most successful scientific theories ever built. Any outsider gesture toward it should begin with humility. The fact that a theory has open questions does not make it weak. Successful theories often create sharper unanswered questions precisely because they explain so much else.

There are two complaints people often blur together.

The first is the free-parameter problem. The Standard Model contains constants that are measured rather than derived: masses, couplings, mixing angles, and related quantities. Asking why those values are what they are is a serious mainstream question. Physicists do not need Null-Origin Reasoning to notice it. They are already working on it.

So the honest statement is:

Null-Origin Reasoning does not add much to the free-parameter problem by itself.

The second complaint is more general: the feeling that the theory describes the pattern of reality without making that pattern feel necessary. It tells us what fits, but not why this structure and not another.

That may be a profound clue.

It may also be an aesthetic discomfort masquerading as an argument.

A final theory might make the observed structure feel inevitable. Or reality may simply contain facts that are not derivable from something deeper in the way we want. The desire for necessity is not proof that necessity exists.

Null-Origin Reasoning can sharpen the question:

Which assumed primitives would have to become derived before the pattern felt explained?

But unless that sharpening produces mathematics, prediction, or experiment, it remains a philosophical pressure, not a scientific result.

That boundary matters.

The method turns on itself

A method that asks other models to expose their assumptions has to expose its own.

Null-Origin Reasoning begins with “distinction.” It asks what minimum difference must exist before a problem can be stated.

But distinction is not obviously primitive.

To say “A is not B” already seems to require two terms and some relation between them. That relation may already smuggle in structure. If we speak of one distinction giving rise to another, we may already be assuming something like sequence. If we talk about transition, we may already be borrowing something time-like.

So the “zero” in Null-Origin Reasoning is not truly zero.

It is the lowest point the method can currently articulate. That is much less impressive, but more honest.

The method does not reach bedrock. It is a stripping exercise, not a proof of what lies beneath reality. Its first move may already contain assumptions it cannot justify.

That does not make it useless. Many useful tools rest on imperfect foundations. But it does mean the method should be used with modesty.

It is not a way to declare what is fundamental.

It is a way to notice what you have been treating as fundamental without checking.

Uses outside physics

The safest uses of Null-Origin Reasoning are not in speculative physics. They are in places where a reframing can change a design decision.

In software architecture, instead of asking:

How do we scale this service?

you might ask:

What assumption created the scaling problem?

Perhaps the issue is not raw traffic, but shared state. Or synchronous dependency. Or ownership boundaries. Or a read model trying to serve two incompatible use cases. The “scale” problem may really be a coupling problem.

In AI product design, instead of asking:

How do we make the agent more intelligent?

you might ask:

What distinction are we using to recognise intelligence here?

Maybe the product does not need intelligence at all. Maybe it needs memory, adaptation, retrieval, planning, explanation, emotional fluency, or simply better workflow integration. “Intelligence” may be hiding several different requirements under one flattering word.

In organisational design, instead of asking:

How do we make teams more autonomous?

you might ask:

What decision is the team unable to make, and why?

The answer may reveal that the problem is not autonomy as a value, but unclear authority, missing information, fear of blame, or incentives that punish local judgement.

In personal systems, instead of asking:

How do I become more disciplined?

you might ask:

What assumption makes discipline necessary here?

Maybe the real issue is friction, fatigue, environment, social expectation, or a goal inherited from someone else.

In each case, the method does not solve the problem. It changes the first question. Sometimes that is enough to make a better solution visible.

What this is actually good for

Null-Origin Reasoning is a tool for finding hidden primitives.

It asks what a problem assumes before it begins. It strips away inherited categories, not because inherited categories are bad, but because sometimes they hide the shape of the difficulty.

Its value is modest but real.

It can expose a false shortcut, as in the propulsion example.

It can separate a real open problem from a vague dissatisfaction, as in the Standard Model example.

It can reveal when an “extra thing” is doing explanatory work and when it is only patching discomfort.

It can shift engineering attention from improving a mechanism to asking why the mechanism is needed.

But it cannot produce truth by sounding deep. It cannot replace domain knowledge. It cannot skip mathematics. It cannot turn metaphysics into science without prediction. It cannot make costs disappear by renaming them at a lower level.

The simplest version is this:

Reframing a problem at a deeper layer can expose assumptions, but it does not abolish the constraints those assumptions were carrying.

That is the discipline.

Not “think back to zero and escape the world.”

More like:

Think back far enough to see what you smuggled in — then return to the world and check whether anything actually changed.

The deepest question the method leaves is not “how do we move through reality?”

It is:

What distinction makes movement, distance, and destination appear as separate facts in the first place — and does answering that change anything we can measure?

If the answer is no, the exercise may still have clarified our language.

If the answer is yes, then it has earned the right to become more than an essay.

---

Further reading: the serious version of asking which transformations are possible belongs near David Deutsch’s constructor theory. The real physics around emergent spacetime belongs with work by Jacobson, Van Raamsdonk, Verlinde, Cao, Carroll, Michalakis, and others. Landauer’s principle and the Maxwell’s demon literature are useful cautions against assuming that deeper-layer control removes higher-layer costs. Bobrick and Martire’s work on physical warp drives is relevant to the claim that exotic geometry does not straightforwardly remove the need for propulsion. Kuhn, Lakatos, and Feyerabend remain useful background for thinking about how mature paradigms harden their categories.