Does completion derive the Born rule?

What must a probability derivation actually establish?

Completion tells us which event is counted as an available result. Probability tells us how frequently the different results occur under a specified preparation. These are related tasks, but a device can finish reliably while giving many different distributions. A deterministic counter, a classical random selector and a quantum instrument can all raise the same finished flag. The flag alone does not determine the statistics of the label it accompanies.

For SMM/DTT, the constructive question is whether source dynamics and the allowed completion tests force the quantum probability law. A satisfactory answer must state the source states, admissible effects, preparation procedures and rule for combining exclusive alternatives. It should then identify which of these ingredients has been derived from the generative construction and which was supplied as an interface. The measurement discussion supplies exact instruments inside quantum mechanics; this page asks about the extra step of reconstructing their probabilities.

The six nested inner levels can organize dependent work and the seventh can expose a completed result. Neither the nesting count nor the name of that final event specifies a numerical measure on alternative histories. A measure is mathematical structure with consequences that can be checked across different experiments.

Normalization permits many laws

Consider a normalized two-level state ψ=√(3/4)|0⟩+√(1/4)|1⟩. Standard projective measurement gives probabilities 3/4 and 1/4. However, a rule that squares those weights again and renormalizes gives 9/10 and 1/10. Both assignments are nonnegative, sum to one and give certainty on the two basis states. Therefore normalization, basis certainty and the availability of a completed outcome do not by themselves select the Born rule.

This alternative is a comparator, not a proposal for physical quantum theory. Its value is diagnostic: any claimed derivation using only those three properties leaves this freedom untouched. The missing assumption must constrain probabilities across more than one preparation or one measurement arrangement. A source model should explicitly exhibit the operation that excludes the alternative instead of simply declaring its weights unphysical.

Equal counting of hidden source paths has the same issue. Splitting one path into two differently named subpaths changes a raw count unless their relation is controlled. A counting measure becomes a probability law only after one specifies which distinctions are physically meaningful, how equivalent descriptions are identified and why a preparation samples those distinctions with the chosen weights.

A finite check: split one outcome into two

Suppose a measurement has raw event weights 3/4 and 1/4. Refine the first event into two identical subevents with weights 3/8 and 3/8; afterwards forget which subevent occurred. Ordinary addition returns the same coarse probability, 3/4. This is a basic consistency requirement when refinement changes only the resolution of the stored label.

The squared-and-renormalized comparator behaves differently. Before refinement its first probability is 9/10. After refinement, the three squared weights are 9/64, 9/64 and 4/64. Their normalized values are 9/22, 9/22 and 2/11. Merging the first two now gives 9/11. The original event has acquired a different probability simply because its label was subdivided.

A physical implementation can realize this example using effects (3/4)I and (1/4)I, then replacing the first effect by two copies of (3/8)I. Here an effect is an operator representing a possible yes–no event, and I is the certain event. The example isolates a demand on event composition. If a source model claims that such refinements are not equivalent, it must explain what additional operation changes the actual apparatus and how that change is observable.

What a probability representation theorem can contribute

Busch's theorem for generalized quantum observables starts with the full set of positive Hilbert-space effects, a normalized nonnegative valuation and additivity whenever effects form an allowed measurement. It obtains the trace form v(E)=Tr(ρE), including the two-dimensional case. This provides a precise conditional route from an operational effect structure to quantum probabilities.

For a source programme, the task is to justify that effect structure and its consistency conditions. Calling an inner history an amplitude does not already establish that every relevant completed question corresponds to a positive operator. Likewise, writing ρ=ψψ† and reading its diagonal as frequencies already imports the essential rule unless a preceding argument has derived the state–effect pairing.

The representation result should therefore be used as a target interface with explicit inputs. One can first show that a finite collection of completion procedures obeys positivity, normalization and refinement consistency. One can then investigate whether the available operations become rich enough to support the required family of effects. This turns a large foundational claim into a sequence of inspectable constructions.

Preparations and composite tests sharpen the requirement

A classical lottery that prepares ρ with probability s and σ otherwise should produce the same statistics as the mixed preparation described by sρ+(1−s)σ, when the lottery record is unavailable and no extra interaction occurs. This affine dependence is an operational consistency condition. It distinguishes a physical mixture from a coherent superposition, whose interference can depend on relative phase.

For example, the equal mixture of |0⟩ and |1⟩ gives a balanced X measurement, whereas the coherent state |+⟩ gives the positive X result with certainty. A completion scheme that retains only the two population weights cannot predict both experiments. The composition guide shows why phase-bearing joint states and their admissible later tests must be specified together.

References matter as well. A device may appear unbiased on a few isolated inputs but distort correlations with another system. Probability reconstruction should remain consistent when a system is embedded in a larger preparation and only one part is measured. This requirement does not automatically choose the standard tensor product; it creates a further connection to the composite-system problem.

A useful first result for contributors

Begin with one source realization of a two-level carrier and a finite set of preparations that includes basis states, interference states and mixtures. Specify two-outcome completion tests plus one physical refinement of an outcome. Compute the entire probability table from the source rule without identifying amplitudes with probabilities in the definitions. Check whether refinement, mixture equivalence and repeatability impose a common positive state–effect pairing.

If they do, state exactly which family of measurements the result covers. If alternative probability assignments remain possible, exhibit one and identify an additional physically motivated source operation that distinguishes it. Either outcome gives a concrete advance: a theorem on a stated operational domain or a counterexample showing why a proposed reconstruction is incomplete. The eventual goal is an explanation of the measure as well as the record, with the two tied to the same source dynamics.

Continue reading

• Measurement, completion flags, and retained outcomes →
• Quadratic completion and phase-insensitive records →
• Hermitian records from complete tests →
• How phase-compatible systems compose and become entangled →

Read this discussion as part of a path

• From completion to quantum predictions →