Does completion explain quantum measurement?

Which part of measurement does completion address?

DTT separates inner work from a completed record available to an observer. This gives a natural question about measurement: when has an apparatus actually produced a result, and what makes that result usable for a subsequent operation? A flag can answer that question only if it is correlated with the required work by an explicit dynamics.

The wider measurement problem includes further questions about probabilities, the status of individual outcomes and the persistence of records. A source model can make progress on one without having solved all of them. The useful comparison is between a certificate merely appended by definition and one generated autonomously from a work register, a fresh apparatus state and a specified interaction. The latter makes quantitative predictions about errors, delays and retained information.

An instrument describes every outcome

Within standard quantum mechanics, an instrument consists of completely positive outcome maps Iₖ whose sum preserves trace. For one Kraus operator per outcome, Iₖ(ρ)=MₖρMₖ† and ΣMₖ†Mₖ=I. The trace pₖ is the outcome probability; Iₖ(ρ)/pₖ is its conditional state when pₖ is positive.

Take projectors P₀ and P₁ onto a two-level basis. On |+⟩=(|0⟩+|1⟩)/√2, the two outcomes each have probability one half. Keeping both outcomes without reading the label gives P₀ρP₀+P₁ρP₁=I/2. Keeping only P₀ρP₀ has trace one half and represents a selected branch. A completion account must retain that success frequency rather than describe the normalized accepted state as the whole process.

Can work complete without measuring the payload?

Yes, in an explicit quantum model. Let V be a fixed unitary operation on an unknown payload and define the accepted map I_acc(ρ)=pVρV† with fixed 0<p≤1. Its trace is p for every input. The acceptance flag therefore does not preferentially choose payload states. The same identity holds when the payload is entangled with an untouched reference.

This separates certification of work from identification of the unknown input. A mechanism could make the controlled work equal V throughout an accepted region of source states. Then source preparation and its acceptance test determine p. The positive task is to derive that channel from the interaction, including its duration and energy budget. Payload independence does not imply independence from the apparatus preparation or clock: those remain measurable aspects of the completion process.

What would make repeated attempts legitimate?

Complete the example with M_acc=√p V and M_fail=√(1−p)W, where W is another known unitary. Their effects sum to identity. After recorded failure, W† restores the unknown payload, including correlations with a passive reference. With fresh independent apparatus preparations, restored retries have the same success probability.

The first success on trial n has probability (1−p)ⁿ⁻¹p. If p=1/4, the mean number of attempts is four, and success within five attempts has probability 1−(3/4)⁵=781/1024, about 0.763. These numbers depend on recovery and fresh preparation. An irreversible failed branch need not permit the same retry law. A physical clock counting all attempts also differs from a record index counting only successes, so elapsed time must include failures and recovery.

Why registration can change later generation

Compare two protocols on |+⟩ using the same Hadamard operation H and unread Z measurement D. If D comes first, the state becomes I/2 and remains I/2 under H. A final Z measurement is balanced. If H comes first, |+⟩ becomes |0⟩; D leaves it unchanged and the final result is certainly zero.

This is an exact ordinary quantum effect of operation order. It makes a useful benchmark for source models because an intermediate registration can erase coherence used by later work. To make it a distinctive DTT prediction, the source law must determine what physical operation implements registration and constrain its joint relation to completion probability, duration or energy. The half-versus-certain outcome alone is already explained by the supplied quantum comparator.

What must a completed record retain?

A certificate written at one instant does not guarantee later readability. The source can re-entangle the register, overwrite its label or allow a control to reveal information omitted from the public state. A completion model should therefore specify retention time, permitted interrogation and the error probability after declared delays. Predictive sufficiency must be checked under the future operations actually allowed.

Quantum correlations expose the issue sharply. The coherent state (|00⟩+|11⟩)/√2 and the equal mixture of |00⟩ and |11⟩ give the same matching Z labels. Their X⊗X expectations differ: one for the coherent state and zero for the mixture. A record that preserves only the Z correlations cannot predict both tests. The model must say whether completion retains the coherence, physically destroys it or restricts the future test family.

What would amount to a fuller measurement explanation?

The instrument examples adopt the Born probability rule and quantum state space. They make completion precise inside that framework; they do not derive the rule or establish an interpretation of individual outcomes. A stronger reconstruction would obtain the allowed positive states and effects from source principles and explain how the apparatus acquires stable accessible correlations.

A concrete next result is still substantial: derive one autonomous source instrument, prove its certificate accuracy and payload disturbance bounds, calculate every outcome frequency, and predict the lifetime of its records using the same parameters. Compare it with a conventional apparatus model under equal resources. The six nested roles can organize the dependent work, while the seventh exposes its result. Completion then contributes a physical mechanism for available records, with its remaining explanatory tasks stated precisely.

An autonomous certificate in a three-position history

A small model shows what it means for a flag to follow actual work. Take history positions |0⟩, |1⟩ and |2⟩, with the payload initially at position zero. Let U₁ and U₂ be its prescribed work operations. Use a time-independent Hamiltonian with forward links a g|1⟩⟨0|⊗U₁ and a g|2⟩⟨1|⊗U₂, together with their adjoints. Here g is a frequency and a is the supplied action scale. Propagation to the last position applies V=U₂U₁.

Changing basis along the history removes the U operators from the clock's three-by-three matrix. The remaining chain has eigenfrequencies 0 and ±√2g. Starting at zero, the amplitude to reach two is [cos(√2gτ)−1]/2. Reading the final-position projector therefore realizes an accepted payload channel with the following probability:

At τ=π/(√2g), p=1; at τ=2π/(√2g), p=0. This finite example produces unbiased work certification from an explicit evolution. It also exposes a limit: a final-position flag is transient. The closed history can move backwards and erase its completion indication. An additional retention mechanism, stopping control or declared observation window is needed for a lasting record.

The finite Hamiltonian can be made nonnegative by adding a sufficiently large scalar multiple of identity; that shift leaves every displayed probability unchanged. Positive energy alone therefore does not pick the observation time, the action scale, or the retention rule. A complete implementation must account for those ingredients, including whatever physical preparation supplies the initial history position.

Which errors does a certificate actually bound?

A useful completion statement needs two different error measures. A false-positive bound asks how often the flag says finished when the required work was not performed. A disturbance bound asks how closely the accepted payload operation matches the desired channel. A flag can be reliable about reaching a history position while the operation stored at that position is systematically wrong. Conversely, accurate work can occur even when a noisy flag fails to announce it.

Calibration should therefore compare the joint flag-and-payload map, including an untouched reference system when entangled inputs are permitted. Testing only a preferred basis misses errors in relative phase. For example, the identity and Z gates agree on the output labels of |0⟩ and |1⟩, but send |+⟩ to different interference states. A work certificate claiming arbitrary-input accuracy must include tests capable of seeing that distinction.

Conditioning also makes small success probabilities demanding. If an ideal unnormalized accepted map differs from the realized map by trace-norm error ε, dividing by a small success probability can magnify the error in the reported conditional state. The precise bound depends on the norm and how the success probabilities are compared. Reporting the accepted state's fidelity alone hides this sensitivity. State both the unconditional branch error and a lower bound on acceptance over the admitted input class.

Repeatable outcomes and stable storage are separate tests

An immediate repeated projective measurement can be perfectly repeatable without producing durable storage. After a Z outcome, measuring Z again yields the same label if nothing intervenes. If a subsequent X gate is allowed on the record, the label flips. If the apparatus undergoes uncontrolled dynamics, readability may decay. Repeatability is consequently a claim about a specified continuation, not a guarantee against every conceivable operation.

For a simple storage benchmark, suppose a classical register flips with independent probability q at each declared time step. Starting from a known label, the probability that the stored label is wrong after n steps is [1−(1−2q)ⁿ]/2. With q=0.01 and n=10, it is about 0.0915. A redundant or protected record may improve this number, but its extra memory and recovery work belong in the resource account.

A quantum record can also preserve the label while losing access to coherence between labels. Whether this counts as successful storage depends on the intended future tests. A public result used to select the next classical instruction has different retention requirements from a coherent control qubit. The source theory should define that task before declaring a record complete.

Where an outcome interpretation enters

A unitary system–apparatus interaction can correlate two orthogonal input states with two distinct pointer states. Linearity then maps their superposition to a correlated superposition. Tracing an inaccessible environment may explain why the accessible pointer behaves like a mixture for a stated set of tests. The calculation, by itself, does not choose an interpretation of why an individual reader experiences one result.

A DTT account can place that interpretive question at outer completion, but it should say whether completion is an ordinary instrument, a new stochastic law, a restriction of available observations or an ontological claim about records. These options have different mathematical obligations. The dedicated probability discussion isolates the remaining task of obtaining outcome weights, while the recovery discussion tests what the unobserved branches physically retain.

The same outcome statistics can conceal different instruments

Consider one instrument with M₀=|0⟩⟨0| and M₁=|1⟩⟨1|. Compare it with N₀=|+⟩⟨0| and N₁=|+⟩⟨1|. In both cases the effects are P₀ and P₁, so every input gives exactly the same probabilities for the first recorded label. Yet the instruments leave different payload states: the first retains the measured basis state, while the second prepares |+⟩ after either outcome.

An immediate second Z test distinguishes them. Conditional on the first label, the projective instrument repeats it with certainty. The prepare-after-reading instrument gives either label with probability one half. Both maps are legitimate complete quantum instruments. Consequently, a catalogue of first-outcome probabilities does not establish which future record dynamics a source implements.

A completion model should therefore supply the state change associated with each outcome and the behavior of the retained label, not only an acceptance effect. This is especially important if the completed payload controls later inner work. Two devices can agree perfectly on their initial flag statistics and disagree on every subsequent phase-sensitive or repeatability test.

Continue with the articles

• A quantum source for generation and completion

  A positive phase–dilation Hamiltonian ties one phase encoder to a powered coordinate and controlled work. It gives an explicit quantum realization of generation and completion while retaining its constitutive interfaces as inputs.

• Completion without biasing an unknown input

  A factorized accepted channel gives the same success probability for every admitted unknown input, including inputs entangled with a reference. The failure channel and the amplification of errors by conditioning require separate checks.

• Undoing rejected partial work

  A failed completion attempt can be coherently undone by reversing its actual source-controlled work prefix. The recovery restores an unknown payload under explicit inverse-access, memory, and control assumptions.

• Reliable completion by repeated attempts

  Fresh recoverable attempts produce a geometric success law and an explicit reliability bound for a finite archive. Improved completion probability consumes prepared sources, flags, control, and runtime.

• A record that carries its own completion certificate

  An autonomous history Hamiltonian can generate both a data record and a flag tied to its completed write. The exact construction retains incomplete outcomes, coherent intermediate stages, and finite observation windows.

• What spatial readout forgets

  Different inner states can share one spatial readout and still have different successors. A concrete counterexample and a kernel-preservation criterion identify the memory needed for autonomous completed evolution.

Continue reading

• Measurement, completion flags, and retained outcomes →
• A record that carries its own completion certificate →
• Completion without biasing an unknown input →
• When a readable record is enough to predict the future →

Read this discussion as part of a path

• From completion to quantum predictions →