Why probabilities alone can be too little information
In a standard coherent two-path experiment, the probabilities associated with the separate paths do not determine every subsequent measurement. Relative phase matters when the paths recombine. The states (1,1)/√2 and (1,−1)/√2 have the same separate-path probabilities, but a suitable mixing operation routes them to different outputs. [B05] [B19]
This is a concrete lesson for completion. A retained description may discard a common global phase without losing any allowed single-system test, yet it cannot discard an experimentally accessible relative phase and still predict interference. The appropriate equivalence of histories is determined by the full admitted test family, not only by one detector display.
Change phase, then reduce coherence
What a retained phase can change
This is a standard two-path interference model. Phase changes the recombined outputs; the coherence factor reduces contrast. It illustrates information that an intensity-only snapshot can miss.
The two output probabilities in this balanced model are P₊ = (1 + V cos φ)/2 and P₋ = (1 − V cos φ)/2. They sum to one for every displayed setting. At full coherence V = 1, changing φ by π exchanges a bright output with a dark one. At V = 0, neither output responds to φ.
The wave pattern in the plot represents phase-dependent intensity across a schematic screen. The two bars separately represent the two outputs of a balanced recombiner. They are related teaching examples, not experimental data and not a simulation of a newly derived SMM/DTT apparatus.
The dual-time interpretation asks how an underlying source retains the structure needed for this phase-sensitive behaviour and how a completed apparatus record becomes available. The interference formula itself uses standard quantum rules; it should not be presented as a consequence of simply naming two temporal roles. [B19] [P03]
What measurement must add
A measurement interaction correlates a system with distinguishable apparatus states. Coupling to the environment can suppress accessible interference between apparatus alternatives. In a familiar schematic form, a superposition becomes a correlation among system, apparatus, and environment. [B11]
When different environmental states have very small overlaps, off-diagonal terms become small in the reduced apparatus description. This explains how a record can become stable for a restricted family of observations. It does not by itself select one unique global outcome from the full correlated state.
DTT can use the distinction between generative alternatives and retained records to organize the measurement question. A stronger dynamical account would need to state its probability law, record-stability mechanism, and any additional outcome-selection law. Conscious attention is not a substitute for those physical specifications.
A published finite recorder
A finite capture window
For the published recorder model, p(t) = 1 − cos²ᴸ(ωt/2). L counts apparatus stages. The full plotted period makes its recurrence visible.
The quaternionic reconstruction includes a finite recorder with a specified Hamiltonian and completion probability p(t) = 1 − cos²ᴸ(ωt/2). At t* = π/ω, the formula gives unit completion probability. Near that time, the error scales as sin²ᴸ(ωu/2), with u the displacement from the capture time. [P01]
The integer L counts stages of this apparatus. It is not the six inner roles of the book illustration, and it is not the three acquisition depths of the receiver recurrence. Those integers belong to different constructions.
The full period shown in the plot also reveals a limitation: the isolated finite model is recurrent. The probability decreases again beyond the capture window. Reliable long-term storage therefore requires an additional retention mechanism or a specified readout protocol. Reaching a certain capture time and keeping the result available indefinitely are distinct requirements.
Entanglement, correlation, and signalling
A common-source account can motivate the study of nonfactorizing correlations, but a common origin alone does not calculate quantum entanglement. A complete model must state the composite state space, permitted local operations, probability assignments, and how the observed joint statistics are obtained. The phase-balancing issue in the preceding chapter is one part of that structure. [P02] [P03]
Nor should an internal generative ordering be depicted as an operational channel for sending controllable messages faster than light. An acceptable physical account must reproduce the no-signalling behaviour of the experiments it models and show how that follows from its accessible operations. The distinction between a hidden construction order and physical communication is crucial.
The productive target is a single source dynamics that yields phase relations, composition, records, and their calibrated propagation together. The present mathematical results constrain parts of that target, while the familiar experiments supply demanding benchmarks for any proposed realization.

