Past, present and future organize our access to events. They also enter physical models, historical inference and philosophical claims about existence. Distinguishing these uses helps us ask what a present record actually tells us about an earlier or later state.
A present record can refer to another time
A photograph exists now while depicting an earlier exposure. A forecast exists now while referring to a possible later event. Their current physical availability does not make their referents simultaneous with the act of reading them.
Likewise, an astronomical image combines signals emitted at different times and received in a particular observation window. Its timestamp labels the recording procedure. Reconstructing the source events requires information about propagation and the instruments.
These distinctions leave a substantive question: how does a surviving record constrain events outside its own immediate presence? The answer involves the processes that created and preserved the record, together with the model used to interpret it.
Compare a reversible update with an erasing update
Consider a register with four possible values, 0, 1, 2 and 3. Let its update add one modulo four. If the present value is 2 and the rule is known, the preceding value must have been 1 and the next value will be 3.
The update is invertible. Each present state has one predecessor and one successor. The mathematics gives equal access to the two directions once a complete state and the rule are supplied. It does not explain why an actual observer has one collection of memories rather than another.
Now use an erasing rule that sends both 0 and 1 to 0, and both 2 and 3 to 1. A present output of 1 has two possible input values. The output determines the next step under a specified rule but does not retain enough information to reconstruct which input occurred.
This loss can be represented without destroying information in an enlarged description. Store an extra bit distinguishing the two inputs. A record containing both the coarse output and that bit can distinguish all four original possibilities. The apparent irreversibility depends on what remains accessible.
Uncertainty about a future is not one thing
Even an invertible deterministic law need not produce a useful prediction from an incomplete observation. Suppose the four-state register is observed only as even or odd. An even reading leaves both 0 and 2 possible. The next parity is fixed, but the next full state is not.
Predictive sufficiency therefore depends on the question being asked. A record can be enough to predict one observable and insufficient for another. The phrase “the future is determined” needs a specified law, state description and target quantity.
A probabilistic model adds a different kind of structure: it assigns conditional probabilities to later outcomes. Whether those probabilities represent fundamental randomness, incomplete information or an effective reduction requires further argument. The mere presence of a probability in a forecast does not settle that interpretation.
Local presence and distant simultaneity differ
Two events meeting at one detector can be compared locally. Assigning a time to a distant event requires a synchronization procedure or another coordinate convention. Relativity constrains those assignments and preserves causal relationships while allowing different orderings of spacelike events.
It follows that a global present cannot simply be read off from every observer's clock at once. It does not follow that local records have no temporal organization. The simultaneity demonstration makes that distinction explicit.
Philosophical positions concerning the existence of past or future events add interpretive commitments to this physical structure. An essay should identify those commitments rather than hide them in a diagram or in the grammatical tense used to describe an equation.
Ask what completion preserves
Within SMM/DTT, a completed record matters because later operations can use it. The book's renewal and persistence example shows how an accumulated memory can survive changes in its carrier. Its reconstruction requirements connect that record to specified future tests.
The register examples suggest a practical standard. State which distinctions survive completion, which are discarded, and which predictions depend on them. If two inner histories give the same completed state, explain the family of later observations for which they are equivalent.
This makes past and future useful research terms: the past constrains the available record, while future tests determine whether that record is sufficient. The broader ontology remains a question to be argued alongside, rather than substituted for, these physical relationships.
Sources and further reading
- The finite-register examples are original constructions. Compare the book's three different arrows.
- Continue with distributed records and relativity and becoming.