On Monday, January 13, 2003, at 02:40 PM, Jesse Mazer wrote:
I think the number of multiple past histories (other than the _actual_ one) which "lead to exactly the same present state" is as close to zero as one would like to calculate.Tim May wrote:On your point about "many pasts are fundamentally caused by quantum uncertainty in memory devices," I strongly disagree. There is only one past for one present, whether RAMs dropped bits in recording them or historians forgot something, etc.What do you mean by this, exactly? In a deterministic universe with time-symmetric laws, there'd be only one possible past history for a given present state, and only one future history as well, while in a universe with stochastic laws, or deterministic laws in which paths in phase space could converge, there might be multiple past histories that would lead to exactly the same present state.
(This is captured by the formalism of observations, as well. Even with Uncertainty, all honest observers will report the same outcome of an experiment. We have not seen a violation of this, nor is one expected. There are various ways to look at this, including the topos-theoretic view of subobject classifiers. But the point is that in our history either an event happened or it did not. This is independent of whether the event was observed, recorded, written about, remembered, etc.)
But this is a topic of great fascination for me, and I hope we can continue to discuss it. I am quite strongly persuaded that "many pasts for a particular present" is not a reality.
Understand that I am not including "current interpretations," as in "Some historians think the Roman Empire fell because of lead in their plumbing" sorts of theories of the past. I am referring to space-time events.
As noted, I also view time and events as a lattice. But lattices have certain properties of importance here. More on this later.
As to why there is only a single past but multiple futures, this is implicit in the measurement process. (I doubt you will find this convincing unless I expand on this.)
Consider the "There will be a sea battle next year" proposition, the favorite of the Stoics and Aristotle. Unknown at this time, and few prospects for computation. Determinism is not very supportable, especially at full detail.
And yet the proposition "There was a sea battle at Jutland during World War II" is answerable. And all will agree on that answer.
The future is not knowable, the past is not disputable.
This arises with quantum measurements of course. Once a measurement is made--path of a photon, for example--all honest observers will report exactly the same thing. There simply is no basis for disputing the past, for Alice to say "I saw the photon travel through the left slit" but for Bob to say "I saw it travel through the right slit."
(If I am wrong on this, please correct me ASAP!)
Honest observers will report the same outcomes of measurements, whether those measurements are of photons in slits or sea battles.
This shows up in the formalism of lattices, especially the orthomodular lattices of quantum mechanics.
Of course, in quantum mechanics it's not even clear that we can talk about the "present state of the universe" as if it's a well-defined entity, in which case it may not make sense to ask whether "the" present has multiple pasts or multiple futures. A MWI advocate would say we could talk about the present state of the universal wavefunction, but that's different from the present state of an individual "world"--I believe there's a fair amount of controversy about what people even mean by "worlds" in the MWI. With a hidden variables interpretation of QM you can talk about the universe's present state, but the exact details of the present state would always be unknowable.From an instrumentalist point of view, the state of an experiment (or of the universe) is recorded by the set of measurements made.
I'm not saying things are not weird...QM by any interpretation gives results not in accord with our "realist" intuition. (E.g., the quantum box with an animal which may have one door opened to reveal whether it's a dog or a cat or the other door opened to reveal whether its a male or a female, but never both at the same time and never even sequentially so that one first opens one door, sees a dog, opens the second door, sees a male. In the ordinary world, both doors are openable, either simultaneously or sequentially, and things are as we would expect them to be. Not so with quantum things, modulo entanglement decoherence, etc.)
Even at the "present" the universe will not have a "defined past," as delayed choice experiments show. But I would argue that those delayed choice "in limbo" states are not part of the causal past, at least not yet. That dog or cat, male or female, is not yet part of our causal past (in fact, were an outcome to interact with other lattice events in any way at all, this would constitute a "measurement").
Rereading my paragraphs, maybe they are unclear. It takes entire chapters of books (I like David Albert's book, or Smolin's "Life of the Cosmos" (from whence the cat and dog example was taken), Bub, Hughes, and Barrett) to talk about these things, so my paragraphs are doing the ideas justice.
What I'll say is that there is indeed an asymmetry in what we "know" (have measured, recorded, seen) about the past versus what we don't know about the future. I'm not a supporter of "block time," a la Huw Price and others.