Right now, this second, an old man is exhaling his last breath. Elsewhere, two young lovers exchange their first kiss. Farther afield, two asteroids silently collide. Sunrise comes to a planet orbiting a neighboring star. This very second, a supernova detonates in a faraway galaxy.
And yet ‘this very second’ across the universe apparently does not really exist! Our best fundamental theory of space-time, Einstein’s Relativity, expressly precludes a single, objective definition of simultaneity. Events occurring ‘now’ by one observer’s estimation can — with equal validity — be said to occur at different times according to another observer who is far away and/or in motion relative to the first. We don’t notice this issue much here on Earth, but it becomes very obvious for example in cosmology, where how one defines ‘now’ can determine whether the universe looks uniform or not, and even if it is finite or infinite!
But wait. Of course ‘now’ is real: Why aren’t there any dinosaurs? Why has the Roman Empire fallen, but the planet around 51 Pegasi not yet been colonized? Why has the iPhone been invented but not yet the iPortal? Why have you read the last paragraph, but not yet the next? Whether or not these events have already happened may be ambiguous to some being in a faraway galaxy, but that being can’t observe these events. For those of us who can, the demarcation is clear.
Then again, is it? Einstein came to believe that passage of time is fundamentally a fiction: like frames of a cosmic movie, events are all laid out throughout spacetime, and it is only our limited nature that requires us to perceive them sequentially. The relativity of simultaneity formed part of that conviction, but it went beyond that. General relativity shares with Newtonian mechanics, Maxwell’s electrodynamics and many other physical theories the feature of mathematical determinism, or ‘unitarity’: given a complete description of a system at one time (however defined), the system is uniquely specified at all earlier and later times. It’s as if one could take any frame of the cosmic movie and by choosing a particular chemical to add, reveal any other chosen frame of the movie. If all the frames are hiding in each one, how can we say that the past or future is any less real than the present? As Einstein put it to a friend who had recently lost a loved one, “People like us, who believe in physics, know that the distinction between past, present, and future is only a stubbornly persistent illusion”. Why mourn someone who lives?
That seems crazy, though, even if Einstein said it. We remember the past, but can’t affect it; we can try to predict and affect the future, but can’t recall it. Physical laws distinguish past from future. In particular, the Second Law of Thermodynamics, one of the most basic, says that entropy — a measure of how ‘generic’ the state of a system appears — increases toward the future, not the past. The universe thus has a clear ‘arrow of time’ built in that points in the direction of time increase.
And yet this arrow of time might be explainable using unitary laws of physics brooking no such distinction, given suitable boundary conditions for the universe. If these conditions specify that there is a time at which the world has much lower entropy than it might, then it is arguably natural for entropy to increase away from this time, which can then be called ‘the past’. This thermodynamic arrow of time might then be used to explain others, such as why we remember the past but not the future.
We can question, though, whether ‘unitary’ really describes all the laws of physics. Imagine a physical system with a property P — ‘color’ for example — with two distinguishable alternative possible values A and A’. Quantum Mechanics, our core theory of how matter behaves, holds that the system can be put in a combined ‘superposed quantum state’ that has both values A and A’. What if we measure P with some apparatus? Theoretically, the apparatus itself can be ascribed a quantum state, and when it ‘measures’ the system, the system+apparatus combined state unitarily evolves into superposition of apparatuses, one having measured A and one having measured A’. A phenomenon known as quantum ‘decoherence’, however, ensures that this apparatus-level superposition behaves as two distinct classical systems, showing no indication of quantum behavior connecting them. It also indicates that the reverse process — two macroscopically distinct systems combining into one — essentially never happens, for reasons closely connected to the fact that entropy essentially never decreases.
While the mathematics points to A and A’, though, what we observe is A or A’ — that’s part and parcel of the apparatus behaving classically. How to reconcile the ‘and’ with the ‘or’ is a fascinating question over which much ink has been spilled. But in relation to the question of time, many proposed reconciliations have a similar feature: while there may be description in which the World evolves unitarily, at the level of observed reality it contained multitudinous forks and branches.
When we step back, we thus seem to have two rather different and contrary views of time’s nature. In one, the ‘Unitary Block’, spacetime and quantum states are laid out ‘all at once’, specified once and for all by some set of boundary conditions. Everything at any time is uniquely determined by — and thus implicitly contained in — any other time, and the world exhibits no distinction between past and future. At the same time, the ‘Experienced World’ we actually inhabit and observe has a very clear distinction between past, present, and future, produces entropy, and allows branching between a single present reality and several possible future realities.
Among knowledgeable and thoughtful people, there seem to be three basic views of this paradox:
- The Unitary Block is the fundamental, and by implication more true description; things such as the arrow of time, definite experimental outcomes, etc., are emergent phenomena that, if we only could make precise enough computations, could be reduced to ‘nothing but’ the fundamental description.
- The Unitary Block is wrong in some essential way; a more correct view would be much more like — and much more readily reconciled with — the Experienced World.
- The Experienced World is more fundamental than the Unitary Block, which is just the correct description of regularities in the Experienced World in very particular regimes.
View 1 is by far the most common amongst my theoretical physicist colleagues, but I’ll make three arguments as to why we should think carefully before embracing it.
The first is that future or past events in Experienced Reality cannot, even in principle, be exactly predicted using any amount of knowledge and precision of the evolution laws and state that comprise the Unitary Block. The ‘branching’ behavior of quantum mechanics is one clear reason, but it goes beyond this. One might argue that we can precisely compute probabilities for events in Experienced Reality. But we can’t. It is a theorem of quantum mechanics that even a hypothetical super-being in a lab cannot determine the lab’s exact quantum state, without dramatically affecting the lab. Even allowing for such alterations, I suspect that it would also be impossible for the super-being to determine its own quantum state precisely. Moreover, even leaving aside quantum mechanics, there are rigorous arguments in Relativity showing that an observer at some spacetime point cannot gather enough information from its past to exactly predict any event in its future. Finally, it has been shown using the theory of computations that if a physical system contains a decision-making apparatus (or being), there are decisions that cannot in principle be simulated or predicted better than letting the system actually work out the decision itself.
Second, a theory F is generally described as more fundamental than an ‘emergent’ theory E if E could not have been any different without F being different, whilst E might have in principle emerged from several different theories F, F’, F’’. That is, there is a many-to-one mapping between Fs and Es. But that seems an inaccurate way to describe the one-to-many relation between the Unitary Block and the Experienced World that exists in terms of, for example, time-direction or quantum experimental outcomes. The laws of physics may even have this relation: string theory appears to comprise a single high-energy theory that has many, many very different low-energy effective theories identified with it. This one-to-many relation means that Experienced Reality contains information that is lacking from the Unitary Block. Consider the ‘now’ with which we began this essay. Consider using a Relativistic Block description of the universe to predict tomorrow’s weather. Without something like ‘now’ or ‘here’ you simply can’t get anywhere. Elements of reality like ‘here’ and ‘now’ and (perhaps) ‘in this branch of the quantum state’ may be called ‘indexical information’ and are central to doing science, but have no place in a Relativistic Block.
A third claim takes this further: in addition to indexical information, there are important and meaningful things that exist in Experienced Reality but are essentially meaningless in the Unitary Block, making the latter too impoverished to be considered a true depiction of reality by any reasonable definition. Consider circles. They exist as defined mathematically, as imagined by people, as approximated by round objects, and all in all sorts of other ways. But there are no circles in a quantum state or a unitary evolution rule.
Imagine, for example, presenting our hypothetical super-being a large box containing a computer, plugged into which is a quantum random number generation card. At time T, on the basis of a random number so generated, a complex algorithm in the computer draws a circle or square on the screen. Contemplating the circle or square, a person decides whether to have Indian or Chinese food for lunch. Can the super-being predict the cuisine chosen? No way! It can’t get the quantum state; even if it had it, it can’t predict a particular number from the card. Nor can it think both in terms of circles or squares, let alone transistors, microchips, programs, etc. and also in terms of the quantum state; nor, even thinking in these terms, can it necessarily simulate the prediction. There seems no meaningful sense, then, in which the decision ‘pre-exists’ in the state of the universe at time T. They simply don’t exist on the same plane of reality.
What does this tell us? To me, our understanding of time is a particularly vivid demonstration of how we can usefully describe the same World in multiple ways that are rather starkly different in character and in import for such questions as free will and agency, universe or multiverse, epistemology, and others. The mode of theoretical physics has tended toward elevating certain descriptions as more ‘true’ and ‘fundamental’, and the other descriptions as ‘in principle derivable’ from these. I think it is worth contemplating a different approach: that just as the Experienced Reality description is far more useful than the ‘Unitary Block’ in some applications (and far less in others), that we take Experienced Reality as an equally true description that is complementary to, but not derivative of, the Unitary Block. Perhaps the nature of time is telling us something about the nature of truth.
- If the speed of light were much slower, so we could really experience the subjectivity of simultaneity implied by Relativity, how to you think it would change our experience of time, space, and the world?
- If you truly believed that the past, future, and present all exist in just the same way, as in the Unitary Block view, would it change your attitude toward life — or death, or decisions?
- If there are aspects of the world that are unpredictable in principle, is there still a sense in which we can say that they are determined? Or do those two ideas go hand-in-hand?
- I’ve suggested that part of the controversy over how to think about quantum mechanics is a controversy over whether the classical, macroscopic world emerges from the quantum world, or whether the quantum world is a particular, stripped-down limit of an essentially classical world. How do you see it? What theoretical or experimental findings might lead you to accept one view as more viable than the other?
I’m again grateful to BQO for the invitation to contribute an essay and participate in the discussion. I enjoyed and learned from both. The online discussion brought up a number of issues. Some were essentially technical scientific questions, which I attempt to address — except that I declined to get into a debate about the validity of Einstein’s Special Relativity. Others raised a number of interesting quasi-philosophical issues.
A key point I tried to make, which many discussants appeared sympathetic to, is that the ‘Unitary Block’ description is not just difficult-to-work with when trying to describe all of reality, but actually incomplete. It must be augmented in order to function, and we (by which I really mean theoretical physicists) often neglect or minimize the important of what we must add. When you read, for example, that a person is ‘nothing but’ an arrangement of atoms, this gives the unfortunate (and perhaps unintended) impression that the arrangement has less ‘reality’ to it than the atoms themselves. But the ‘arrangement’ is everything! And we know from quantum mechanics that the atoms are far less material and ‘real’ than our intuitions like to make them.
Physics is, unfortunately, rather ill-equipped to describe this information in a useful way. The concept of ‘Shannon Information’ is wonderful in many contexts, but in the Unitary Block it is neither created nor destroyed. This is useful (just as the conservation of energy is useful) but in mechanics or quantum mechanics the conservation of energy is related to the general time evolution of a system, so energy conservation tells you a lot about how the system evolves. Information theory is not presently able to do something similar in terms of capturing the crucial difference between, say, the arrangement of atoms that is a person, versus another special but nonliving arrangement.
One nice question in the discussion brought this to light: is Hamlet ‘implicit’ in the initial state of the universe — does it ‘exist’, in some sense, in the initial state. A second question hit parallel note: do things like ‘stars’ exist in a universe in which there is no observer, or consciousness, to perceive them. My answer to both questions was ‘Yes’…and ‘No.’ There is a strong case for ‘yes’ in both. In the first, if the state of the Universe now can be directly 1-to-1 mapped to that at some early time, and Hamlet exists now, how can Hamlet not ‘implicitly’ exist at the beginning, albeit in perhaps scrambled or encoded form. In the second, it seems absurd to think something ‘objective’ like a star could depend on being perceived.
Yet in both cases I assert that we bring a tremendous amount with us when we say that Hamlet or a star exist, that we tend to neglect what we have brought, and that it’s unclear that these objects really exist without it. I think it would be very interesting to devote more thought to these questions. Phrased a bit more generally:
New Big Questions
1) What theoretical tools could be developed that could allow us to recognize and distinguish an object (such as a Piano) relevant to Experienced Reality and developed by a meaningful historical process, from a different meaningless arrangement of the same atoms, which may even contain more Shannon information (lower entropy).
2) What is the ontological status of ‘arrangements’? How do we pick something out of a wavefunction, say? Is it always in relation to an observer, and if so, what type of observer?