Carlo Rovelli on challenging our common-sense notion of time
nautil.us
nautil.us
I get it that it is convenient to talk of change in something, like movement from state a to state b, under the guise of wrapping it in "time", but is there some physical argument in support of time in general? Honest question.
Living beings getting older is their cells becoming more and more inefficient in division and cell repair. One could likely achieve the same effects with chemicals, but doing so would not mean time has run faster.
The concept of time for humans seems to be all about observable change, which needs an observer with a memory to compare the current state with previous state to be able to say: time has passed.
A rock changes via erosion and such, but it has no memory, and cannot observe anything. Does a rock feel time? Of course not. Does it exist "in time"? Does it, without an observer that somehow measures the flow of time (via changes in Cesium atoms or something)? Or is the rock just existing and under the whims of all forces of nature that might impact it and change it into smaller pieces and eventually to sand, and so on.
I guess my question is: what exactly is time, physically, and why should it have to exist as some sort of a physical process in the first place.
I realize you might mean you're not sure how they're relevant to the topic at hand, rather than how they're relevant to each other. I'll attempt to illustrate this connection by providing the simplest possible internally-consistent model of an abstract system that appears to resemble reality as we experience it, including the passage of time. I will define the base components of the system, and compare the emergent interactions of these components with the phenomena we experience.
First, we observe that any given physical space can be said to represent some positive quantity of information; a physical space can thus represent the "state" of a system.
Second, we assume the existence of "causality", patterns for recognizing and amending portions of a state. We may then refer to a possible application of these patterns to a portion of a given state, which we'll call an "event".
For every pairing of initial state and set of possible events, there exists a space of computable states that we may call a "totality". A totality can be represented as a directed graph. The root of this graph is the initial state, edges are detected events from which new states can be computed, and child nodes represent states transformed by events.
Now, when we talk about "the passage of time", we mean the iterative computation of successive levels of totality. An event's distance from the root node is its chronological order, and thus a "point in time" is a depth in the graph of totality. The full state of the system at a given point in time is then naturally the union of the state nodes at that depth.
Our experience of the passage of time is represented similarly in the brain, with neurons building complex models by generating successive branching layers. We hinge knowledge on key experiences, and we perceive time fractally, in heirarchical levels of detail. It seems intuitive that our brains, which function by reflecting our outward reality internally, would in fact structurally reflect reality. Furthermore, the conclusion that the causality graph becomes more dense as mass in a space grows more dense seems to reflect general relativity.
Here's where Gödel and Turing come in. A totality is defined iteratively: there is no way to know its complete value except to evaluate its events, and the only thing to do with it is evaluate its events according to its patterns of causality. Similarly, given a Turing machine (or a consistent system of axioms), there's no way to know the final state (or final set of conclusions) except to run it with some input, and the only thing it can do is run.
So back to the question:
> what is time, physically, and why should it have to exist as some sort of a physical process in the first place
Time, physically, is the continuous evaluation of possible states of a physical space (like ours). It has to exist because it is, by definition (to the point of tautology), the only thing that can possibly happen. It is still a "mathematical construct", but only to the degree that we also are ¯\_(ツ)_/¯
Ultimately as with most physical phenomena the most profound explanation is that "It's a simple way to describe how things work". That and anthropic necessities.
Relationships between events are almost certainly an anthropic necessity. First without change at all, of course there would be nothing of interest. Second an universe that's simply a random-looking collection of events {A,B,C,...} without some close functional relationship (e.g. {A,B=f(A),C=f(B),...}) also cannot allow a conscious subsystem (because conscious subsystems are defined as a particular kind of changing system, i.e. with close functional connections between the events that temporally evolve). There's an infinitude of mathematical constructions that may harbor consciousness, so some things are arbitrary.
Our arbitrariness is that there appears to be this continuous dependant real variable (time) which permits localizing events in a complicated way described by general relativity and quantum mechanics, and describing relationships between them (temporal evolution).
Because time -is- a location? All things move down this one-way (or so we assume) tunnel, and time is their location in that tunnel.
Imagine a paused video player where everything moves back and forth only when you move the playback slider/scrubber, or a lenticular photo that changes when you move your head.
You have to consider the human ability, tendecy and limitation of perceiving everything in 3 dimension and in terms of distinct points, particles or “snapshots”, as opposed to perpetual streams of potentially many more dimensions, each with different “directions” through which they may be navigated.
We can only see and intuitively reason about a very filtered slice of reality.
And this is why we need to find and meet other intelligent beings who perceive reality in fundamentally different ways, to expand our own understanding of it. :)
In those cases, it might be possible for someone (e.g., a hypothetical primitive person) to notice that the regularities are related to one another, e.g., there are 28 solar cycles for every lunar cycle, and so many lunar cycles between harvests. It might occur to someone that the solar cycle is a useful way to measure something, even if it's not clear exactly what the "something" is. Knowing when to plant and harvest, or when to expect certain kinds of animals to be plentiful for hunting, might be pretty useful.
I tend to think about time in terms of the observable relationships between different forms of periodic behavior. This is at least easier for me to comprehend than problems such as memory and gradual change.
If you want an example of humans able to determine time incredibly accurately over a lifetime, just look a "perfect pitch" musicians who can tune their instruments to match other musicians to a percent from across the world. Although most people would have difficulty matching that accuracy, they wouldn't have difficulty telling an alto from a soprano at any point in their life after they have first heard the difference.
Our civilization grew up with music, dance and all sorts of "absolute" human temporal measures.
The problem with your theory is that the main idea of Special Relativity is that time is a good tangible dimension, that has (almost) the same properties than the other three dimensions. And this is fully backed by experiments, and the effects are measurable in satellites, atomic clocks on planes, the color of gold, and many many many additional experiments.
One important property of the usual dimensions (x, y, z) is that you can mix them. Let's say that we choose z to be the vertical direction. Now we can choose x to be pointing to the east and y to the north. But we can mix x and y
x' = x * 1/ sqrt(2) - y * 1/ sqrt(2)
y' = x * 1/ sqrt(2) + y * 1/ sqrt(2)
Now, it is not a good idea to imagine that the east-west axis is a tangible dimension, but the north-south axis is something else. The main problem is that someone else can choose other directions, like (north-east)-(south-west) and (north-west)-(south-east) and get the same experimental results. Now, which one is the tangible dimension (north-east)-(south-west) or (north-west)-(south-east)? So we assume that all coordinates x, y and z are essentially the same thing and no one of them is special.
In most common situations, the vertical axis is different of the two horizontal axis. In normal situations, it is clear where it is up and down, so the "z" axis is special. It is not clear the direction of the "x" axis, and for most experiments you simply choose the direction that makes the calculations easier, but "x", "y" or any horizontal direction is as good as the other. [Just ignore again that the Earth is spinning and that it has a magnetic field. The spinning of the Earth and the magnetic field define two special directions that are both the "north" in some sense.]
--
Now, even if there Earth is not spinning, and if the Earth had no magnetic field, the problem with the "z" direction going up is that two persons in different continents will disagree, about where is up and down. If you choose (x,y,z=up) in other continent will choose (x', y', z'=up') where x', y' and z' are obtained mixing x, y and z. And both will be totally convinced that "up" is special and there is a clear meaning of "up", in spite each one has a different "up". So it's better to assume that the universe has no preferred directions (x,y,z) and all the differences are due to the details, like a big chunk of dirt we call Earth.
If we assume that some of the directions x,y,z have some essential properties that the other doesn't have, it would not be possible to mix them and construct x', y' and z'.
--
Now, about the problem with time ...
In Special Relativity you can mix the special coordinates and time. It's easier if you always multiply the time by c (the speed of light), and you use the coordinates (ct, x, y, z). Ad it's more simpler if you think that you measure the distance in x, y, z in light-years and that you measure the time intervals in t in years, so you get the same number in ct.
In Special relativity you can mix for example x with ct
ct' = x * ? + ct * ?
x' = x * ? + ct * ?
The important part is that the formulas are mixing x and ct. In one mix you get ct' that is a new time-like coordinate and in the other mix you get x' that is a space-like coordinate. These are the coordinates that someone else sees when is moving at a different speed than you. You and the other person (in a train, plane, spaceship) will disagree about what how the time flows. You see t and the other person will see t' that is a mix of your t and your x. But both you and the other person can do any experiment and get equivalent results, because the universe don't prefer t to t' or vice versa. There is no experiment to determine if t or t' is better.
The main difference between the example with (x,y) and (ct,x) is that x and y are mixed in a slightly different way than x and ct. In particular, you can exchange x and y (or better x and -y for technical reasons). But you can't exchange completely x and ct. All the mixes have a new ct' that is somehow more similar to ct than to x, and a new x' that is somehow more similar to x than to ct. Moreover, all the observers agree that there is one time and three special dimensions, but they will not agree about how the time flows (as the will not agree where the directions x, y and z are pointing to).
The explanation of why can't mix "completely" x and ct is part of the mathematical details that I'm not writing here. It's somewhat related the fact that you can't move faster than the speed of light. Just get any introductory book about Special Relativity, but continue reading until you reach the chapter about Minkowski spaces. The initial formulas doesn't make too much sense until you reach Minkowski spaces.
Emphasis mine.
I'd never heard of this before, so I did a quick search and thought this might be interesting to other people as well:
https://phys.org/news/2015-10-einstein-theory-relativity-fun...
I didn't understand why gold was similar to copper, while silver (between the two on the periodic table) is not. Does it have to do with how electron shells are filled? Definitely beyond my level of chemistry :-)
While this is true it accidentally implies that that was what Einstein invented in his Special Relativity paper. In fact this was an idea of Hendrik Lorentz, who (among many others) had also already been working on the question of why the rate of time is different in different frames for more than 15 years before Einstein figured out how to explain it.
By the way, jnurmine, special relativity isn't super complicated and is quite approachable. No calculus, just straightforward algebra. (sadly this statement is very much not true for General Relativity). I highly recommend "Spacetime Physics" by Taylor and Wheeler, which is readable, full of pictures, and yet written by two amazing physicists. And as I just discovered it's even on the Internet Archive!
In 1D, you only need one number.
In 2D, two.
In 3D, three.
But in all these scenarios, the point in question is "static". What if the position of the object was also changing in time?
Then, it wouldn't be sufficient to provide just the spatial coordinates, you'd also need to specify what time it is. In 1D for example: at 12:00 the point was at 0. At 13:00, the point was at 1.0. At 14:00 the point was at 2.0.
Thus, "4D" in special relativity comes from 3 spatial dimensions + time, since we care about how things move through "space and time".
Our definition of time seems therefore to be two things: first, a snapshot of physical state (in which we imagine points in time a bit like Back to the Future), and secondly (as other comments have noted) a measure of periodicity.
The latter is that which we seem to be measuring experimentally (e.g. with differences in atomic clocks within gravitational fields). However what we are literally measuring seems to be something like “how many times can a given particle move/irradiate relative to another given particle before we bring them back to each other”. Reducing to a pendulum for example, we could imagine that the space was “more dense” so it took longer to travel the same apparent distance, a bit like how light takes longer to travel through particular materials. Another way to see this would be a kind of universal framerate.
So based on this possibly weak conceptual understanding, I can certainly imagine a fourth property of any given spacetime location that reflects how frequently events can occur in that space (a kind of EM-field view of time), but to consider this as a geometric “position” or coordinate is something I conceptually struggle to imagine.
See the Andromeda paradox: https://en.wikipedia.org/wiki/Rietdijk%E2%80%93Putnam_argume...
There is no absolute simultaneity in relativity. So the Andromeda Paradox only becomes a problem if you invent a magical FTL chronovisor that allows you to bypass causality and see what's happening anywhere in the universe at your idea of the present time.
There's no reason to believe such a device could ever exist. So I don't see why there's any confusion.
Both hypothetical observers discover events that happened in the past when the light cone of the events intersects their position in space time. Events in the past aren't uncertain, they're simply unknown in the reference frames of the observers.
And because the observers have no causal influence on distant events, they have to watch them unfold passively.
It's no different to an incredibly laggy online game. If you're well behind the action, events happen before you can influence them. The fact that someone may suffer twice the lag you do doesn't make the events any more philosophically ambiguous, mysterious, or acausal.
Just to take a silly example, what if all the "points in time" just exist (and are ordered -- I don't propose to break causality)? They don't flow at all. From my perspective, at every "point" along the time axis, I can recall the past and it will be ordered as if it "happened", but each point could be independent (though constant). If I could remember "forward" through time, then this would be obvious, but since I can only remember "backward" through time, at every point it appears as if I have progressed through time.
I suppose the interesting thing is that causality is uni-directional. Things can only happen in a certain order in the past. But this is not true of the future. Even if I have perfect knowledge of the present, there are some things I can not predict about the future (quantum mechanics FTW). I wonder why that is (because we are flowing through time? Ha ha!)
Sorry for the diversion, but if someone that is better educated than me could shed some light on the matter, I'd be grateful.
- https://rationalwiki.org/wiki/Timeless_physics
- https://www.lesswrong.com/posts/rrW7yf42vQYDf8AcH/timeless-p...
The theory that tries to explain why we nevertheless only remember the past is called thermodynamics. It has to do with how the universe started out in a very orderly state. If you start the simulation in an orderly state it's likely to become less orderly as you run the simulation. Whether you run time forward or backward doesn't matter: in both directions it will get less orderly. According to thermodynamics that's why we remember the past and not the future. If god had started the simulation with time running backward we'd still have the same experience: we experience the past as whatever direction is toward the more orderly state.
CP violation is an experimental fact since 1964 [1] and by the CPT theorem, this implies that time reversal invariance is broken [2].
This sounds similar to the "block universe", where the universe is a 4D "block" with a certain height, width, depth and duration; where change over time (e.g. an object fading from black to white) is just like change over space (e.g. an object with a colour gradient)
> since I can only remember "backward" through time, at every point it appears as if I have progressed through time.
I've seen a related idea come up in discussions of "Boltzmann brains" (can't find a reference ATM): if the 'arrow of time' is due to increasing entropy, then regions where entropy is decreasing (e.g. an open system radiating heat) could be thought of as experiencing time backwards. What we see from the outside as, say, a collection of photons entering a camera lens and bleaching an existing image off the film; from the inside would be experienced backwards as the taking of a photograph.
> Even if I have perfect knowledge of the present, there are some things I can not predict about the future
That might also be true about the past! There's an unsolved problem in physics called the "black hole information paradox", which points out that if we have a bunch of mass (like a star, or a vast number of elephants floating in space) we might calculate that in the future they'll collapse down into a black hole. Yet if we have a black hole, there's no way to know what it was formed from in the past (stars, elephants, etc.). Black hole formation appears to lose information (like an AND gate), but that's strange since all known microscopic laws of physics (i.e. not the second law of thermodynamics) are reversible.
(Please keep in mind, most or all the words following are incorrect, in a way reminiscent of recounting the history of earth by saying: "A fish walked out of the water and became the emperor of China.")
We can imagine our local universe is a thing like a multidimensional boiling bubble. It recursively sub-divides itself, which makes it interesting, and also creates space-time. The start of our local universe's time is the so-called "big bang", which is when it started this wild and lovely fractal. After a little bit there was enough complexity to implement something like the physics we can observe now.
It turns out, a lovely universe can be created from just recursive subdivision. We're smart enough now that we've defined computability in terms of this, for example S K combinitors, etc. Math and physics will eventually catch on. (The astute reader will notice there are things we can imagine, that are uncomputable, math that is un-representable, and physics that won't work with this system. Which isn't really a problem yet.) Thankfully, this configuration happens to be very interesting.
But wait. How did time start? Well, um, it didn't. What we note as time is the constant, change (adding new nodes), expansion, and complexification of recursive subdivision, which happens in units of the "things" dividing, which is like the quantization that we observe in time, matter, and energy. It's "frequency" is also somewhat dependent on it's local environment, since it's "pulled on" by what attached and intertwined with, exhibiting the relativistic effects of gravity and motion.
Perhaps you've heard of the universe described as the three dimensional surface of a four dimensional sphere. But what's inside that sphere? Actually it's the past. Sitting there right now, for real, probably just as it was, fully connected. Of course we're zooming away from it a varying rates, but pretty fast, with the "force" of the universal expansion behind it.
So if the past is actually sitting right there, not that far away from us, we could just go to it right? Well, maybe. It would take a lot of "energy", and stretch the "universal surface" out of whack. But perhaps one could. But what about those paradoxes? Date your mom and your hand starts fading away? Accidentally kill your grandpa and poof out of existence, or not? Well, thankfully those paradoxes can't happen, because even if you're altering the physical past, you still lived through the past you lived through. The altered future that it creates will be "forever" behind your present. Another way to think about it is: any time you effect the past of this universe, it pushes it slightly in the direction of another dimension, in which there is a whole "copy" (or copy-on-write) of the universe with just that thing altered (so far).
This still doesn't answer why we don't bump into a lot of time travelers. Are we doing this the first time through? Or is it just really expensive, boring and useless to muck about in the past? Or are none of neighbors from the universes next door interested? This presupposes that you can re-run through the time dimension. Only at the edge are we "moving", or are new nodes being added. You would have to fork a whole other universe starting at that time slice, which is probably unfeasible unless you have "massive" extra-"universal" compute power. Is it really that important? Especially given that due to the fractal nature of nature, you can fork another sub-universe in this universe! Much easier. Or even better, just take the tiny part you want and copy and paste. Done. Or just get it right the first time. Come on now people.
Of course the future doesn't exist. Which makes it fun when we can seem to predict it using such things as Newton's old equations, and speculative execution.
Basically, the flow of events in time is computed recursively by applying the laws of causality, and we experience the passage of time through continual iterative changes applied to the portion of the spacetime event causality tree that represents us.
The x,y,z directions are connected to each other, because the x direction can be rotated into the y direction by rotating in the xy-plane. What special relativity is about is that the t direction can be rotated into the x direction by "rotating" in the tx-plane. Rotating in the tx-plane is acceleration in the x direction. That's what causes all the weird relativistic effects such as time dilation and length contraction when you accelerate objects. Creatures that don't intuitively understand that the x and y directions form a whole could come up with a similar theory: if you rotate an object in the xy-plane it gets x-dilated and y-contracted. The only mathematical difference between a rotation in the xy-plane and a rotation in the tx-plane is a minus sign in the formula for rotation. So the t direction is not exactly like the x,y,z directions, but the only difference is that terms involving t^2 get a minus sign. For example, the squared distance in 3D is x^2 + y^2 + z^2, but the squared "distance" in the weird 4D spacetime geometry is x^2 + y^2 + z^2 - t^2.
https://en.m.wikipedia.org/wiki/The_Unreality_of_Time
http://edwardfeser.blogspot.com/2018/05/godel-and-unreality-...
It's rarely discussed because the implications are more philosophical.
The math, and expiriments are very useful when time is a variable. So is daily life, clocks are very useful.
This is the bit that irks me. Quantum mechanics, the real QM that physicists actually use, involves collapse of the wavefunction. This is absolutely time-asymmetric. But in all these discussions of "why is time one-way?" this never seems to be mentioned. Apparently QM is not a real theory, and just a placeholder until we can work something out properly. Irk!
I don't think so.
> "piles of words"
I definitely agree about the piles. Maybe ordinary language is just to feeble for this stuff.
If you could run an experiment where you account for the interaction between the apparatus and the observer, a meta experiment if you will, you would not see the wave function collapse that the observer thinks they observed.
The basic problem with your suggestion is that it does not explain how anything actually "happens". Unitary evolution is a theory where nothing changes. People seem to think that they can evoke decoherence, or similar, and these issues disappear. But they do not.
Anyway, I'm not positing any solution, I am just saying that it irks me that all these ideas are dismissed whenever it comes to the problem of the asymmetry of time. It just seems bizarre to me.
<u|U^\dag U|v> = <u|v>.
OTOH, I can understand the argument that nothing ever happens. One reasonable requirement for something to happen is that it should not be able to "unhappen," which is why some are looking for the possibility of irreversible decoherence.
This is not unitary. Quantum theory has two parts: unitary evolution, and measurement processes. That's it.
Let u and v be qubits. Apply a Hadamard gate to u while leaving v unchanged. Then they have evolved independently (but unitarily): u according to H, and v according to I.
There is no QM theory. There are interpretations (like the Copenhagen interpretation). We don't know the truth as of yet. (which is exactly what Hawking has been trying: creating a unified physics between QM and General relativity).
I don't understand the rest of your question.
A crystalline solid has relatively low entropy because you can be reasonably sure where each nucleus is, as they are highly ordered. A gas or plasma can have the nuclei distributed nearly randomly. Both systems can be measured generally (stochastically, or of their overall or average properties) but the number of ways you can organize all the nuclei to get those results very different between the two systems.
Feel free for someone more knowledgeable to correct or expand.
Given that the gravity must have been immerse then how should we understand this time?
We don't really know. There are speculative theories, but they don't all agree.
It's very difficult because the temperature would have been so hot at the very start that instead of four fundamental forces (of which gravity is one), we would have had only one. There's also the Hartle-Hawking State, which is a theory that proposes at the very "start" there simply was not time, only space, and that time coming into existence was part of the process that occurred during the planck epoch.
There's also the Grand unification epoch, inflationary epoch, electroweak epoch, quark epoch, and hadron epoch that occurred in the universe before a second had passed, all with very different implications for how physics could have worked, so the answer would change multiple times during that period.
I enjoyed the presentation. Talking for an hour on this kind of material with just a piece of string and a couple of watches is fairly impressive. The thought experiment with the wooden box and green/red balls (in two size categories unrelated to their colour) was thought provoking.
I'd like to see someone define/construct two sets of macrostates that give a different time scales in a system with one set of microstates. For all I know someone may have actually done that. The paper referenced does have a section on the 'thermal time hypothesis'.
They shouldn't, if they are properly made. They should tick and measure at their usual rate. Their local timeframe is out of whack with somewhere else, but that's got nothing to do with the clocks.
It's just one of my peeves, "clocks slowing down in hight gravity". Nonsense, it's time itself which "slows down".
For anyone that is dubious: Black holes are one of the common ways this gets discussed, because it's such an extreme example, and you can read more about it at http://www1.phys.vt.edu/~jhs/faq/blackholes.html#q11 - this goes into a bit about photons and the event horizon - you can ignore that and just consider it in regards to the time dilation effects of gravity.