Causality ain't what you think it is

Do you know what causality is?

If you do, let me know, because I’m not sure.

I’ve never come across a conception of causality that makes sense to me.

After all, our universe seems to follow simple equations like Einstein’s equations, and there’s no mention of causality in these equations.

It makes me think that there’s no such thing as causality.

Unless...

Well, here’s the thing.

I’m no longer sure that our universe does follow these continuous equations.

I’m beginning to think that at the smallest scale, our universe might evolve through discrete computations.

If that turns out to be true, it allows for a limited conception of causality after all.

It’s causality, Jim, but not as we know it.

—

References:

Images:

—

Kootenay Village Ventures Inc.

[ludinodreamsmith]

Thank you for making The Last Theory series, and this video in particular.

The main problem I have with what you say in this episode is that I think of computation itself as a causal process.
Computers are machines made specifically for ensuring that causality holds, and that you can control and manipulate discrete causal connections. Analog, continuous and basically untrustworthy components are engineered to behave like discrete, completely deterministic systems. As long as the system is in place, you can think of computation independently of the physical mechanism that instantiates it. I think the same goes for causality.

Think of it this way: an instance of causation is an event in which one or more causes produces an effect. A simple model of this would be where one or two two-valued causes produce a two-valued effect – or put differently: it takes one or two binary inputs and produces one binary output. In a universe with only one such event, the boolean operators, NOT, AND, OR and so on, together would constitute the complete ruliad for this universe (if I understand the term correctly). There are no other possibilities. I guess this also means that any possible discrete causal relation can be modeled by combinations of these rules.

Of course, if computation is a causal process, it makes no sense to say that computation can replace causality as an explanation or model of the universe – or indeed be an explanation for causality itself. In this view, computation and causality is fundamentally the same thing.

If causality is to be used as a fundamental explanatory principle, then it must itself remain unexplained, and it makes no sense to ask what causes causality or what mechanism instantiates it. In this sense, causality actually is “only one thing after another” (according to some rule).

Where the Wolfram model says computation gives rise to physics, I would think causality can do the same work. It is causality that produces the apparently “physical” world. Rather than doubting causality, one should doubt the existence of space and objects. These are theoretical constructs, inferences the mind makes based on its causal interaction with the world (sensations and actions).

I remember the first time I recognized a boolean truth table in at text about digital electronics. I had taken some elementary course in philosophy, but could never quite digest the idea that a sentence like “The moon is made of cheese or Socrates is a man” should be considered true, or to have any truth value whatsoever. But considered as a map of causal relations instead of something to do with sentences and language, it totally made sense.

- Ludino

[jrkirby]

It's never just equations, though. We like to suppose you could get a perfect representation of the state of the universe, and then model the evolution of that state over time with equations. But we know this is untrue on many levels. For one, it's impossible to get a perfect representation of the state of even a single element of the universe, according to the Heisenberg Uncertainty principle. For two, we know that the universe is a chaotic system, so any uncertainty about the current state must in fact increase exponentially over time. We also know that subsystems that are not entangled with an observer appear to take into account every possible sequence of events (feynman diagrams), only to 'choose' one at 'random' before reentangling with the observer. The 'Many Worlds' hypothesis of quantum mechanics suggests that every possible outcome of a quantum event exist, but that we merely cannot interact with the events that contradict the event we observed.

There isn't just one future that is consistent with the observed state of the universe. There are many futures. "Free will" is a reasonable name for the process of selecting the singular future that will be experienced from the plethora of futures that were consistent with the past observations. But attributing "free will" to human beings alone, is probably a misunderstanding. We know that even the smallest subsystems of our universe seem to make "choices", so why would it be a stretch to suggest that larger subsystems, made up of these components, also make choices?

I did, when I was 22, and for many years afterward, see the universe and free will the way you describe. " 'Free will' is a useful concept to believe, but not an accurate depiction of the universe." I said. But I have since come to realize that my reality is not a well defined state that has been iterating like clockwork since the beginning of time with a single trajectory into the future. Instead, my reality consists of a finite set of observations (my personal past experience), and all the dynamics, possible past states, and possible future states that are consistent with those observations.

So causality primarily needs to be understood not within the context of any given model of the universe, but with respect to the limited observations we have on hand. What futures are still possible, given the observations I have access to? When a man is holding a ball, my observations cannot give accurate prediction on what he will do with the ball. Information about the dynamics at play inside his skull has not entered the scope of my observations in such a way that would give predictive power. But once the man throws the ball into the air, my observations imply a far more constrained set of futures for that ball's near future.

I don't see this issue as one where the wolfram hypergraph model of physics would disagree with other observations or models of physics. The many possible continuations of the hypergraph should correspond to the many possible continuations of a quantum physics model of a system. There isn't anything about the discrete nature of the hypergraph model that make causal inference have any fundamental differences with a continuous model.

[jaddaj]

Einstein’s equations are causal. If we imagine perturbing the position of the sun say by a large explosion like at the moment when the star goes supernova, the change in gravitational field will propagate at the speed of light by gravitational waves. The event of the explosion causes the change in the motion of the earth. So you can see the causality in Einstein’s equations from the linearised approximation of gravitational waves. You can also prove Causality for globally hyperbolic spacetimes.

[Acryte]

Hey, I love to watch your videos!

I've been a fan of the Wolfram Physics project ever since I read A New Kind of Science. I've taken issue with the current direction of science in creating increasingly detailed definitions of how things behave in an attempt to answer the question of first cause. Presupposing fields, vacuums, virtual particles, and other increasingly detailed definitions is a whole lot of something, and I've always had an issue with the fact that in order to explain how things come from nothing, or why maybe something is the default and nothing is what we shouldn't expect, we instead continue to increase the rigidity of the requirements for how actual stuff must behave in order to make it happen. Imposing a multiverse where every possible outcome is realized to simply address the sheer odds that all the fundamental constants of nature managed to align with the outlandish precision necessary for life to exist in this universe seems unreasonable... Postulating that one common iteration of a simple rule or set of rules yields all the complexity we see around is an exciting concept. (Not sure where the Ruliad fits into all that when considering necessary conditions for this universe. I suppose I have more reading to do).

It's encouraging to see Wolfram and Jonathan's work discover new connections between commonly observed physical laws and constants. Just as Jonathan said, there's probably nothing special about hypergraphs, it's simply emergent from any network or thing that is general enough. It's interesting to think about how we failed to see what's in front of us if we can see it in so many places. It's easier to discover the solutions when they're manifest in so many forms. So cheers to that! Thank you for all your hard work, your script writing is excellent and I appreciate the effort you put into crafting each video. Here's to many more to come!

[Chris-opyt]

free will has never been proved to exist, but whatever. lets keep to physics. people perceive a cause and effect relationship when it comes to collisions, eventhough (like you stated) there is no discrete turns taken by this or that object in formulas.
i've come to conclude that (in physics) looking at isolated events leads us to think there's causes and effects, but in reality no such isolation is possible. rather than any bit of rock in space having a turn to causally bump into another, the entire universe--also down to microscale--is in a deterministic dance all in parallel, with no single thing preceding another.
this is not useful for physics, as for every thing you're interested in would require computing all forces and all objects in whole universe, so physics makes approximations and separates things to provide useful output for particular purposes that are close enough on a short time scale.

[Sam-wezj]

this is deep i shall be pondering this video for days!

im wondering if you are getting rid of causality or are instead assigning it to one place: the computational rules for the universe. this reminds me of a passage David Deutsch has in the beginning of infinity where he says if someone asks why is there a bronze statue of Winston Churchill in parliament square and you answer by talking about elementary particles and the big bang then you've lost explanatory power.

but this feels like something deeper. like you are saying cause is a human overlay on something more fundamental. i guess where im confused is if there is no cause in the universe then why should we care about computational rules? doesn't the word rule invoke a notion of cause? or are the computations something which are 'just happening at the same time' as i decide to throw the ball.

i like your notion of causality as it applies to the hypergraph but here's what im wondering. there's not only one rule being applied after the other, there's also parallelism. as with a non-deterministic Turing machine, in the Wolfram picture there is a choice of what rule gets applied next. i think this is critical.

in a parallel setup not only does the output of any given computation matter, but the propagation of information about that output through the network matters as well. it determines what rules get applied next. Wolfram has talked about an entanglement speed, an upper bound to the flow of information in the network. we also know regions can form in the graph which slow or trap information. so called black holes where regions of the graph are temporarily disconnected. markov blankets are another example.

in a setup where information is causal can macro features of the data structure which slow or trap information be considered causal also? and do they recover my grandmother's concept of causation ha?

another great video this is one of my fav channels

[pilotnamealreadytaken]

At the end of time, this channel only uploads one video. Becomes the most watched and important channel of all time 😂❤

[greenfinmusic]

You didn't cause the ball to fly through the air, you caused your spirit to choose a path through possibility space in which you experienced your hand tossing the ball (and we experienced seeing you do it). By invoking a multiverse/many worlds ontology, we can see how each particular path can appear to have a purely physical causally closed explanation, but from our god's eye view we can see that you exercised your free will not in altering any particular path (which is impossible), but in choosing between the paths, i.e. choosing which of the infinite/near infinite 'zombie yous' branching off at each moment that you 'follow' and 'stay in tune/phase with', which allows you to experience 'what it's like' to be that zombie you in that moment.

Deep, libertarian-style free will (i.e. the kind we naturally tend to think we have) is overwhelmingly likely to exist. The almost perfect linkages between the 'goodness' or 'badness' of particular qualia of phenomenal consciousness and the 'goodness'/'badness' it would have if it were causally efficacious, is uncanny to the point of being practically impossible from a statistical perspective. For example, if phenomenal consciousness didn't have any sort of non-overdetermined causal efficacy, then there's no reason that touching a hot stove shouldn't feel fun and wonderful. But it doesn't. It feels awful. Which is strong evidence that the phenomenal consciousness of the experience (the purely mental qualia of pain and discomfort) has non-overdetermined causal powers.

Physicalism implies epiphenomenalism (and a lack of non-overdetermined causal efficaciousness), but phenomenal consciousness almost certainly has non-overdetermined causal efficaciousness, so epiphenomenalism is almost certainly false, and physicalism is thereby also false. We traditionally think of causal power as being the ability to influence physical reality, but that's impossible, since a purely mental Spirit can't interact with physical reality (since something that is purely mental has no actionable properties in common with something that is purely physical) but what we do have is something near enough: the ability to steer yourself through possibility space so that you have control over the phenomenal consciousness you experience. (and not in a hallucinatory sense, but in a very real way in which all the people around you have experiences of phenomenal consciousness that are logically consistent with each other, e.g. we all experience seeing you toss the ball up when you experience doing it).

[keithbessant]

Studying History, I found that the clear ideas I'd had a about historical causes broke down. It seemed that there were just trends and events existing at the same tme. There was only correlation. Also, surely in a four-dimensional block universe, the whole idea of movement, change and causation becomes a problem. There are things exisitng side by side forever, eg a living being at one point in spacetime, their remains at another. But both exist eternally and one doesn't become another. Things can't become what they're not. I agree with Parmenides and Zeno of Elea. Being, rather than becoming. What is, is.

[audiodead]

It's an interesting video about an interesting topic. Causality is linked to theories of time, space and consciousness. So when people talk about 'theories of everything', you need to think about theories that unify time and consciousness, not just quantum and classical.

For example, if you believe in a universe which is causal at a fundamental level, then you have to reject the idea that quantum randomness if fundamental. If you accept the idea of the 'block universe', then causality goes out the window altogether (along with any notions of fundamentality).

[PeeGee]

Making a distinction between control and influence may be useful. Where control means manipulating something directly, while influence means manipulating something indirectly (by changing its environment).

At the scale of human behaviour, we recognise these as say domination and seduction, instruction and imitation, constraining and incentivising, and so on. The distinction being whether what's being controlled or influenced has any degree of freedom in deciding what new state it will take (no freedom for control, at least some freedom for influence, though you can see how this may be subjective).

[NightmareCourtPictures]

Hello Last Theory,

Thanks for the video, as always. much appreciated.

[peterschmidt]

I think the observer phenomenon is the essential example of causality. An object in a superposition collapses into an object in a discrete state, and it suddenly becomes relative. Perhaps relativity is the grown-up concept of causality. The naive causality which you speak of seems to be an artifact of our language, trying to describe in literal objective words a seamless motion which only a function with complex terms can describe-- the story of the universe indeed... _this_ universe. The redemption of free will would lie in the confirmation of the Many Worlds Hypothesis. If every signal that traverses every synaptic cleft in our brain introduces quantum uncertainty through the breaking of hydrogen bonds during neurotransmitter reactions, our neural networks represent classical information projected into a sort of chaos space. If there are an infinite number of universes encapsulated within a prime singularity, and our minds are interactive with information from other universes, even a little, then although the novelty of ideas exhibited by our "free will" would still be limited at least by what universes it can interact with, that novelty could still far exceed the complexity of the classical information alone, contained by the brain.

[TechyBen]

I do! I'd love to chat about Wolfram's model. I've thought too deep on it, but it can be visualised! Which is great. We could define causality. We'd have to be rigorous at it. The computational universe still has "causality", there is a defined boundary condition to what can be modified and what cannot (computationally), even if we can't fully track *what* did the modification. :)

[nomcognom]

Most people tend to think there is a difference between just preceding something versus bring it about. If you put fire under a pot of water and make it boil, it is difficult to see both things as merely correlated. It is a lot easier to understand that thermal energy was transmitted to the water molecules, which started to manifest this kinetically. In other words, heat caused water to boil. Heating explains molecular agitation, water convection, radiation, and evaporation (i.e. boiling).

The issue of free will doesn't bother me too much, but free will seems to require neither an entirely deterministic world, nor an entirely non deterministic one.

[monkerud]

I liked your take, i'm not sure why you think discreet variables and operations make any difference, so I outlined my thoughts on the subject. The short story is, causality is what nature does, ultimately no there is no difference in kind between general relativity and history, they are both just approximations to the full, pattern of what nature does. But im also quite skeptical about arguing that the equations of general relativity are relevant to the question at all. They don't need to represent any causation at all, they just have a pattern of solutions, possible worlds if you will that are approximations to what the really causal pattern of nature does. And by real, i simply mean that it, what nature does in full detail is the only thing that can serve as a proper definition of causality, and ultimately any sub pattern like general relativity might have just as plausible mechanisms we could apply language similar to the language we apply to history, that is our casual use of the word for history might be just as applicable to any of the details of the real pattern. To understand what i mean you really have to read what i write about analog variables.

[SamsaraRevolves]

The earth orbiting the sun is a system. The system didn't cause the system. The system is the system.

Causality is the set of preceding situations and conditions that led to the emergence of the earth-sun system. The system stays the same until interfered upon by external, interactions.

Seems straightforward to me, but I'm just an ignorant layman...

[geoffclements]

Where did you get the data that once you reach 22 you don't change your mind about free will? I'm 63 and I changed my mind about it around 5 years ago. Somewhat appositely I felt I had no choice but it change it :)

[setaihedron]

Rupert Sheldrake would say that causality is a list of reliably predictable habits. I would think that Wolfram would say that causality is the aggregation of increasing uncertainty (entropy).

[donmanley]

Causality refers to our perception of the arrow of time