Big Picture Part One: Cosmos

One of a series of quick posts on the six sections of my book The Big Picture — Cosmos, Understanding, Essence, Complexity, Thinking, Caring.

Chapters in Part One, Cosmos:

  • 1. The Fundamental Nature of Reality
  • 2. Poetic Naturalism
  • 3. The World Moves By Itself
  • 4. What Determines What Will Happen Next?
  • 5. Reasons Why
  • 6. Our Universe
  • 7. Time’s Arrow
  • 8. Memories and Causes

It wasn’t easy to settle on how to organize all the material in The Big Picture. Ultimately I decided to start with the universe — a bit of cosmology, yes, but also some basic features of how the universe physically operates. Features that played a crucial role in the transition from an ancient view of the world, which still lingers on in our informal “manifest image” of how things work, to our modern scientific view.

Ibn-SinaOne important step in that transition was the seemingly-innocuous realization that momentum is conserved. In Aristotle’s physics, to keep something moving you had to keep pushing it. That’s a very sensible thing to believe, since it’s absolutely true in our everyday experience. It took many centuries of thinking by very smart people (including Persian polymath Ibn Sina, right) to realize that things tend to move by themselves, and are only slowed down by external forces such as friction. That’s important for physics, of course, but there is a deeper implication for our picture of what kinds of things we need to invoke to explain the universe. As I put it in the book:

Aristotle’s argument for an unmoved mover rests on his idea that motions require causes. Once we know about conservation of momentum, that idea loses its steam. We can quibble over the details — I have no doubt Aristotle would have been able to come up with an ingenious way of accounting for objects on frictionless surfaces moving at constant velocity. What matters is that the new physics of Galileo and his friends implied an entirely new ontology, a deep shift in how we thought about the nature of reality. “Causes” didn’t have the central role that they once did. The universe doesn’t need a push; it can just keep going.

It’s hard to over-emphasize the importance of this shift. Of course, even today, we talk about causes and effects all the time. But if you open the contemporary equivalent of Aristotle’s Physics — a textbook on quantum field theory, for example — words like that are nowhere to be found. We sometimes talk about causes, but they’re no longer part of our best fundamental ontology.

What we’re seeing is a manifestation of the layered nature of our descriptions of reality. At the deepest level we currently know about, the basic notions are things like “spacetime,” “quantum fields,” “equations of motions,” and “interactions.” No causes, whether material, formal, efficient, or final. But there are levels on top of that, where the vocabulary changes. Indeed, it’s possible to recover pieces of Aristotle’s physics quantitatively, as limits of Newtonian mechanics in an appropriate regime, where dissipation and friction are central. (Coffee cups do come to a stop, after all.) In the same way, it’s possible to understand why it’s so useful to refer to causes and effects in our everyday experience, even if they’re not present in the underlying equations. There are many different useful stories we have to tell about reality to get along in the world.

The idea that “cause and effect” isn’t fundamental to the workings of the universe hasn’t spread as widely as it should have, despite the efforts of smart people such as Bertrand Russell. In this first section of the book I sketch how we moved from a picture of the universe animated by causes and reasons to one that obeys patterns, without the need for anything to cause it or sustain it. Of course the idea of causality is still crucial to our everyday lives, so I talk a bit about how cause-and-effect relations are emergent phenomena in a macroscopic world with a pronounced arrow of time.

This entry was posted in Big Picture. Bookmark the permalink.

21 Responses to Big Picture Part One: Cosmos

  1. Reader297 says:


    You write “Of course, even today, we talk about causes and effects all the time. But if you open the contemporary equivalent of Aristotle’s Physics — a textbook on quantum field theory, for example — words like that are nowhere to be found.”

    Actually, due to relativity, causality is a very important constraint on the structure of QFTs, and comes up frequently in textbooks. (Quite a bit of the opening chapters of Peskin and Schroeder are concerned with the consequences of causality!)

    We need our QFT Hamiltonians to be local and to commute with each other at spacelike separation (making it natural to build them out of local quantum fields that satisfy the corresponding microcausality constraint [phi_(x1), phi(x_2)] = 0 ) precisely because action-at-a-distance leads to conflicts with causality in relativistic physical theories. Without the microcausality condition, the time-ordered products appearing in S-matrix elements would, for example, be frame-dependent, density matrices could influence each other across nonlocal separation distances, and all that leads to causal paradoxes — grandfather paradoxes, etc.

    If logical chains of cause-and-effect were truly irrelevant, nonlocal interactions wouldn’t be such a headache, and people would certainly be less bothered by the seemingly nonlocal influences in certain interpretations of quantum theory.

    As for the Aristotelians, I think the larger problem wasn’t that they couldn’t imagine a universe that went back infinitely far in time (after all, they could perfectly have conceived of an infinite chain of causes and effects), but that **if** the universe began at some moment in time, then it “had” to have a prime mover. Science as a whole has moved some distance toward dealing with that question, but conservation of momentum (even in its more abstract sense that things can keep going forever) doesn’t offer an answer.

  2. Sean Carroll says:

    Of course the word “causality” appears in physics books quite a bit. But not “cause and effect,” or notions that actually match onto Aristotle’s four causes; it’s a different use of the same word.

  3. Reader297 says:


    That’s an interesting point. So what do you precisely mean by causality, and why should we care about violating it if it has nothing to do with cause and effect?

  4. Reader297 says:

    I mean, I know it’s hardly a canonical reference on the subject, but Wikipedia defines causality thusly:

    “Causality (also referred to as ‘causation’, or ’cause and effect’) is the agency or efficacy that connects one process (the cause) with another (the effect), where the first is understood to be partly responsible for the second, and the second is dependent on the first. In general, a process has many causes, which are said to be causal factors for it, and all lie in its past. An effect can in turn be a cause of many other effects, which all lie in its future.”

    In the Wikipedia article specially labeled “Causality (physics),” the definition appears as follows: “Causality is the relationship between causes and effects.” That’s what most people (physicists included) appear to have in mind when they talk about the necessity of avoiding causal violations.

  5. Sean Carroll says:

    You should read the book! But “causality” in a QFT context has to do with correlators inside/outside the light cone, and is completely time-symmetric. Aristotle, especially with “final causes” (the end for which an event occurs) is teleological.

  6. Reader297 says:

    It is true that causality in QFT has to do with correlators, and it’s true that if they vanish at spacelike separation, then the result is time-symmetric.

    But why do we care that correlators have to vanish outside the light cone? Textbooks just say that it’s “because of causality,” but that’s not an answer if causality isn’t given a clear definition and isn’t about cause and effect.

    I think we may be talking past each other here. Again, if causality is not about cause and effect, then **why do we care?** And why, similarly, does anyone care about nonlocality in certain interpretations of quantum theory?

  7. Sean Carroll says:

    I think we care because signals propagating inside light cones seems to be a true fact of nature (albeit a time-symmetric one). Not because of our intuitive desire to attribute “causes” to “effects.”

  8. Reader297 says:

    I see. So, if I understand you correctly, you are asserting that “causality” as a constraint on physical theories is just because, as an empirical statement, we simply don’t ever see signals going outside light cones, just as a bare, observed fact of nature. And, correspondingly, you’re saying that there would be no logical problem if they could. We just never see it happen in practice.

    That’s certainly not how we teach this in physics classes and textbooks! Every book on special relativity expends quite a bit of ink on the idea that superluminal signals would lead to grandfather paradoxes, etc. I suppose you might also say that this is the dominant point of view among physicists, but I haven’t really encountered it before this conversation with you here.

  9. Lord says:

    If time doesn’t exist, you could call it the arrow of consciousness.

  10. Crawford MacCallum says:

    I guess I better read the book. The way I thought I learned it was, whereas Aristotle said force causes motion, Newton said force causes acceleration. Don’t understand why that is so deep.

  11. The Thinker says:

    It seems to me that given spacetime, objects are worldtubes in the four-dimensional block, and that means that what causality really is would seem to have to be the relationships of intersecting worldtubes as then precede each other or intertwine with another. For example, asking ‘why do I exist now?’ would be explained by the fact that at an earlier event in spacetime my parents had sex. That was the ’cause’ that resulted in my birth and existence now – but only in the sense that if you trace my worldtube back in spacetime to its origin it’s preceded by my parent’s worldtubes and thus that establishes the ‘causal’ relationship. But just in how a road splitting into two roads doesn’t really ’cause’ the new roads in how we normally think of causality, relationships among worldtubes are not really causal in the temporal sense we think of them.

  12. Christian says:

    You wrote:
    > Of course, even today, we talk about causes and effects all the time.
    > But if you open the contemporary equivalent of Aristotle’s Physics —
    > a textbook on quantum field theory, for example — words like that
    > are nowhere to be found. We sometimes talk about causes,
    > but they’re no longer part of our best fundamental ontology.

    Uhm… If I open a textbook on quantum field theory, I am likely to learn about the S-matrix, which relates the initial state (cause) and the final state (effect). Basically, the S-matrix encodes the relation between cause and effect. The S-matrix is pretty fundamental to quantum field theory, and I don’t see how we can claim that causes and effects are “not present in the underlying equations”. The S-matrix pretty much underlies all of it, dosn’t it?

    I find myself confused.

  13. Torbjörn Larsson says:

    No causes, whether material, formal, efficient, or final. … In the same way, it’s possible to understand why it’s so useful to refer to causes and effects in our everyday experience,

    Very good analysis!

    It reminds me of how it can be convenient, but confusing on a deeper level, to attribute causes and effects in biology. Famously, there is no global ’cause’ why a trait evolves, no ‘goal’, it is a space and time localized mechanism that works on a global population over extended times with many contingencies at play.

    Potato, potatoe.

    @Reader: I think you may still be talking cross purpose.

    Sean notes that light cone causality is a fact, but does not delve into why that is. It has no ’cause’ by itself I think.

    I have seen papers on why signaling faster than the universal speed limit would destabilize – make impossible – gauge theories used to describe fields and particles. If so, no causality, no Core Theory. no conservation laws, no physics.

    In the end then, we face the problem of likelihoods on an infinite space and/or time universe. Why would there be a likelihood for “no physics”, when all we know is that “physics” exists as far back (and far out) as we can see? There is no ultimate cause here.

  14. Pete says:

    Can’t wait to pick up the book Sean.

    Was just wondering something in relation to all this talk of “causality.” Did you ever research anything by philosopher Judea Pearl (he’s actually the father of slain journalist Daniel Pearl)? He’s considered a top contemporary philosopher and his work on causality ( has been nothing short of extraordinary.

    I know you’re really good about understanding how philosophical assumptions are incorporated into everything (including scientific methodology itself), and I just wasn’t sure if your research led you to Pearl’s work.

  15. Sean Carroll says:

    I’ve read Pearl, but in the interest of keeping things simple I didn’t try to explain his ideas in this book.

    Also I clearly underestimated the extent to which physicists don’t realize that their use of “causality” (which is something like “global hyperbolicity”) isn’t the same as Aristotelian final causation or macroscopic notions of cause and effect.

  16. Elto Desukane says:

    In the blockworld view of spacetime, the whole of spacetime exist, fixed, past and future, all determined, no indeterminacy, no probability, no cause and effect, no causality. Of course, if we insist to analyse this 4D spacetime as a 3D space changing over time, then we must somehow generate the future from the past, with causes producing effects, which may not be possible in a deterministic way.
    To view the 4D spacetime as a 3D space changing over time is obviously the natural common sense thing to do, but it may well be a fundamentally wrong point of view.

  17. Mark S. says:

    Hey folks, fyi about the biologist Aristotle. His physics was refuted in his time but it was centuries later Christians and others that made it seem that his physics was what came before Newton. To listen about what science was really like back in the times of the Greeks and Romans then go here:
    “The Sciences in Ancient Greece & Rome: How Far Did They Get?”

  18. Jayarava says:

    “The idea that “cause and effect” isn’t fundamental to the workings of the universe”… has blown my mind! Thanks 🙂

  19. Rhett says:

    Currently in this section of the audiobook…

    Sean, I’m really quite grateful that you, of all people, took this on. I enjoy your approach to poetic naturalism and given your background how you are able to stand pretty firmly within the realms of both the scientific (and specifically in the fields of cosmology/particle physics) and the philosophical.

    As someone with a religious upbringing, and over the last ten years through my 30’s, has emerged into a firm naturalist, I really really appreciate this approach. I’ve consumed Sagan, Dawkins, Hitchens and Harris, who have taken on a more overtly anti-religious approach, and appreciate that approach. But what they did was to, in a sense, lend some credence to their antithesis… They focused, sometimes stridently, on knocking down religion. Those have been, and will be, worthwhile in their own right. They will serve a purpose within a certain sphere…

    But, again given my own background, I think this approach to the topic is really needed in this time. The approach of simply building the naturalist ontology from the ground up based upon all the knowledge we have up to now. Referencing, simply because it’s there, religious ontologies, but not so much taking them on as simply saying that the naturalist ontology is clearly rationally superior so we can, while understanding why they have sway based on historical and psychological reasons, begin to simply discount religious ontologies as “also-rans” in the evolution of human ontologies, not as true competitors…

    The Dawkins/Hitchens/Harris/Krauss/et al. approach just isn’t necessary if the goal is to make naturalism more available… These are generational issues… To try and break grown adults out of long held beliefs is just a huge challenge and honestly, not one really worth waging.

    The real task is to imbue those coming into adulthood with this new way of understanding the world, one which can make a more stable bridge to the that important section of the book…Caring.

    So, thanks again for this effort, Sean. I think it’s really important and something others have attempted (i.e. Krauss’s “A Universe from Nothing”), but not really succeeded at until now….

  20. Glen Marquis says:

    Hello one and all,

    Since we’re touching on causailty, in what sense firstly does Relativity define, a begin to exist?…Isn’t it entirely plausible for cosmos-ingredients to phase transition out of quantum-effects, given that Relativity does not account for any QM effects? Or, perhaps, pre-existing time-like-space transitions into spacetime?…Making uncaused-physical-causality an entirely coherent option that never begun to exist. “Nothing” being a minima, but never Nothingness….your thoughts Gentlemen?


  21. sayak banerjee says:

    Hi Sean,

    I have purchased your book on Kindle and am enjoying my read through it. I am currently reading the section in COSMOS regarding time’s arrow and I have some difficulty in accepting your proposition that without the entropy asymmetry of the Big Bang there will be no way to distinguish the future from the past. I am creating a thought experiment to make my point clear.

    Assume a vast tank filled with only electrons in perfect thermal equilibrium at a certain temperature. (I am using electrons as they are free standing very long lived elementary particles devoid of all the complex internal quark structures that protons and neutrons have. ) I am the precocious Laplace’s demon tasked with recording any and every information I can about all that the electrons do in the bath through time. At a certain instant, I, with my demoniacal power record the energy, momentum and position of all electrons with as much accuracy as I can muster given the quantum uncertainties and write up the appropriate quantum bra’s for them. Because we have uncertainties in these quantities, my prediction of the future evolution of quantum state of the particles (as they repeatedly collide with each other, decohere and become newly entangled) will be probabilistic, and as I try to predict further and further into the future, we will have lots and lots of possible configurations with lower and lower probabilities in each. That means that the future will get fuzzier and fuzzier. Now suppose the demon continues to moniter the state of the electrons through time and makes 1000 such monitorings through time, each time minimizing the uncertainties as much as possible. Now at the 1001-st monitoring, the demon looks back at the “past” , constituting of the record of all the 1000 observations he had made so far and looks forward in the future at his predictions of what he will find over the next 1000 observations he is going to make in the future. It seems pretty clear to me that the uncertainties associated with the predicted state of 1000 future observations will be FAR GREATER than the recorded uncertainties of the past 1000 observations whose records he has. This is the case even though the electron bath is in perfect thermal equilibrium with maximum entropy at all states. This seems to separate the future from the past in such cases as well. The demon still knows the past far better than the future.

    Its a rough thought experiment and doubtless there are inaccuracies in how I describe the quantum theory. But if you comprehend the idea behind the thought experiment, could you explain where the idea I have proposed has gone wrong? Thank you