The Big Picture: On the Origins of Life, Meaning, and the Universe Itself

by Sean Carroll

We will ultimately understand the world as a single, unified reality, not caused or sustained or influenced by anything outside itself. That’s a big deal.

That doesn’t mean we can’t talk about ships just because we understand that they are collections of atoms. It would be horrendously inconvenient if, anytime someone asked us a question about something happening in the world, we limited our allowable responses to a listing of a huge set of atoms and how they were arranged.

The poet Muriel Rukeyser once wrote, “The universe is made of stories, not of atoms.”

As far as our best current physics is concerned, each moment in the progression of time follows from the previous moment according to clear, impersonal, quantitative rules.

With Boltzmann’s definition in hand, it makes perfect sense that entropy tends to increase over time. The reason is simple: there are far more states with high entropy than states with low entropy. If you start in a low-entropy configuration and simply evolve in almost any direction, your entropy is extraordinarily likely to increase. When the entropy of a system is as high as it can get, we say that the system is in equilibrium. In equilibrium, time has no arrow.

How well do we know what we think we know? If we want to tackle big-picture questions about the ultimate nature of reality and our place within it, it will be helpful to think about the best way of moving toward reliability in our understanding.

The Bayesian approach is much more general than this, however. It reminds us that we assign prior credences, and update them appropriately, to every factual proposition that may or may not be true about the world. Does God exist? Can our inner conscious experiences be explained in purely physical terms? Are there objective standards of right and wrong? All of the possible answers to such questions are propositions for which each of us has a prior credence (whether we admit it or not), and which we update when relevant new information comes in (whether we do so correctly or not).

Bayes teaches us (1) never to assign perfect certainty to any such belief; (2) always to be prepared to update our credences when new evidence comes along; and (3) how exactly such evidence alters the credences we assign. It’s a road map for coming closer and closer to the truth.

Radical skepticism is less useful to us; it gives us no way to go through life. All of our purported knowledge, and all of our goals and aspirations, might very well be tricks being played on us. But what then? We cannot actually act on such a belief, since any act we might think is reasonable would have been suggested to us by that annoying demon. Whereas, if we take the world roughly at face value, we have a way of moving forward. There are things we want to do, questions we want to answer, and strategies for making them happen. We have every right to give high credence to views of the world that are productive and fruitful, in preference to those that would leave us paralyzed with ennui.

It makes sense, as Wittgenstein would say, to apportion the overwhelming majority of our credence to the possibility that the world we see is real, and functions pretty much as we see it. Naturally, we are always willing to update our beliefs in the face of new evidence. If there comes a clear night, when the stars in the sky rearrange themselves to say, “I AM YOUR PROGRAMMER. HOW DO YOU LIKE YOUR SIMULATION SO FAR?” we can shift our credences appropriately.

The different stories or theories use utterly different vocabularies; they are different ontologies, despite describing the same underlying reality. In one we talk about the density, pressure, and viscosity of the fluid; in the other we talk about the position and velocity of all the individual molecules. Each story comes with an elaborate set of ingredients—objects, properties, processes, relations—and those ingredients can be wildly different from one story to another, even if they are all “true.” Each theory has a particular domain of applicability. The fluid description wouldn’t be legitimate if the number of molecules in a region were so small that the effects of particular molecules were important individually, rather than only in aggregate. The molecular description is effective under wider circumstances, but still not always; we could imagine packing enough molecules into a small enough region of space that they collapsed to make a black hole, and the molecular vocabulary would no longer be appropriate. Within their respective domains of applicability, each theory is autonomous—complete and self-contained, neither relying on the other. If we’re speaking the fluid language, we describe the air using density and pressure and so on. Specifying those quantities is enough to answer whatever questions we have about the air, according to that theory. In particular, we don’t need to ever refer to any ideas about molecules and their properties. Historically, we talked about...

The reason why emergence is so helpful is that different theories are not created equal. Within its domain of applicability, the emergent fluid theory is enormously more computationally efficient than the microscopic molecular theory. It’s easier to write down a few fluid variables than the states of all those molecules. Typically—though not necessarily—the theory that has a wider domain of applicability will also be the one that is more computationally cumbersome. There tends to be a trade-off between comprehensiveness of a theory and its practicality.

The molecular and fluid descriptions of air in a room provide an innocent, uncontroversial example of emergence. Everyone agrees on what is happening and how to talk about it. But its simplicity can be misleading. Seeing how relatively easy it is to derive fluid mechanics from molecules, one can get the idea that deriving one theory from another is what emergence is all about. It’s not—emergence is about different theories speaking different languages, but offering compatible descriptions of the same underlying phenomena in their respective domains of applicability.

Is an emergent theory just a way of repackaging the microscopic theory, or is it something truly novel? For that matter, is the behavior of the emergent theory derivable, even in principle, from the microscopic description, or does the underlying stuff literally act differently in the macroscopic context? A more provocative way of putting the same questions would be: are emergent phenomena real, or merely illusory?

There are several different questions here, which are related to one another but logically distinct. Are the most fine-grained (microscopic, comprehensive) stories the most interesting or important ones? As a research program, is the best way to understand macroscopic phenomena to first understand microscopic phenomena, and then derive the emergent description? Is there something we learn by studying the emergent level that we could not understand by studying the microscopic level, even if we were as smart as Laplace’s Demon? Is behavior at the macroscopic level incompatible—literally inconsistent with—how we would expect the system to behave if we knew only the microscopic rules?

We’re now entering into the realm known as strong emergence. So far we’ve been discussing “weak emergence”: even if the emergent theory gives you new understanding and an enormous increase in practicality in terms of calculations, in principle you could put the microscopic theory on a computer and simulate it, thereby finding out exactly how the system would behave. In strong emergence—if such a thing actually exists—that wouldn’t be possible. When many parts come together to make a whole, in this view, not only should we be on the lookout for new knowledge in the form of better ways to describe the system, but we should contemplate new behavior. In strong emergence, the behavior of a system with many parts is not reducible to the aggregate behavior of all those parts, even in principle.

A poetic naturalist has another way out: something is “real” if it plays an essential role in some particular story of reality that, as far as we can tell, provides an accurate description of the world within its domain of applicability.

Poetic naturalism sits in between: there is only one, unified, physical world, but many useful ways of talking about it, each of which captures an element of reality. Poetic naturalism is at least consistent with its own standards: it tries to provide the most useful way of talking about the world we have.

Human beings are not nearly as coolly rational as we like to think we are. Having set up comfortable planets of belief, we become resistant to altering them, and develop cognitive biases that prevent us from seeing the world with perfect clarity. We aspire to be perfect Bayesian abductors, impartially reasoning to the best explanation—but most often we take new data and squeeze it to fit with our preconceptions.

Robert Aumann, an Israeli American mathematician who shared the Nobel Prize in Economics in 2005, was able to prove a wonderful mathematical theorem: two people, both acting rationally, who start with the same Bayesian prior credences for their beliefs, and who have access to the same information, including knowing what the other knows, cannot disagree about the updated credences for those beliefs.

We are left with, not absolute proof of anything, but a high degree of confidence in some things, and greater uncertainty in others. That’s both the best we can hope for and what the world does as a matter of fact grant us. Life is short, and certainty never happens.

This quantum probability is very different from ordinary classical uncertainty. Think once again of playing poker. At the end of a certain hand, your opponent makes a big bet, and you need to decide whether your hand can beat theirs. You don’t know what their hand is, but you know what the possibilities are: nothing, a pair, three of a kind, and so forth. Given their behavior so far in the hand, and the odds that they received certain cards to start, you can be a good Bayesian and assign different probabilities to the various hands they could have. Quantum states sound kind of like that, but they are crucially different. In the (classical) game of poker, you don’t know what your opponent has, but they have something definite. When we say that a quantum state is a superposition, we don’t mean “it could be any one of various possibilities, we’re not sure which.” We mean “it is a weighted combination of all those possibilities at the same time.” If you could somehow play “quantum poker,” your opponent would really have some combination of each of the possible hands all at once, and their hand would become one specific alternative only once they turned over the cards for you to look at them.

The observed and predicted values aren’t exactly the same, both because of experimental error and because of theoretical approximations. But the lesson is clear: quantum mechanics isn’t some loosey-goosey, anything-goes kind of operation. It is relentlessly specific and unforgiving.

Antirealism is a pretty dramatic step to take. It seems to have been advocated, however, by no less of an authority than Niels Bohr, the grandfather of quantum mechanics. His views were described as “There is no quantum world. There is only an abstract physical description. It is wrong to think that the task of physics is to find out how nature is. Physics concerns what we can say about nature.”

So the wave function of the universe assigns a number to every possible way that objects in the universe could be distributed through space. There’s one number for “the Earth is here, and Mars is over there,” and another number for “the Earth is at this other place, and Mars is yet somewhere else,” and so on.

At last, then, we have a candidate for a final answer to the critical ontological question “What is the world, really?” It is a quantum wave function. At least until a better theory comes along.

There’s a lot to love about the Everett/Many-Worlds approach to quantum mechanics. It is lean and mean, ontologically speaking; there’s just the quantum state and its single evolution equation. It’s perfectly deterministic, even though individual observers can’t tell which world they are in before they actually look at it, so there is necessarily some probabilistic component when it comes to people making predictions. And there’s no difficulty in explaining things like the measurement process, or any need to invoke conscious observers to carry out such measurements. Everything is just a wave function, and all wave functions evolve in the same way.

Our goal is to offer a plausibility sketch that the world can ultimately be understood on the basis of naturalism. We don’t know how life began, or how consciousness works, but we can argue that there’s little or no reason to look beyond the natural world for the right explanations. We can always be wrong in that belief; but then again, we can always be wrong about any belief.

Our human-scale world is relatively calm and predictable. Throw a ball on a day with good weather, and you can estimate with some confidence how far it will travel. Cells, by contrast, operate at the scale of nanometers, billionths of a meter. Conditions in that world are dominated by random motions and noise—what biophysicist Peter Hoffmann has dubbed a “molecular storm.” Just from ordinary thermal jiggling, molecules inside our bodies bump into one another trillions of times a second, in a maelstrom that puts ordinary storms to shame. Scaled up to human size, living in the equivalent of the cell’s molecular storm would be like trying to throw a ball that was constantly being bombarded by other balls, each of which carried hundreds of millions of times the energy that your arm could impart.

It’s one thing to see how living organisms can harness free energy to maintain themselves and move around. It’s quite another thing to understand how life ever got started.

The lipid’s search for contentment is a metaphorical way of talking about the fact that the system evolves so as to minimize free energy. Entropy increases, which suggests to us a certain emergent vocabulary, in which the molecules “want” to find a state with low free energy. The arrow of time leads us to speak a language of purpose and desire, even though we’re only talking about molecules obeying the laws of physics.

This is where the hydrogenation of carbon dioxide comes in. Russell’s comment alludes to the fact that there is free energy locked up in a mixture of carbon dioxide (CO2) and hydrogen gas (H2), both of which were abundant in certain environments on the young Earth. If the carbon could somehow shed its two oxygen atoms and replace them with hydrogen, we could end up with methane (CH4) and water (H2O). That’s a configuration that has less free energy; as far as the second law of thermodynamics is concerned, it’s a transformation that “wants” to happen.

Albert Szent-Györgyi, a Hungarian physiologist who won the Nobel Prize in 1937 for the discovery of vitamin C, once offered the opinion that “life is nothing but an electron looking for a place to rest.”

“We know there’s life; how did it start?” Instead, it’s suggesting that life is the solution to a problem: “We have some free energy; how do we liberate it?”

So the real world is a beautiful mess. Is this kind of undirected mechanism—just what we would expect in a universe governed by unthinking underlying laws and with a strong arrow of time—sufficient to account for all the spectacular intricacy of our planet’s biosphere? “There is grandeur in this view of life,” Darwin writes in On the Origin of Species. But is his simple mechanism really enough to make dolphins and butterflies and rain forests from a meager collection of organic molecules fighting for free energy? Can the wonders of efficiency and ingenuity we see in biological organisms really come about from random variation plus time? (Hint: yes.)

It’s a natural thing to worry about. The process of evolution is unplanned and unguided. Whether or not genetic information gets passed on to future generations depends only on the conditions of its immediate environment and random chance, not on any future goals. How can an intrinsically purposeless process lead to the existence of purposes?

The American cultural anthropologist Ernest Becker, commenting on Danish philosopher Søren Kierkegaard, once characterized consciousness this way: What does it mean to be a self-conscious animal? The idea is ludicrous, if it is not monstrous. It means to know that one is food for worms. This is the terror: to have emerged from nothing, to have a name, consciousness of self, deep inner feelings, and excruciating inner yearning for life and self-expression—and with all this yet to die.

Or in the words of neuroscientist David Eagleman, “Your consciousness is like a tiny stowaway on a transatlantic steamship, taking credit for the journey without acknowledging the massive engineering underfoot.”

A poetic naturalist says that we can have two very different-sounding ways of describing the world, a physics-level story and a human-level story, which invoke separate sets of concepts and yet end up being compatible in their predictions concerning what happens in the world. A libertarian thinks that the right way to talk about human beings ends up making predictions that are incompatible with the known laws of physics. We don’t need to do such dramatic violence to our understanding of reality just to make peace with the fact that we make choices as we go through the day.