As with all such documents, a revision history is included at the bottom.
Lear: … what can you say to draw a third more opulent than your sisters? Speak.
Cordelia: Nothing, my lord.
Lear: Nothing will come of nothing, speak again.
— William Shakespeare, King Lear Act 1 Scene 1
The “argument from first cause” is a particularly common argument used to attempt to demonstrate the necessity of a divine creator. There are several versions of this argument. William Lane Craig, one evangelical Christianity’s most common defenders in debates, nearly always includes the Cosmological Argument as part of his opening statements. His formulation of this argument is among the most concise, and so has gained a great deal of traction among Christians.
Statement of the Argument, and the Short Version of my Responses
Trace the line of cause and effect backward. Every effect had a cause. Every cause is itself an effect of a prior cause. Traced back far enough, we must conclude that there was some “First Cause” that did not itself need anything to cause it. This “First Cause” we call “God.” Craig’s formulation is:
- Whatever begins to exist has a cause.
- The universe began to exist.
- Therefore the universe had a cause.
This is certainly a simplified version of the argument, but I believe it contains the essential elements of the idea. An even more simplified version of the argument can be expressed as the statement from King Lear shown above… “Nothing will come of nothing…” but I feel this version leaves a bit out, so I’ll address the slightly longer version from the previous paragraph. If any of my readers think that there are essential elements that I have left out, then I hope they will comment so that I can ensure that I am addressing the real argument rather than a strawman.
For such a short, seemingly-straightforward line of reasoning, I think you will be surprised how much there is to discuss about it. The short-version of the points that I want to make is:
- This linear chain of causation oversimplifies the idea of causation dramatically; many effects have multiple causes, and many causes have multiple effects.
- Causation itself may simply be a psychological shorthand rather than a real phenomenon.
- If this argument provides the definition of “God,” then there is little reason to suspect that this “God” has key characteristics of what most people mean when they use the word, “God.”
- If this argument is used to justify the existence of a separately posited or demonstrated “God,” it is not clear that the addition of God to explain the first seemingly-uncaused event actually adds any explanatory value.
- The current understanding of the so-called Big Bang includes time itself being created in that event. It is not clear how cause-effect relationships work at the boundary of time itself.
- It is not true that every effect has a cause. The most widely-accepted interpretation of quantum mechanics, the Copenhagen Interpretation, features uncaused effects in a very central way.
- The Heisenberg Uncertainty Principle, a central part of quantum mechanics, provides a mechanism for “something” to arise from “nothing,” vacuum fluctuations. These vacuum fluctuations have been experimentally verified.
- The type of “begin[ing] to exist” most relevant to Craig’s first premise has never been observed to have a cause.
The application of the scientific points to the origins of the universe, as you will see, isn’t unassailable, since there is a significant amount of speculation involved. You should notice that I am not saying, “This is what science says happened,” but rather, “This is a plausible scenario that is consistent with established science.” Given the choice between such a plausible scenario and an explanation that involves divine, or at least supernatural, agency, the more likely explanation is the one that doesn’t posit the divine (readers may find it useful to explore my essay on miracles and the supernatural concerning this point).
Causality isn’t a Linear Chain
Consider buying a used car. There are multiple considerations that simultaneously go into the decision, including price, age of the car, mileage on the car, make and model, apparent condition, and fit of the car with its intended uses. No one of those can generally be identified as the consideration that led to the purchase; they all played a role. Similarly, most effects can be traced to multiple causes. In an analogous argument, a single event, like dropping a pebble into a still pond, produces ripples that spread out over the entire surface of the pond. The movement of the surface of the pond on the left side can just as easily be traced back to the dropping of the pebble as can the movement of the surface on the right side. Hence, multiple effects can come from a single cause. This suggests that causality is really more of a web of interactions than it is a linear chain. Of course none of this suggests that there are effects without cause at all (that’s coming in a later point), so this objection is really just a clarification of the argument rather than a rebuttal to it.
Causation may be a Psychological Phenomenon
While it is definitely a convenient psychological shorthand to label causes and effects, it may be that the reality of the “causation,” at least in the conventional sense, is highly questionable. Let us take the example of general relativity, and examine the motion of a planet about a star. In the conventional way of describing cause and effect, the cause for the planet to move in its orbit is the gravitational field of the star. However, in the general relativity description, there isn’t truly a gravitational field… space itself is bent, or warped, in such a way that from the planets perspective it is actually moving in a straight line. This is similar to the “straight” lines on the surface of a globe intersecting themselves. The planet is doing nothing except following Newton’s laws of motion, just in a somewhat nonintuitive environment. Is it the star that causes the orbit? Is it the gravitational field that causes it? Or is the orbit itself simply a mental (and mathematical) construct of ours to describe the observed motion? I think the latter is the most accurate.
Causation is a tidy way for our minds to generalize observed phenomena, yet it introduces levels of approximation that may make it philosophically indefensible. That doesn’t make it useless, of course, just as Newtonian mechanics, while simply an approximation to reality, are very useful for engineers. It is simply that it is a mistake to conflate the description of a phenomenon with the phenomenon itself.
If God is the “First Cause,” Is He/She/It Recognizable as God?
If the argument from first cause were to be coherent, and if we were to simply define whatever that first cause was to be God, it is not all that clear that the God defined in this way matches up with the God of any of the established religions. There is nothing in the argument itself that would require the first cause to be a sentient (i.e., self-aware) consciousness rather than a non-sentient force. Let me try to explain it by analogy… a dangerous prospect, of course, since all analogies fall apart when stretched too far. Look at water that has been heated almost to the point of boiling. What is the “cause” of the first bubble forming? The nucleation of a phase transition from liquid to gas, initiated at a particular point unpredictably because of fractal mathematics. Is that nucleation caused by a sentient being? Not at all, particularly if you look at water boiling near a lava vent, so there isn’t some “cook” that turned up the heat. Why is it difficult to see our universe as a bubble that has spontaneously nucleated in a process called the Big Bang? Given my scientific points below, this seems the most reasonable thing in the world.
If even the sentience of the first cause isn’t implied by the argument, then I hope it is clear that many other features that are commonly associated with God are even less implied, including that he in any way influenced the world after the creation event, that he is the inspiration for any particular holy book, that he has anything to do with morality, and that he has anything to do with a human afterlife. All of these aspects of the first cause, and many more besides, would have to be attributed to this God by other arguments.
The “First Cause” as God Holds no Explanatory Power
If, instead, the existence of God were to be posited or demonstrated by other means, saying that God is the first cause of the universe doesn’t actually explain anything that needs explaining. The problem that the argument from first cause strives to address is the issue of causality requiring a cause. Whatever the first effect is, presumably the Big Bang, we seemingly have to ask what caused it. The problem with the argument from first cause for God is, however, that if we were to say that God caused the Big Bang, we have simply shifted the question, not answered it. If God caused the Big Bang, then what caused God? If, on the other hand, God doesn’t need a cause, then why are we unable to say the same thing about the Big Bang? Philosophically, there is no good justification that I have found for claiming that God is exempt from causation that could not equally be applied to the Big Bang itself, and applying that special attribute to the Big Bang rather than to God makes for a simpler explanation, thus suggesting that we invoke Occam’s Razor (about which I will probably have a separate essay sometime in the future) and eliminate the God hypothesis.
The one argument that I have heard that begins to address this objection is the following: causality is only required within a sequential timeline; God exists outside of our timeline, and thus does not require a cause. There is a subtle problem with this argument, however. If God exists outside of our timeline, he nonetheless must exist within some timeline, and therefore within his timeline, he still requires a cause as much as the Big Bang does within our timeline.
It may seem to be a pretty bold statement for me to assert that God must exist within some timeline. However, I hope you will all agree that any theistic vision of God includes the notion that God does things… that he takes actions. Without a timeline of some kind, the entire concept of an action is incoherent. For the concept of an action to be a meaningful concept, there must be a state-of-affairs that was before the action and a state of affairs that was after the action. Q.E.D. If any of my readers would like to try to argue for a theistic God that exists in no timeline at all, then I would love to see the explanation and try to address it. Lacking such an argument, however, I believe that these last two philosophical points demonstrate the unreasonableness of the first cause argument as a whole, and the distinct lack of connection of the supposed first cause with any specific theistic version of God.
What Does “Before the Big Bang” Even Mean?
Many people trying to argue the first cause position accept, at least for the sake of argument, the Big Bang as the first event in the universe. The natural question is, what was there before the big bang? Numerous physicists, including Stephen Hawking (as described in his popular book, “A Brief History of Time”) consider this to be a faulty question. Hawking’s illustration by analogy goes as follows: what happens if you go to the North Pole and then head north? Well, that’s not really a meaningful question, because there is no north at the North Pole. Every direction that keeps you on the surface of the planet is south. Similarly, Hawking argues, there is no “before” at the Big Bang.
While this response may be correct based on the equations of general relativity, it is quite hard to understand. There are a lot of scientifically demonstrable concepts that are quite difficult to reconcile with our intuitive, everyday expectations. The behavior of atoms and molecules, for example, are governed by quantum mechanics, which includes incredibly counterintuitive but directly-measurable effects. Our expectations about how the world work were developed and honed on a size scale and timeframe where quantum mechanics and general relativity mathematically reduce to classical Newtonian mechanics. But move to the extremes, like the very small and light or the very heavy or very fast, and our intuitive understanding falls apart. I will go into quantum mechanics in quite some detail in the next couple of sections; but it is primarily relativity that is the issue here, and since relativity is not one of my areas of expertise, and since even accepting this notion that questions about “before the Big Bang” must be disallowed still doesn’t provide a theoretical framework within which the origins of the universe can be understood, I’ll leave this point here and move on to the more quantum-mechanical questions that both are more within my expertise and are more able to provide a justification for the origins of “something” from “nothing.”
The Unavoidable Sidebar 1: Quantum Mechanics
In order for me to make my next couple of points, a basic understanding of quantum mechanics is going to be necessary. Unfortunately, even a basic understanding of this field is quite difficult to grasp. I will do my best to restrict my discussion to the most relevant of the concepts, and will as well try to keep the presentation as accessible as possible, but it is inevitable that some aspects will be unclear, so I welcome any questions from readers that will let me know where I can flesh out or clarify my explanations.
The basic idea of quantum mechanics is that matter… the physical things in the universe… move like waves rather than like particles. The larger and faster the objects move, the more particle-like their motion becomes. For example, the “wavelength,” or the characteristic distance between the crests of a wave, of a baseball is, because the baseball is so large itself, much smaller than that of an individual atom. This means that for everyday objects, we are unable to notice their wavelike behavior. But once you get down to molecules, atoms, or subatomic particles, the wavelengths we are dealing with can be huge compared to their own size, and thus it becomes much more appropriate to describe their motions using wave equations than it does to describe them as having trajectories.
This is undoubtedly a bizarre concept to try to wrap one’s brain around, and yet countless experiments have validated this approach. The most powerful microscopes in existence, transmission electron microscopes, rely on this wave-like motion of the electrons in order to operate, and operate well they do; they have resolutions that can be better than a tenth of an atom across. The entire field of chemistry relies heavily on describing the motions of electrons as waves. Transistors in modern computer chips similarly rely on the wave-nature of electrons.
One of the most direct observations of this wave-like behavior of electrons is the electron double-slit experiment. I recommend highly the video that I just linked to, but I will give a short description here as well. First, let’s look at a couple of comparison cases. If we fire a machine gun at a bullet-proof plate that has two parallel slits in it (figure 1), and then look at the pattern of bullet holes on a wall some distance behind the plate, we will see two lines of holes, consistent with the notion that the bullets follow straight-line trajectories from the gun to the wall, and are only allowed to pass through the two slits. This is classic particle-like behavior.
If, however, we set up a similar experiment using something that behaves like waves, such as the water waves in a ripple-tank or electromagnetic radiation (a.k.a., light), we see instead on the far wall (the detector) an interference pattern (figure 2). We see multiple regions of high intensity, where the crests of the waves passing through one slit match up with the crests of the waves passing through the other slit. We also see these high-intensity regions separated by low… or zero… intensity regions, where the crests from the wave passing through one slit match up with the toughs of the waves passing through the other, thereby cancelling out (figure 3). This interference pattern is a hallmark of classic wave-like behavior. If this is hard to visualize, I again highly recommend the video I linked to.
Now set up the same type of experiment with a beam of electrons. What kind of pattern do we see? Well each electron hits the detector in a specific spot, like a particle, but after accumulating a large number of hits, we see that the pattern formed by these hits is an interference pattern. “Ah,” some say, “clearly the electrons passing through one slit are bouncing off the electrons passing through the other slit.” But in a clever tweak on the experiment, scientists have slowed down the rate of electron emission to the point that only one electron hits the slits at any given time… and an interference pattern was still observed. The natural, though bizarre, explanation is that the electron passed through both slits simultaneously, and the wave-like propagation of the electron as it moves toward the detector interferes with itself. This experiment shows just how poorly our everyday experiences have prepared us to think about the behavior of matter on these small size-scales. And lest you think this bizarre behavior is limited to electrons, this same double-slit experiment has actually been performed with molecules of buckminsterfullerene, C60, instead of electrons, and indeed an interference pattern was observed.
In classical Newtonian physics we describe the motion of a particle using a trajectory. Given a starting position, an initial velocity, and a set of forces acting on the particle, we can map out where the particle is at any point in time. This is how the trajectory of cannon balls is determined, and how the orbits of planets is determined. However, in quantum mechanics there is no trajectory. There is wave-like motion, where all we can do is use the particle wavefunction (the mathematical expression that in quantum mechanics replaces the trajectory) to determine probabilities of where the particle might be. We can say that there’s a 50% chance of finding the particle in region 1, a 25% chance in region 2, and a 25% chance in region 3, but we can’t know for certain until we make the measurement. In fact, in a very real sense (that doesn’t factor into the issues we are dealing with in this essay) the particle itself doesn’t go to a specific spot until the measurement is made; this is the essence of the famous Schrödinger’s Cat thought experiment.
These probabilities can be determined quite precisely, which is why quantum mechanics has such an incredibly strong track record experimentally, but we have to make measurements on lots of individual particles to check them; any one measurement could end up in any of the allowed positions. This means that the motion of quantum-mechanical particles is inherently probabilistic rather than deterministic, and any given outcome isn’t specifically caused. The last hope for causation in quantum mechanics is the “hidden variable” hypothesis, where there is some type of deterministic effect underlying the probabilistic quantum mechanical equations that we haven’t found yet. None of the proposed mechanisms for this have become widely accepted, and some types of possible hidden variables have been experimentally disproven. Thus the prevailing scientific view on quantum mechanics is what is known as the Copenhagen Interpretation, wherein the underlying reality itself is probabilistic rather than deterministic.
There are examples of discrete phenomena that can be determined to either happen or not happen, such as the radioactive decay of a specific atom, where the occurrence depends entirely on the outcome of these quantum mechanical probabilities. These phenomena are therefore good examples of effects that occurred without cause. Thus, in a quantum mechanical world, the argument from first cause fails because the premise that every effect must have a cause is demonstrably violated.
The Unavoidable Sidebar 2: The Heisenberg Uncertainty Principle
Hopefully I’ve convinced you by this point that we have to use wave mathematics to describe the behavior of small particles rather than trajectories. Let’s think a bit about what this implies. If you have a wave with a single wavelength, that wave is spread out in space. The wavelength is related to the velocity of the wave (see figure 4). This means that if you specify the speed of the particle precisely, it doesn’t even make sense to ask about where the particle is, because it is spread out over a large area of space. On the other hand, there is a way to localize the particle in space by adding together (in what is called a linear combination) a bunch of waves of different wavelength. As you can see from the figure 5, the more waves we add in, the more it starts to make sense to ask, “Where is the particle?” But, correspondingly, it becomes less sensible to ask, “How fast is the particle moving?” Because each of the waves that has been added together has a different speed. The particle, in this scenario, is spread out over velocity-space.
This notion is expressed mathematically in quantum mechanics by the Heisenberg Uncertainty Principle. This principle states that there are certain pairs of properties, in the example above speed and position, that cannot be specified precisely at the same time. The more precisely we know speed, the less precisely it is possible to know position, and vice versa. Mathematically this is expressed as ΔpΔx ≥ C, where Δp is the uncertainty in momentum of the particle (momentum is related to velocity as p = mv, where m is the particle mass and v is the velocity), Δx is the uncertainty in position of the particle, and C is a constant of nature, specifically Planck’s constant divided by 4π. When Δx is small, Δp must be large in order for their product to be larger than the constant. This isn’t a measurement limitation, it is a limitation imposed by the nature of the particle itself.
While speed and position are the most widely known pair of quantities that are governed by the Heisenberg Uncertainty Principle, they are far from the only pair of such quantities. For addressing the origins of the universe (yes, I am working my way back to that topic), the pair that matters the most is energy and lifetime. In the rest of this section I am going to describe something about how we can verify that the Heisenberg Uncertainty Principle applies to these quantities, and then in the next section will describe how this relates to the argument from first cause.
“Spectroscopy” is the science of measuring the light absorbed by or emitted by substances. Spectroscopy is particularly powerful because it is a direct probe of the energy states that a substance can take on. If a particular substance absorbs light with a wavelength of 532 nanometers (a green photon), we know that each photon being absorbed transfers 3.73 x 10-19 Joules of energy to the substance. By changing the wavelengths of light we are looking at, we can see that some energy changes within the substance seem to be allowed, but others are not. This brings up the notion of, “quantized energy levels,” which is the “quantum” in quantum mechanics. The allowed energy states in a substance must be thought of as a ladder or staircase rather than as a slope; the substance can be on the first rung or the second rung, but it cannot be on the 1.5th rung, because that energy isn’t allowed. Please note, I am not saying that we don’t allow it, I am saying that experimentally we observe that such an energy isn’t possible in that substance.
Once a substance absorbs a photon of light, it is in what is known as an “excited state,” meaning that it on a higher energy level than the bottom rung of the ladder, which is known as the “ground state.” In many cases, the energy gap between the excited state and the ground state is very well defined, which means that the photons that are emitted by the substance are very specifically of one energy. The emission spectrum is what we call “sharp,” because the emission energies are very sharply defined.
However, suppose we have a system that should have a sharply-defined emission spectrum, but where the excited state is extremely unstable, meaning that it will very quickly emit a photon to let the substance revert back to its ground state. The more quickly this happens, the shorter the “lifetime” of the excited state is. Experimentally we observe that as the lifetime of an excited state goes down, the less sharp the emission spectrum becomes. This makes perfect sense from the Heisenberg Uncertainty Principle, because the shorter the lifetime is, the more poorly defined the energy has to be. Thus we have direct experimental support for the idea that lifetime and energy are related by the Heisenberg Uncertainty Principle.
Vacuum Fluctuations and the Origins of the Universe
Imagine a complete vacuum, void of any matter or energy. The Heisenberg Uncertainty Principle for energy and lifetime suggests that particles of low mass can spontaneously pop into existence within that vacuum as long as they don’t survive very long. The most clear version of this would be electron/positron pairs that then annihilate each other within 10-43 seconds. Here ΔE is the uncertainty in energy (which we can view as mass by Einstein’s famous equation E=mc2) and Δt is the uncertainty in lifetime, and their product must be greater than a constant. This means that for sufficiently small times, we are inherently unable to specify the mass to be zero within a certain amount of mass, and therefore that much mass can in fact spontaneously pop into existence. Counterintuitive as this might be, it has been experimentally verified through the Casmir Effect to happen. This phenomenon is known as a vacuum fluctuation. Experimentally, something can in fact come from nothing.
Now, it is unquestionably an extremely long way to argue from vacuum fluctuations of electron/positron pairs to a vacuum fluctuation so large that it creates the big bang and thus the entire universe. After all, there is a lot of mass in the universe (currently estimated to be 1.5 x 1053 kg), and it has existed for a very long time (current best estimate is 13.8 billion years). However, remember that the Heisenberg Uncertainty Principle expression in question refers not to mass but to energy, where we understand some of that energy to be mass by the Einstein equation. There are lots of other kinds of energy in the universe that have to be taken into account, including gravitational potential energy. The thing is, gravitational potential energy is, when the mathematics is followed through carefully, formally a negative number. It is generally considered plausible that if we add up the total energy of the universe and appropriately express the gravitational potential energy as a negative value, that total energy is equal to or very close to zero. With a total energy of zero, the lifetime of the vacuum fluctuation can be enormous.
Craig’s first premise
Most “things” that we think of have very obvious causes. A baby is caused by its parents. A painting is caused by its painter. A watch is caused by its watchmaker.
But Craig is using the argument in a much more fundamental sense than these examples suggest. The baby, the painting, and the watch, are arrangements of atoms, every one of which existed well before the baby, painting, and watch were “made.” The “makers” made a new arrangement of those atoms, but no new atoms were made in the process.
But that’s not the type of “beginning” that Craig is talking about in his Cosmological Argument. He is applying the concept of cause to creation ex nihilo, or creation from nothing. He is trying to apply the argument to the origins of the universe, where matter, space, time, and even possibly the laws of physics themselves came into being.
As I discussed above, we have in fact observed one type, and one type only, of creation ex nihilo. And that one type is, according to the by far dominant interpretation of quantum mechanics, the Copenhagen Interpretation, is very literally uncaused. These are the virtual particle pairs (a.k.a., vacuum fluctuations).
So what we come to in analyzing Craig’s argument is that the only real examples we as a species have ever observed of something beginning to exist came into being without a cause. This means that not only should we doubt Craig’s first premise because there are exceptions to it, but we should reject it entirely because we have never seen a single case where it is correct, and have seen cases where it is incorrect.
Where does this leave us?
This argument suggests that a purely naturalistic explanation for the origins of the universe is possible that is consistent with current science. Certainly not all of the details have been worked out, and it may in fact be a different mechanism than vacuum fluctuations that turns out to be correct; there is, for example, a fascinating alternate scenario that suggests the precipitating event might be a fluctuation in space-time curvature. However, what the scenario I have outlined does do is demonstrate that the argument from first cause for the existence of God doesn’t work. Causality demonstrably doesn’t operate in the way that the argument assumes, at least within certain constraints that are relevant to the origins of the universe. Something can in fact come from nothing, again within certain constraints that are relevant to the origins of the universe. And most tellingly, the only cases of creation ex nihilo that have ever been observed have been uncaused. In short, the argument from first cause fails to demonstrate that a sentient entity is necessary to be a universal first cause.
I therefore declare the Cosmological Argument well and truly dead, at least until there is any experimental example of creation ex nihilo that proceeded with a cause.
- May 12, 2015: Revised to directly mention William Lane Craig’s formulation of the Cosmological Argument and to add the last scientific point.
- March 12, 2016: Corrected a capitalization error.