Scientists love running simulations. Climate modelers try to simulate how Earth will respond as carbon dioxide levels increase. Particle physicists simulate collisions of particles to understand measurements made in the complex detectors surrounding their particle accelerators. During graduate school and postdocs, I used computers to simulate more gamma rays than I ever detected with telescopes. The prolific use of simulations through all scientific disciplines raises the question of whether we (humanity, Earth and all the life it contains, and even the universe) exist only as simulations in complex computer programs run by our advanced descendants. Remarkably, scientists can test this idea by making measurements of our universe.
Oxford philosopher Nick Bostrom was the first to develop the concept that humans might exist in a complex simulation run by our descendants.1 In his paper, he outlined three possible options: (1) the human species will likely go extinct before “advancing” enough to simulate a large universe; (2) any sufficiently advanced civilization is unlikely to run a significant number of simulations of human development; or (3) we almost certainly live in a computer simulation. Testing the third possibility requires knowledge of how to develop a simulation of the universe.
The first step in programming a good simulation requires modeling the physics that determines how things behave. The simulation operator must then build a grid or lattice that organizes where the calculations of the physical interactions happen. These lattice points are like the dots on a very large computer monitor that shows what happens and when. Typically, one envisions a static lattice, but more complex simulations permit elastic and even movable lattices.
The number and spacing of the lattice points sets the resolution and limits of the simulation. For example, consider a simulation of Earth’s atmosphere with two million lattice points. Simulating the entire atmosphere gives a distance around ten miles between lattice points (assuming an atmosphere ten miles high). This spacing limits the information regarding cloud cover because clouds are much thinner than ten miles. Reducing the spacing to less than a mile (more realistic for cloud cover) means that the simulation will cover only 200,000 square miles of Earth’s surface (less than the size of Texas).
It’s one thing to simulate Earth’s atmosphere, but quite another to simulate the many physical characteristics of the universe. Similar limits apply to the most fundamental simulations of the universe. Specifically, physicists model the behavior of the strong nuclear force using such a lattice and accurate results require a lattice spacing of fractions of a femtometer (10-15m—or the size of a proton). Current speed, storage space, and other computer limitations result in the largest simulations encompassing a spatial size of a few femtometers. However, as computer power and algorithm sophistication increase, the size of the simulations will also increase. Reasonable extrapolations of technological trends indicate that humanity could simulate volumes up to one meter in size (a much larger, but still very small chunk of the universe) within the next 150 years.
Of course, the hypothesized advanced civilizations capable of simulating our universe will possess much greater, but still limited, resources. Consequently, these civilizations are likely to rely on similar lattice systems for simulation-building.
Lattices impose a cubic symmetry on results, but “real world” quantities calculated by physicists assume a rotational symmetry (a cube only exhibits symmetry when rotated by specific angles, a sphere exhibits symmetry regardless of the rotation angle). Thus, determining which quantities are most sensitive to differences between rotational and cubic symmetry holds the most promise for assessing what physical measurements (aspects of the universe) might probe the lattice spacing.
According to recent research, two promising measurements include searching for discrepancies in the anomalous magnetic moment of the electron and muon as well as comparing the value of the fine structure constant using two different measuring techniques. However, the most sensitive method involves looking for asymmetries in the directions of the most energetic cosmic rays. In a grid-based simulation, these cosmic rays would travel diagonally across the grid rather than along the edges and they would preferentially interact in certain directions.2 Current measurements of the highest energy cosmic rays indicate no such asymmetries, but scientists expect operational experiments (like Auger) to substantially increase the database of cosmic rays available for analysis.
For now, the data provides no indications that we reside in a complex simulation. However, the possibility of simulated realities implies that they could exist—like a multiverse of simulations. I find it provocative that some students of the Bible consider this universe as a “testing ground” for the battle of good vs. evil and that we will enter a new realm upon completion of the battle. Maybe it is not so unreasonable to consider our universe one massive “simulation” with God as the Master Simulator.