The universe is very likely flat because it was created flat. A flat universe starts with minimal information, which is fine since our cosmic creation process can create all the information that we have today. Gottfried Leibniz’ question, “Why is there something rather than nothing?” might be “the universe is made out of something and the exact opposite of that something.”

When I was a first-year graduate student in astrophysics at Harvard University in 1958, I encountered two problems that have remained with me all these years. One was the fundamental problem of information philosophy - “What creates the information structures in the universe?” The other was the flat universe.

At that time, the universe was thought to be positively curved and "closed." A particle traveling in a fixed direction would eventually return to its starting position. Edwin Hubble’s red shifts of distant galaxies showed that they did not have enough kinetic energy to overcome the gravitational potential energy, so they should all eventually collapse back together. Textbooks likened the universe to the surface of an expanding balloon decorated with galaxies moving away from one another.

That balloon popped for me when Walter Blade came to Harvard to describe his work at Mount Wilson. Baade took many images with long exposures of nearby galaxies and discovered there are two distinct populations of stars. And in each population there was a different kind of Cepheid variable star. The period of the Cepheid’s curve of light variation indicated its absolute brightness, so they could be used as “standard candles” to find the distances to star clusters in the Milky Way.

Baade then realized that the Cepheids being used to calculate the distance to Andromeda were 1.6 magnitudes brighter than the ones used in our galaxy. Baade said Andromeda must be twice as far away as Hubble had thought.

As I listened to Baade, for me the universe went from being positively curved to negatively curved. It jumped right over the flat universe! I was struck that we seemed to be within observational error of being flat. Some day a physicist will find the reason for perfect flatness, I thought.

I used to draw a line with tick marks for powers of ten in density around the critical density ρ_c to show how close we are. Given so many orders of magnitude of possible densities, it seemed improbable that we were just close by accident.

We could increase the density of the universe by thirty powers of ten before it would have the same density as the earth (way too dense!). But on the lighter side, there are an infinite number of powers of ten. We can’t exclude a universe with average density zero, which still allows us to exist, but little else in the distance.

In the long run we are approaching a universe with average density zero. Some say all the non-gravitationally bound systems will slip over our light horizon as the expansion takes more and more of them faster than the velocity of light. At this time, galaxies with a redshift greater than z = 1.8 are already over our light horizon. We can never exchange signals with them.

But note that we may always be able to see back to the cosmic microwave background, all the same contents of the universe that we see today will always be visible, just extremely red-shifted!

When Alan Guth presented his theory of inflation at Harvard as an explanation of how the universe appears to be flat, I asked him the simple question, "What if the universe has always been flat?" "That's too easy," he replied.

When I wrote to Steven Weinberg about the possibility of a flat universe he replied that flatness might be the simplest solution. Years later in his 2008 book Cosmology he wrote...

[W]e will see later that the total energy density of the present universe is still a fair fraction of the critical density. How is it that after billions of years, ρ is still not very different from ρ_crit? This is sometimes called the flatness problem.

The simplest solution to the flatness problem is just that we are in a spatially flat universe , in which K = 0 and ρ is always precisely equal to ρ_crit. A more popular solution to the flatness problem is provide by the inflationary theories... In these theories K may not vanish. and ρ may not start out close to ρ_crit, but there is an early period of rapid growth in which ρ/ρ_crit rapidly approaches unity.

Cosmology, p.39

The simplest solution to the flatness problem is therefore correct within observational error. There is nothing in general relativity theory that can explain that observational fact, but there is another simple observation that might throw light on the overall curvature of the universe.

We can start from the simple fact that the observable universe looks more or less the same in all directions.

As far back as 380,000 years after the Big Bang, when free electrons were combining with protons to form hydrogen atoms, light that was being scattered by the previously ionized gas (the last remnant of the original plasma) was now free to travel across the now transparent universe.

Visible light at that time had temperature 5000K (about the same as today's solar surface). It now appears to us as cooled down microwave radiation at 2.7K, as roughly predicted by George Gamow in the 1940's. This cosmic microwave background radiation (CMB) was first discovered by Arno Penzias and Robert Wilson in 1965 as faint noise coming in from all directions.

Now the latest sky surveys (see especially the Sloan Digital Sky Survey) show areas of the sky a bit more populated than others, and the CMB is marked with tiny regions that are slightly cooler than others (which may correspond to closer regions with fewer galaxy clusters), but overall the universe is remarkably uniform.

We can ask what the gravitational force would be on a particle of light or matter traveling through space today. If it's traveling near a large gravitational object, its path will be curved or bent toward that object as Einstein explained. But what if it's traveling in some part of space far from most masses?

If we averaged out the matter in all directions, the universe density would be uniformly ρ_crit in all directions, and the net gravitational force on our test particle would be zero!

We can now see that a spherically symmetric universe would exert no net force on a test particle. It would travel in a straight line. It would experience no spatial curvature!

We conclude that the average curvature of the universe is simply zero, except in the neighborhood of large gravitational masses. When Walter Baade's observations in the 1940's suggested that the universe was not closed and finite as previously thought, the universe first could be seen as open and infinite in all directions.

Normal | Teacher | Scholar