Where Did Galaxies Come From? In Cosmology We Can See the Starting Conditions

A theory of how something evolves needs to know where it started. Almost always, the starting point is the hard part: you cannot see it, so you reconstruct it, idealize it, or leave it as a knob to tune. Cosmology has a rare gift. The starting conditions for all the structure in the universe are not guessed at. They are on display, photographed in the cosmic microwave background.

What is the cosmic microwave background?

That background is the light set free when the young universe first turned transparent, about 380,000 years after the Big Bang. The faint differences in its temperature from place to place across the sky are a near-direct picture of the differences in density that existed back then, the very seeds that later grew into galaxies. What COBE actually pinned down is statistical: not the exact map of where every lump sat, but the amplitude and pattern of the ripples. In 1992 the COBE satellite, with George Smoot's team, measured those temperature differences at about one part in a hundred thousand (a fractional temperature change, which on these large scales tracks the density seeds up to a known factor), with a pattern across scales consistent with the same strength at every scale. This is not a starting spectrum someone assumed to make the numbers work. It is a measured one: its amplitude is pinned down tightly, its scale dependence more loosely.

Why a photographed start changes the science

That changes what we actually know. Because the law that acts on those seeds is also known, gravity amplifying the denser spots as the universe expands (Jeans, 1902; Peebles, 1980), the later arrangement of matter becomes a genuine prediction from observed starting data, not a fit with the start left free. Unlike most of physics, here the input (the starting conditions) and the law are both fixed independently of each other, and the output is checked against a different observation, the galaxy surveys. That is a stronger position than most of physics gets to stand in. The seed is a measurement, and cosmic structure is what follows from it.

It is worth noting how unusual it is to observe a system's beginning rather than infer it. It works here because light travels at a finite speed, so looking far is looking back. Whether anything like that generalizes is a separate question I am setting aside. The solid claim is narrow: the seed of cosmic structure is seen, not supposed.

Can we test it?

It is testable, with no wiggle room in the start itself. The lumpiness of today's universe should follow from the measured primordial ripples run forward through gravity, using the rest of the cosmic ingredients (how much matter, how much dark energy, how fast the expansion) pinned down by other measurements. The seed stays fixed at its measured value; any mismatch has to be charged to those ingredients, not to the starting ripples. The claim fails if no allowed set of those independently measured ingredients can run the fixed seed forward into the galaxy pattern we actually see. They line up, and that agreement is the account passing its test.