The Origin and Fate of the Universe

Core idea

The hot big bang model tells you what happened from one second after the beginning onward: cooling, nucleosynthesis, recombination, galaxy formation, stars, planets, us. It does not tell you why the universe started so hot, so uniform, so finely tuned, or with the particular fluctuations that became galaxies. Inflation answers most of those how-questions by inserting a brief phase of accelerating expansion that smooths the universe and amplifies quantum noise into structure. The deeper why-question — what set the initial conditions — Hawking and Hartle reframe with the no-boundary proposal: in imaginary time, the universe has no boundary at all. There was no moment of creation, because there was no edge to time. The universe simply is.

Hawking's argument: Asking what set the initial conditions of the universe presupposes that the universe has a boundary. But if quantum gravity allows space-time to be finite in extent yet without any edge — as the surface of the earth is finite without an edge — then the question of initial conditions evaporates. "The boundary condition of the universe is that it has no boundary."

Why it matters

The hot big bang has held up, but it begs questions

In the standard picture the universe cooled from infinite temperature at zero size. By one second it was at ten billion degrees and full of photons, electrons, neutrinos, and a few protons and neutrons. By a few minutes, fusion had locked about 25% of the protons and neutrons into helium — a prediction (Alpher, Bethe, Gamow, 1948) confirmed to high accuracy by the abundance of helium we measure today. A few hundred thousand years later, electrons combined with nuclei to form neutral atoms, releasing the photons we now see as the 2.7 K microwave background (Penzias and Wilson, 1965). The model fits — but it leaves four puzzles. Why was the universe so hot? Why is it so uniform across regions that could never have been in causal contact? Why was the expansion rate so absurdly close to the critical value that even today, ten billion years later, it has not recollapsed or thinned to nothing? And what set the tiny density fluctuations that grew into galaxies?

Inflation as a single answer to four puzzles

Alan Guth's 1981 proposal: in the first 10⁻³⁵ of a second, the universe went through a phase of exponential, accelerating expansion. Its size grew by a factor of roughly 10³⁰. The expansion was driven by an effectively false vacuum — a configuration of fields with positive constant energy density that acts like a cosmological constant, repulsive rather than attractive. Inflation smooths irregularities away (any region observed today was once tiny and in thermal equilibrium), drives the expansion rate to within a hair's breadth of the critical value (because in an exponentially expanding space, any deviation from flat is exponentially diluted), and quantum fluctuations during the inflationary epoch get stretched to cosmological scales — exactly the seeds galaxies later need. Guth's original "old" inflation had a technical problem (the bubbles of broken symmetry could not coalesce); Linde, Steinhardt, and Albrecht patched it with "slow-roll" inflation; later Linde proposed the simpler "chaotic" version that survives today.

Imaginary time and the sum over histories

To do quantum gravity at all, you need Feynman's sum over histories — every possible space-time geometry contributes to the probability of any observable. Working in real time, those sums diverge. The mathematical trick is to rotate the time coordinate by 90° to imaginary time, where the calculation becomes well-behaved (in this rotated space, time looks just like another spatial direction). The trick is not physical so much as mathematical, the way you can integrate a real function by stepping out into the complex plane and then coming back. But the geometry that solves the equations in imaginary time turns out to have a striking feature: it can be finite in extent and yet have no boundary or singularity at all — the way the surface of the earth is finite in area but has no edge.

The no-boundary proposal

Hawking, working with Jim Hartle in 1983, proposed that the universe's quantum state is given by summing over all such finite-without-boundary Euclidean geometries. In this picture, the universe in imaginary time looks like the surface of a globe: it starts at a "north pole" that is perfectly smooth — no singularity, no edge, the laws of physics hold there exactly the same as anywhere else. In real time, the rotation back from imaginary time produces a universe that begins in a smooth, ordered, low-entropy state, expands inflationarily, and eventually behaves like the hot big bang model with all the right initial conditions baked in for free. The no-boundary proposal makes a quantitative prediction — the tiny temperature fluctuations in the microwave background — which COBE detected in 1992 in agreement with the theory. It is therefore not just metaphysics but a falsifiable scientific claim.

Where this leaves the question of a creator

If the universe has no boundary in imaginary time, there is no moment of creation to be created — no edge for an external act to bring into existence. The cosmological argument for a creator (something must have started it) loses its grip, because "starting" presupposes a boundary the no-boundary proposal does not have. Hawking is careful to call this a proposal, not a proof; it is testable by what it predicts about the present universe, not by direct inspection of imaginary time. But the implications for the theology Hawking grew up around are pointed and intentional, and he ends the topic on the line: "What place, then, for a creator?"

Key takeaways

Mental model

Copy sourceZoom

Practical application

Distinguishing what is solid from what is provisional

When you read cosmology coverage, sort claims by how far back they probe. Anything about hydrogen, helium, deuterium, and lithium abundances is constrained back to one second and is unlikely to move. Anything about the CMB temperature, polarisation, and large-scale structure goes back to a few hundred thousand years and is also robust. Claims about inflation — when, how long, by what mechanism — are well-motivated but the specific model is not settled. Claims about a multiverse, the wave function of the universe, or eternal inflation are theoretical extensions that may or may not be testable. The further past one second you go, the less direct the evidence becomes.

How to think about "before the big bang"

If the no-boundary proposal is right, the question is mis-formed. "Before" requires a time coordinate that does not exist outside the universe. The north pole of the earth has no point that is "north of" it. This is a useful pattern to notice generally — when an apparently profound question depends on a coordinate (time, space, identity) that does not extend the way you assume, the question dissolves rather than gets answered.

Reading inflation news

Inflation predicts a near-flat universe, near-Gaussian density fluctuations with a slight tilt toward larger scales, and a faint background of primordial gravitational waves. The first two have been confirmed to high precision (Planck mission, 2013–2018). The third has not — the BICEP2 claim in 2014 turned out to be galactic dust. Detecting primordial gravitational waves would be the next major confirmation of the inflationary picture; their absence at improving sensitivity is gradually constraining which models of inflation remain viable.

Example

Consider the cosmic microwave background fluctuation map released by Planck in 2013. To the eye it looks like noise on a uniform sphere — temperature variations of about one part in 100,000 around a mean of 2.725 K. Decompose that map into spherical harmonics and plot the power at each angular scale, and you get a spectrum with characteristic peaks and troughs. The position of the first peak fixes the spatial curvature of the universe — flat to within half a percent. The relative heights of the peaks fix the baryon-to-photon ratio, dark matter density, and the spectrum of the original fluctuations. The slight tilt of the spectrum (slightly more power at large scales than at small) is the inflationary signature.

What is striking is that all of these numbers — curvature, expansion rate, fluctuation amplitude, spectral tilt — are what the simplest inflationary models predicted before they were measured. The no-boundary proposal goes one further and predicts that the fluctuations should be the minimum allowed by the uncertainty principle, with a specific shape. That prediction, too, is consistent with the COBE, WMAP, and Planck data. None of this proves the no-boundary proposal — there are other inflationary scenarios that fit equally well — but it shows the proposal is doing real scientific work, not just rhetorical.

Related lessons