Nobody was there, and nobody ever will be. Everything on this page was worked out backwards, from things that can still be measured today. So this is not a story about what happened — it is an account of what is known, what each piece of evidence proves on its own, and the several large places where the honest answer is that nobody knows.
Almost everybody pictures the same thing: a dot in the middle of an empty blackness, going bang, and bits flying out into the space around it. Every part of that picture is wrong, and holding onto it makes the rest of the page impossible to follow.
There was no dot sitting in space, because there was no space around it to sit in. There was no middle, and there is no middle now. Nothing flew out into anywhere. What happened is that space everywhere started getting bigger, and it has been getting bigger ever since.
Which means the Big Bang did not happen at some far-off spot you could point to. It happened here. It happened in the room you are sitting in, and equally in every other room everywhere, all at the same moment. Everywhere is where it happened.
The gold dot on the right is not special — it is just the one we happened to stand on. Everything runs away from it, and the farther off it is the faster it goes. Now put yourself on any other dot and redraw the arrows: you get exactly the same picture. Every dot is equally right about being the one everything is fleeing, and none of them is the middle. That is not a trick of the drawing; it is the actual situation, and there is a balloon at the bottom of this page that will let you check it for yourself.
The measurement itself belongs to rung 7 of the distance ladder: every distant galaxy's light arrives stretched towards the red, and the farther away it is, the more stretched it is. Space has been getting longer the whole time the light was on its way. That page has the rubber band you can do at the kitchen table; there is no need to repeat it here.
What this page does is run it the other way. If everything is getting farther apart, then yesterday it was closer together, and a billion years ago closer still. Squeeze a gas and it gets hotter — you can feel this in a bicycle pump. So the further back you go, the denser and the hotter it gets. Keep going and you reach a moment, a little under fourteen billion years ago, when everything that is now spread across the sky was packed together and unimaginably hot.
Less than you would think. For thirty years there was a serious rival idea, called the Steady State, which accepted the expansion completely and still had no beginning at all. Its answer was that a little new matter quietly appears as space grows, just enough to keep everything looking the same forever. An expanding universe with no start is a perfectly consistent thing to believe, and clever people did believe it. So expansion on its own does not prove a beginning. It only tells you that things used to be closer together.
Running the expansion backwards gives an age, and when it was first done in 1929 the answer came out at about two billion years. That was an embarrassment, because rocks on Earth had already been dated to more than three billion. The universe appeared to be younger than the ground. It took another thirty years of better distance measurements to find the mistake — the galaxy distances were far too small — and the age grew to what we now think it is, about 13.8 billion years. Worth remembering when a number on this site looks final.
In 1964 two engineers, Arno Penzias and Robert Wilson, were testing a very sensitive radio antenna in New Jersey, and they could not get rid of a faint hiss. They checked the wiring. They pointed it in different directions and got the same hiss. They waited for a different season, in case it was something in the sky, and got the same hiss. They found pigeons nesting in the antenna, evicted them, and scrubbed out the mess. The hiss stayed exactly where it was.
It was not a fault. The hiss is real, it comes from every direction equally, and it is the oldest thing anyone will ever detect.
For its first few hundred thousand years the universe was too hot for atoms to hold onto their electrons. Loose electrons bounce light around, so the whole universe was an opaque glowing fog — not dark, but you could not see through it, the way you cannot see through a cloud even in daylight.
Then it cooled to about 3,000 °C, which is cool enough for electrons to be captured and held. The fog cleared — everywhere, over a fairly short stretch of time — and all the light that had been bouncing about suddenly had nothing left to bounce off. It flew free. It has been flying ever since, and some of it is arriving here now.
The light left as a fierce orange-white glare. Stretch a wave a thousand times and it stops being light you can see and becomes a microwave — and a sky full of microwaves is a cold sky. This one measures 2.7 degrees above absolute zero, the coldest temperature there is. That number is not a separate discovery; it is what you get when you take the glare of the fog and stretch it by the amount the universe has grown.
This one is much harder to argue with. A glow of exactly this kind, arriving equally from every direction, is extremely difficult to produce any way except a hot dense past that cooled. The Steady State idea had no way to make it, and that is essentially what ended the argument — not a debate, but a hiss in an antenna.
But notice what it cannot do. You cannot see through a fog, and this is the fog. The microwave background is a wall: no telescope that has ever been built or ever will be built can see light from before it, because there was no light travelling freely before it. Everything earlier has to be worked out some other way.
One more thing about it, which is true and slightly absurd: an old analogue television tuned between channels showed a screen of snow, and a small part of that snow — a per cent or so — was this glow. The oldest light there is, turning up as static in a living room.
Take the ordinary matter anywhere — the Sun, a galaxy so far away that its light is billions of years old, a cloud of gas that has never got round to making a star — and weigh what it is made of. You get about three quarters hydrogen and one quarter helium, with everything else in the leftovers. Not roughly. The same, everywhere anyone has looked.
That is strange, because stars make helium out of hydrogen, so you would expect the proportion to differ from place to place depending on how much star-making has gone on. It does vary a little. But the variation is small, and two things give the game away: all the stars that have ever existed cannot account for more than a few per cent, and the very oldest stars — made from the least-processed gas anywhere — already started life with about 24% helium.
So nearly all of it was made before there were any stars. There is only one moment available: the first few minutes, when the whole universe was as hot as the middle of a star is now. Hydrogen fused into helium everywhere at once. Then the expansion cooled things below the temperature where fusion works, and about twenty minutes in, it stopped for good. The proportions have been fixed ever since.
A hot, dense first few minutes — and, because the exact recipe depends very sensitively on how dense things were, a rather precise measurement of that density. It is the sharpest of the three pieces, and it also predicts how much of a couple of rarer ingredients there should be. Deuterium, heavy hydrogen, matches the prediction beautifully.
What it cannot tell you is anything about what came before those minutes, or what has happened since. On its own it is a snapshot of twenty minutes, with no before and no after.
The same calculation predicts how much lithium there should be, and the amount actually measured in old stars is roughly three times too small. Nobody has resolved this. It has been open for decades, and it is quietly listed in every honest account of the subject, including this one. It is not enough to overturn the picture — hydrogen, helium and deuterium all match — but it is not nothing either.
Put them side by side and the shape of the argument becomes clear. Each piece of evidence, taken alone, leaves a door open. It is the overlap that closes them.
That is what "we know" means here. Not certainty, and not a proof — one surviving explanation, and a graveyard of alternatives that each failed on at least one of the three. If a fourth measurement turned up tomorrow that this picture could not accommodate, the picture would have to go, and everybody in the field knows it.
The three pieces above do not all reach equally far back, and it is worth being clear about which part of the story rests on what.
You need a balloon and a marker pen. Rung 7 has the other half of this: there you split a rainbow off a CD and stretch the lines in it, which is what expansion does to light. This one is what it does to places, and it is the question at the top of this page — where did it happen?
Now do the part that matters. Pick a completely different dot, pretend that one is home, and look at the balloon from its point of view.
So which dot is the real centre? None of them. There is no point on that surface that anything expanded away from — the whole surface simply got bigger. And that is the answer to "where did the Big Bang happen?". It happened at every dot, including yours.
Keep hold of one thing though: the balloon's surface is the model, not the balloon. The air inside and the room outside have no counterpart in the real universe, and going looking for them is exactly the mistake at the top of this page.