According to the standard Big Bang model, the universe was born during a period of inflation that began about 13.7 billion years ago. Like a rapidly expanding balloon, it swelled from a size smaller than an electron to nearly its current size within a tiny fraction of a second.

Initially, the universe was permeated only by energy. Some of this energy congealed into particles, which assembled into light atoms like hydrogen and helium. These atoms clumped first into galaxies, then stars, inside whose fiery furnaces all the other elements were forged.

This is the generally agreed-upon picture of our universe's origins as depicted by scientists. It is a powerful model that explains many of the things scientists see when they look up in the sky, such as the remarkable smoothness of space-time on large scales and the even distribution of galaxies on opposite sides of the universe.

But there are things about it that make some scientists uneasy. For starters, the idea that the universe underwent a period of rapid inflation early in its history cannot be directly tested, and it relies on the existence of a mysterious form of energy in the universe's beginning that has long disappeared.

"Inflation is an extremely powerful theory, and yet we still have no idea what caused inflation—or whether it is even the correct theory, although it works extremely well," said Eric Agol, an astrophysicist at the University of Washington.

For some scientists, inflation is a clunky addition to the Big Bang model, a necessary complexity appended to make it fit with observations. Nor was it the last such addition.

"We've also learned there has to be dark matter in the universe, and now dark energy," said Paul Steinhardt, a theoretical physicist at Princeton University. "So the way the model works today is you say, 'OK, you take some Big Bang, you take some inflation, you tune that to have the following properties, then you add a certain amount of dark matter and dark energy.' These things aren't connected in a coherent theory."

"What's disturbing is when you have a theory and you make a new observation, you have to add new components," Steinhardt added. "And they're not connected … There's no reason to add them, and no particular reason to add them in that particular amount, except the observations. The question is how much you're explaining and how much you're engineering a model. And we don’t' know yet."