Almost all the laws of physics are time-symmetric; play the movie of the universe backwards and everything still obeys the laws of physics. For the few laws that aren't exactly time-symmetric, a true symmetry can be made by flipping charges (positive to negative) and mirror-imaging the universe, all at once. If you flip charge and time, while putting the universe in a mirror, all the laws of physics are the same. (This is known as CPT-symmetry).

But wait, you say; backwards movies look very different than real life. No one has ever seen a shattered teacup suddenly reassemble itself. No one has seen an oak tree shrink down into an acorn. So what does it mean to say the laws of physics are time symmetric?

As many people know, this "directional" aspect to time appears because Entropy, a statistical measure of disorder, always increases. But doesn't that mean that entropy obeys a time-asymmetric law of physics? No -- entropy is simply the statistical view of microscopic processes which are all time-symmetric. From this perspective, one would imagine that entropy should be time-symmetric as well. So why does it always increase?

The commonly-accepted solution to this paradox is that the laws of physics are time-symmetric, but the boundary conditions of the universe are not. The Big Bang, which started the universe off, was a very low entropy boundary condition (for reasons which aren't at all well understood). A time symmetric entropy law says this: entropy must always increase away from any low-entropy boundary condition -- in BOTH time directions! But there was no time before the Big Bang (only after the Big Bang), so therefore entropy increases away from the boundary condition at the beginning of the universe. Paradox solved.

But, still, you may ask, why would this be asymmetric? If all the laws of physics are time-symmetric (or CPT-symmetric, to be more accurate), then why wouldn't the boundary conditions of the universe also be CPT symmetric? Instead, modern physics assumes there is an initial boundary condition of the universe, but no final boundary condition.

In 1958, Thomas Gold proposed that perhaps the symmetry of the laws of physics also applies to the boundary conditions of the universe. [T. Gold, Am. J. Phys. 30, 403 (1962)] That would mean that the universe would start from a low-entropy initial state, expand to a high-entropy middle-state, and then contract to a low-entropy final state! The arrow of entropy would change direction in the middle of the universe!

This was an intriguing idea, and many people bought into it for awhile. Even Stephen Hawking believed this to be true when he wrote his book "A Brief History of Time". But he has since changed his mind, as have most other physicists; the concept which Gold presented simply has too many logical flaws.

There things sat for awhile, until a Philosopher named Huw Price wrote a book about time; [H. Price, "Time's Arrow and Archimedes' Point," Oxford Univ. Press, New York (1996).] In this book Huw Price briefly mentions another possibility of a time-symmetric universe, which he calls the "mixing model", and distinguishes it from Gold's model, which he calls the "meeting model". But the distinction is the end of the discussion; Price does not speculate on any consequences of this new model, perhaps because he is not a physicist.

My paper discusses in some detail what such a universe might look like and how to experimentally test the cosmology. Without going into the math, here's an analogy:

A footrace can be symmetric in two very different ways, even if all runners must begin at the start line and end at the finish line. The first possibility is to have the runners begin together, diverge from each other in the middle, and then run in such a careful manner as to all converge together at the finish line. The second possibility is to have two equal groups of runners. One group must start together and can finish independently (a usual footrace), while the other group can start independently but must finish together (the time-reverse of a usual footrace). [The mixing model] proposes the second option.

Instead of putting two boundary conditions on everything in the universe -- the source of the difficulties in Gold's model -- this new model only puts ONE boundary condition on any given particle in the universe. Half of the universe is "ordinary" matter, with an INITIAL boundary condition. The other half is something that is not known to modern physics -- it has a Final Boundary Condition (FBC), but no initial boundary condition.

So what are the implications? After all, if the above theory is correct, where's the other half of the universe? How could we detect them? What cosmological effects would they have? Have the proper experiments been performed and if not what experiments could shed light on this topic?

To answer these questions requires a lot of tough thinking about time that isn't easy for us IBC humans who are used to thinking about time progressing in only one direction. Consider how you might detect a a time-reversed star. If you look at the star, you don't see anything, because the photons are leaving your eyes and going to the star. But if you look in the opposite direction from the star, you still don't see anything, because now the photons are emerging from the back of your head and going to the star. Instead, you need indirect techniques to detect this "future-boundary" radiation; normal cameras are designed to detect photon absorptions, not photon emissions. But there are indirect ways to see this radiation, and in the paper I propose an actual experiment which could be done with present-day technology.