Demystifying the “Arrow of Time”

The “arrow of time” is often presented as one of the great mysteries of physics. The argument usually starts from the observation that the fundamental laws relevant to ordinary macroscopic processes are largely invariant under time reversal, while the world around us seems to evolve in a preferred direction. A glass falls from a table and shatters, but we do not see the pieces spontaneously assemble themselves and jump back onto the table. Entropy is then invoked to explain why time itself has a direction.

There is considerably less mystery here once several different statements are separated.


Consider the simplest possible example: an object moving through the space. At some moment it has a particular position and velocity. Given this state, the laws of motion tell us how it will evolve. If it is moving to the right, it will continue moving to the right. And the notion of entropy has nothing to do with it.

This distinction between time-reversal symmetry of the laws and the state is very important. The laws governing a system may possess a symmetry that a particular state does not. Rotationally invariant laws, for example, do not prevent the existence of chairs. The state selects the direction while the underlying laws remain rotationally invariant. There is nothing fundamentally different about time reversal.

This is in fact the essence of the spontaneous symmetry breaking, which is one of the central concepts of fundamental physics.


Where, then, does entropy enter the story?

Entropy becomes relevant when we stop following every microscopic constituent of a system and describe it using a smaller number of macroscopic variables. An intact glass corresponds to a very particular organization of an enormous number of atoms. There are vastly more microscopic configurations corresponding to what we would simply call a shattered glass. Once the glass breaks, a spontaneous return to the original organized state is therefore extraordinarily improbable.

It is not forbidden, however. If we could take the complete microscopic state immediately after the glass shattered and reverse all the relevant motions with perfect precision, time-reversal-invariant microscopic laws would evolve the system back toward the intact glass. The pieces would come together and the glass would return to the table. Preparing such a state is absolutely unrealistic, but the fundamental dynamics does not prohibit it.

The apparent irreversibility arises because we do not prepare such extraordinarily special microscopic states. We describe macroscopic systems only approximately, ignoring almost all of their microscopic details. For a typical microscopic state compatible with that coarse description, evolution toward macrostates that can be realized in many more ways is overwhelmingly more likely. This statistical tendency is what we describe as increasing entropy.

This also explains why examples such as breaking glasses, mixing gases or melting ice should not be confused with evidence that time itself only permits evolution in one direction. We can and do observe processes in which particular quantities increase or decrease depending on the state of the system. What we can say is that for sufficiently large systems, evolution toward higher total entropy is overwhelmingly typical under the conditions in which macroscopic systems are normally prepared.

There is nevertheless an interesting question hiding underneath the usual discussion. If increasing entropy characterizes the history of our universe, then tracing that history backwards leads us toward a state of lower entropy. The early universe must therefore have been in a very special state compared with the enormous number of states that would have been possible.


Why the universe had such an initial state is a legitimate question. It is ultimately a question about cosmological initial conditions and perhaps about physics we do not yet understand. But it is not the same as asking why time has a direction. The laws provide the dynamics, while the state of the universe determines which particular history is realized.

The so-called arrow of time therefore does not require entropy to push time forward. Entropy characterizes a statistical property of the macroscopic evolution of our universe. Time-reversal invariance tells us that the underlying laws also permit appropriately reversed histories. There is no contradiction between the two.

What remains to be understood is not why time knows which way to go, but why the universe began in the particular state from which the thermodynamic history we observe could unfold.


See also: What is time?
How does complexity arise?


© 2026 Physics and Reality. All rights reserved.
About