Time’s Second Arrow: Evolution, Order, and a New Law of Nature Robert M. Hazen and Michael L. Wong W. W. Norton (2026)

Like evolution by natural selection, some ideas that start out as revolutionary can become so intuitive that they seem banal to future generations. Mineralogist Robert Hazen and planetary scientist Michael Wong must hope that the central idea behind their book becomes similarly mainstream.

Time’s Second Arrow explores how the natural world’s shift from humble beginnings to intricate order was inevitable. According to the authors, science has overlooked this “obvious truth” for too long. It is time to establish guiding principles that explain the complexity in all evolutionary systems — living and inert.

How the story of life and our planet unfolded — together

The authors open their case by outlining how laws of physics allow scientists to make predictions about how objects will behave anywhere in the Universe. The main rule in Hazen and Wong’s sights is the second law of thermodynamics. It describes how energy tends to dissipate over time, such that disorder (also known as entropy) always increases.

For example, a snowman melts because warmth loosens the hydrogen bonds that hold together its rigid ice crystals, turning the snow into messy liquid water. Similarly, when organic matter decays, its chemical bonds are broken and its atoms spread out into the soil, air and water.

This second law is often called time’s arrow because it lends all events a direction. A snowman’s fate on a warm day will always become a puddle, and summer’s green leaves all eventually turn brown. Hazen and Wong acknowledge this law. But they think that they have uncovered another: one that works in the opposite direction.

Natural selection of all things

Hazen and Wong ask scientists to recognize the fact that order has repeatedly sprung into existence from simple, randomly distributed beginnings. Some examples include the creation of elements through nuclear fusion, the formation of chemical compounds and even the evolution of intelligent life.

The tendency towards increased organization is, they contend, so ubiquitous that there “must be a second arrow of time” that explains “the universal tendency for certain systems to display increases in order”.

It’s an intriguing concept. Just as life on Earth has become more complex over time — developing from single cells to multicellular organisms — atoms, stars and minerals might follow their own evolutionary principles.

Hazen and Wong argue that order will inevitably arise from relative disorder as long as three prerequisites are met: the potential for many possible configurations, the capacity to create new forms and the means to select between them.

The secret relationships that govern life on Earth

Order can be measured, the authors suggest, using ‘functional information’ — which is estimated by comparing the number of forms that exist with the number of configurations that are theoretically possible. Functional information, and therefore order, increases when emergent forms occupy a smaller proportion of possible configurations.

For example, human language evolved out of many possible sounds, our ability to make noises and the utility of those noises for communicating ideas and things. Over time, only particular sounds caught on.

The book is engagingly written and remarkably thought-provoking. I can see generations of thinkers debating Time’s Second Arrow for years to come, and I imagine that the authors’ ideas will go on their own evolutionary journey.

Their proposal is grounded in some peer-reviewed literature, but it is by no means a slam dunk. There are counterarguments, and the authors admit that they have not found “a satisfying solution” to all of them.

What drives complexity?

One of the thorniest problems arises from the second law’s requirement that disorder increases overall. This must remain true even when local order increases — but how? Is increasing orderliness merely an eddy on the river of entropy’s merciless flow?

To address such questions, scientists would have to measure the functional information of natural systems, but this is not straightforward and leads to another problem.