⁂ Entropy

Counting, not mess

Boltzmann's formula, S = k ln W, is carved on his grave: entropy S grows with W, the number of microscopic arrangements consistent with what you see. 'Disorder' is a misleading shorthand — oil and water separating, or crystals forming, still raise the total entropy once the heat they release is counted. A better picture: energy spreading into more of the ways it can be shared. Established

Words people mix up

Six words that often get used for each other — in popular writing, textbooks, and even by physicists speaking loosely. Each has one job. Established

Two examples of the mix-up. The universe cooling as it expands is a fall in temperature, not entropy: the early light keeps its count; each photon just carries less energy. And a star heating up as it loses energy is about heat capacity, not entropy. "Disorder", "randomness" and "dispersal" are analogies for entropy, not definitions.

Is entropy a thing?

Not the way gravity is. Entropy is a count, not a force or a substance: how many arrangements look the same from outside. Nothing pushes a gas to spread; there are just overwhelmingly more spread-out arrangements for it to wander into. The second law is odds, not a push. Established

A still gas has entropy, but entropy only means something for time when states change: the arrow needs a before and an after to compare. A metre is still a metre if time stops; the arrow of time isn't.

Keep a second word for potential. Physics calls it free energy: the part of a system's energy still available to do work — its energy minus its temperature times its entropy. For a system left to itself at a steady temperature, as entropy rises, free energy falls. A stretched rubber band pulls back because letting go frees its long molecules into far more arrangements; the pull is a free-energy gradient, and physicists call it an entropic force. Two words, two jobs: free energy is what can still happen; entropy counts how spread out things already are. Established

That's also why "gravity from entropy" isn't absurd: if rubber bands can pull for statistical reasons, perhaps gravity does too. It's unproven — see Gravity as thermodynamics. Contested

Why it grows

Most arrangements look 'spread out', so a system wandering at random among its arrangements almost always ends up there. For a handful of particles, backward steps happen and have been measured (fluctuation theorems); for the 10²³ molecules in a room, they're never seen. Established

A subtle point: the count depends on how coarsely you describe the system — which details you ignore. Physicists still discuss what that means for how objective entropy is. Contested

It matters most when comparing. A fair comparison keeps three things fixed: the same whole, the same forces, the same level of detail. "The early universe had low entropy" compares the whole observable universe then with the most the same matter could hold, with gravity switched on. Change any of the three mid-argument and the comparison stops meaning anything. Established

Information costs energy

Entropy and information are two sides of one coin. Landauer (1961) showed that erasing one bit of information must release a minimum amount of heat — measured in 2012 and, in 2025, in a quantum many-body system. Established

Life runs on an entropy flow

Earth receives sunlight as relatively few high-energy photons and sends the same energy back to space as many more low-energy infrared photons — roughly twenty for every one received. Plants, weather and people live in that gap, building order locally while total entropy rises. Established

Black holes and the universe

Black holes turn out to carry enormous entropy, set by the area of their horizon (Bekenstein and Hawking, 1970s) — widely accepted theory, not yet tested by experiment. They hold most of the entropy in the universe today. The smooth early universe, by contrast, had very little of its entropy in black holes or clumps — which is why everything since has had room to run downhill. Established (the smooth start) · Contested (why it was so).

One entropy, two ledgers

Physicists mean one entropy, but it has parts that behave very differently, and mixing them up causes most of the confusion about the early universe. Two matter here. The thermal part: how energy is shared among the particles' motions and the light. The gravitational part: how the matter is arranged, when gravity is pulling it together. Analogy as a way of dividing the bookkeeping: for black holes the entropy has an agreed formula, but for gravity in general it doesn't — see the Weyl curvature hypothesis Contested.

At the start the thermal part was nearly full and the gravitational part was nearly empty. Almost all the growth since then has gone into the gravitational part. That empty headroom is what "low entropy early universe" means.

How we know the thermal part was full: the microwave background has the most perfect black-body spectrum ever measured, the mark of light and matter in thermal balance. How we know the gravitational part was empty: the same light is smooth to about 1 part in 100,000. The photon and neutrino entropy set in the first minutes has barely changed since; nearly all the growth is clumping, and above all black holes. Even today the universe is nowhere near its maximum. Established for the inventory (light and neutrinos then; black holes now); the numbers are estimates to the nearest power of ten, for the observable universe.

Which viewpoint? The trap

The trap is switching viewpoint mid-argument.

• With gravity ignored, a smooth gas is high entropy, like a box of air.
• With gravity on, the same smooth gas is low entropy, because it wants to collapse.

Both are right; they answer different questions. Between a few atoms, gravity is far too weak to matter, so in a room spreading wins. Over enough mass it adds up — above the Jeans mass, a cloud's pull beats its pressure — and then clumping wins. The clump looks more ordered, but it heats up as it falls in and radiates, and the light it sends out carries away more entropy than the tighter arrangement gave up. Count the whole, with the same forces on, and the total still rises. Established

See both in the Entropy box.

Why a star heats up as it loses energy

Most things cool when they lose energy. A star, or any cloud held together by its own gravity and close to balance, does the opposite: it has negative heat capacity. As it radiates, it shrinks a little; falling inward speeds its particles up, and faster particles are a hotter gas. The virial theorem fixes the split: of the energy released by shrinking, half heats the gas and half leaves as light. Established

That makes gravity the one force that grows temperature differences instead of evening them out: lukewarm gas becomes hot stars in cold space. It still makes no new potential; it spends the smoothness. And two such systems in contact never settle into balance: the hotter one loses heat and gets hotter still (Lynden-Bell & Wood, 1968). Established