I came across the question after watching a video (Heat Death) regarding the heat death of the universe and my question serves as such:

I understand this universe is constantly going from a state of order to chaos i.e., a state of high entropy to lower entropy and in such a way I understand that entropy is the probabilistic measure of order in a closed system within a certain window of time, but this personally brings up a lot of questions in the philosophical sense.

  1. It is mentioned in the video that the origins of the universe had high entropy, what caused such high entropy if it seems like the universe is going towards a state of low entropy, in our current time? Is it God? Is this the only indication/proof we have for the beginning of the universe? What if we didn't have an understanding of entropy which will in turn might affect our understanding of the Cosmic Microwave Background, would we not have been certain as to the existence of the 'beginning' of the Universe?

  2. I understand the intuitive reason as to why scientists claim that we move from order to chaos but can one not have a negative perspective on the words and so consecutively attribute higher entropy to chaos and lower entropy to order, in where one can view the spreading out of energy throughout a given system could be said to be "fair/orderly and uniform" and the hoarding of energy to a given local would be "selfish/chaotic and disordered". I understand that this is highly anthropomorphic but why wouldn't such descriptors work?

In such a way instead of viewing the heat death as something that is quite chaotic, one could view it as uniform and spread out equally, poetically speaking "peaceful, balanced and uniform". Is this a viable/justifiable view to have?

P.S. I understand it is a lot of questions but an answer to a few of my questions will suffice. Thank you.

  • 3
    This question pushes on one of the problems of plugging "physicalism" into the hole left ontologically when Einstein refuted materialism. Many of the key aspects of physics, such as energy, entropy, and information, are not "material objects" or anything like them. They instead seem to fall into the family of abstract objects. Physics, therefore, is basically ontologically dualist, with both material and abstract idealist features. These are Popper's worlds 1 and 3, as opposed to the world 1 and 2 dualism of Descartes. And if physics is dualist, then physicalISM is dualist too!
    – Dcleve
    Mar 11 at 0:28
  • 7
    "Entropy is the probabilistic measure of order" (more commonly - disorder) and "universe is constantly going from a state of order to chaos" are both wrong, those are just common pop-culture misconceptions. See e.g. Michaelides or Denbigh:"entropy increase... and increase of 'orderliness', increase of 'organization' or of 'complexity'... are far from meaning the same thing, but have all been supposed... to be contraries to the process of entropy increase."
    – Conifold
    Mar 11 at 0:42
  • 1
    "What it really is in its essence" is nothing grand or pop-culturally interesting, it measures how typical some classes of states are given a probability distribution. That the universe "is going to" more typical states says nothing about what those states might be like without much other information about it. Styer in Entropy as Disorder: History of a Misconception illustrates it with stacks of pennies, salad dressings and dot patterns to show how entropy values come out differently from intuitive expectations of order and disorder.
    – Conifold
    Mar 11 at 1:08
  • 1
    @Howwhye The change of entropy is a concept from physics, it is defined via the physical quantities heat and temperature. What do mean by asking whether entropy is "idealistic"?
    – Jo Wehler
    Mar 11 at 7:04
  • 6
    Re "I understand this universe is constantly going from a state of order to chaos i.e., a state of high entropy to lower entropy": This is just a terminology issue you have got backwards -- entropy is a measure of chaos, not order. The universe (or any "closed" subsystem within it) evolves towards more chaos, that is, higher entropy. Mar 11 at 10:22

7 Answers 7


to complete RodolfoAP's answer with a simple image of what entropy is that might resolve the confusion of OP.

Entropy is not about "order" or "chaos", those terms are just used by PhDs to explain to the general public in a 15 minutes video what took years of dedicated study to figure out. They are trying to describe complex equations with labels we assign to states of matter in order to express our feelings about it. Largely, people tend to consider orderly what they can understand and figure intuitively, and "disorderly" what is too complex to figure out, but there is no objective reality behind those labels. After all, as OP says, one could consider the thermal death of the universe, when everything is in so much equilibrium that nothing happens, to be the most orderly possible state of being.

What entropy is actually about is the most probable evolution of the state of a system after some time has passed.

It can be figured by imagining a heap of sand: the most stable, difficult to alter state for that sand is actually not to be put together as a heap, but spread on the ground as flat as possible, because grains of sand tend to fall down, not climb on top of each other. The overwhelming majority of what happens to the heap, gusts of wind, tremors, people walking on the heap... will push grains down. It might sometimes happen that some grains get higher, but those occurences are few and far between, so the most probable, to the point of overwhelming certainty, state of the heap after time has passed is to be lower and wider, not higher.

What the equations of entropy define is the "flatness" of the heap. Which of the two states of the sand, being in a neat heap or neatly spread evenly on the floor is the most "orderly" state is left as an exercise to the reader.

  • 2
    +1 What a beautiful analogy, I can say this is the closest I have come to completely grasping the idea of entropy.
    – How why e
    Mar 11 at 2:13
  • Nice indeed! I would add that when the sand is in a heap there exists the possibility of doing something with it, like letting it run through a funnel and drive a tiny machine. In general, what happens to the heap reduces it usefulness.
    – Philip Roe
    Mar 11 at 19:21
  • 1
    That is a good analogy but you might also point out that without intervention, given bounded area (let's say a pile in a sealed box) it will, over time, only flatten, never grow back into a mound. This seems a strange thing to say for sand, but the sand is representative of energy. All "Work" done with the energy leads towards the "Flat sand" state, and once it's at the "Flat" state no more work can be done (Nothing can happen) for that isolated area without external input.
    – Bill K
    Mar 11 at 21:38
  • 1
    that's the problem with analogies, if you stretch them to much they tend to burst. I guess one could build on the sand (pun intended) and move to a steam engine or something. Maybe use the image of an hourglass with a small mill wheel. I suspect creationists would also jump on to suggest that this is proof that someone had to put the heap together in the first place, etc. But I think the sand example is enough to dispel the whole "order/chaos" misconception, which was the original goal.
    – armand
    Mar 12 at 0:14
  • 1
    @leftaroundabout analogies break, that's life. A reader who does not have the notion that closed systems do not lose energy and can't have "low energy states" is indeed very unprepared. I can't take the responsibility to babysit people.
    – armand
    Mar 13 at 13:36

Arguably everything that's "physical" but not "fundamental" can be argued to be idealistic. If heat doesn't exist then neither do chairs - they are both measures of groups of particles, and not particles themselves.

Physicalism doesn't actually need to be opposed to idealism, as long as the implementation of those ideas happens because of physical things. Evolution is an idea, and it's implemented in our universe by physical things. Algorithms are idealistic, and they're implemented in our universe by physical things.

Physicalism isn't the denial of all abstract ideas, it's the statement that, if those abstract ideas "exist" in our universe, they exist in the context of physical things - they are implemented within physics, by the interactions of physical things.

  • 1
    Agreed. The central role of abstract objects in our universe is a problem for materialism, but we know from the modern physics revolution of the early 20th century that materialism is wrong. There is matter, but not everything is matter, and matter is not most fundamental. Physics is very comfortable being an abstract/material fusion. Abstract objects are only a problem if one tries to use physicalism as a monistic ontology. Treat physicalism as a dualist fusion, and abstractions are no problem for the view.
    – Dcleve
    Mar 22 at 13:25
  • The best way to characterize what physicalism is claiming is that Popper's world 2, that of experiences, is not fundamental -- that it is entirely dependent on either world 1 or 3. It is a denial of triplism. Note that functionalism, and the AI approach to consciousness, rely upon an identity theory to abstract functions/algorithms, not to matter/neurons. These are still recognizably physicalist efforts to explain away consciousness, and deny its causal power, but they are also explicitly dualist in that the way this is done traces to an abstraction not to matter.
    – Dcleve
    Mar 22 at 13:30

A couple of points to begin with. As others have pointed out, you have entropy backwards. The universe started in a state of low entropy and is evolving towards a state of higher entropy. We don't know why the universe was ever in a state of low entropy. The other point is that the connection between entropy and disorder is not straightfoward.

The concept of entropy has in fact developed through three distinct iterations.

  1. In the middle of the 19th century, the concept was introduced by Clausius and Carnot in the context of understanding the operation of engines. Entropy was conceived as a bulk property of matter that provides a handle on how much work a system is capable of performing. It was defined as dS = dQ(rev)/T.

  2. At the end of the 19th century, Boltzman and Gibbs used statistical mechanics to provide a statistical understanding of entropy in terms of the properties of the microstates of an ensemble of qualitatively identical particles. Boltzman's equation is S = k log(W). Gibbs' is S = -k Sigma(p log(p)). The two are not identical, but they both account for entropy in terms of probability distributions defined over sets of possible combinations of microstates.

  3. In the middle of the 20th centry, Shannon introduced the concept of information entropy, which is H = -Sigma(p log(p)). Information entropy is a concept rooted in communication theory and is closely related to the concept of compressibility. This would seem unrelated to thermodynamic entropy, but it has the same equation. This has led many theorists to suggest that they are actually fundamentally the same concept. On this view, entropy is a measure of a lack of information. It is not a property of matter as such, but a property of how much information we are lacking about a given system. If this is correct, then the second law of thermodynamics is not so much a law of physics, but an application of Bayesian statistics. This position is disputed, but it has a lot going for it. For a defence of it, I recommend Arieh Ben-Naim's book, A Farewell to Entropy (World Scientific, 2008).

So I think your question should really be not, "Is entropy physical or idealistic?" but, "Is entropy physical or information theoretic?" The answer is still a matter of dispute.

  • +1 Great explanation and response. Thank you
    – How why e
    Mar 12 at 3:01
  • 1
    History is so important. +1. See point 3 in my profile
    – Rushi
    Mar 12 at 3:52
  • @Rushi I strongly agree
    – How why e
    Mar 12 at 4:52

There are many misconceptions about entropy.

  1. Direct answer to the main question: thermodynamic quantities are essentially macrostates, or ideals, which are the counterpart of microstates, or what is physical, factual. Temperature is a clear example, it is an average of microstatic energy levels. Heat, the energy that can be transferred, is also an ideal, what is physical is just molecules bouncing and moving at different speeds. So, literally no part has the same temperature of the whole: differences will be large, or infinitesimal, but never the same. If you think you can find a molecule that would be at the average speed, you are being subjective. Just add more precision to the speed measuring, adding more decimal levels, and measures will always differ. The decision of using, for example, eight decimal numbers, is arbitrary and subjective, that is, ideal.

  2. Entropy is an ideal measure of a system, that reaches its minimum value when the system is formed. The increase in entropy essentially means the dissipation of the system. Physically, there are no systems, molecules are not bouncing balls, systems have no boundaries, and systems dissipate from the whole (e.g. a light bulb) to the minimum possible part (e.g. atoms, quarks, strings or fields).

  3. Entropy implies an extreme idealization of systems, which must be closed, must comply with the first law (its energy can't change), and which has only one level of order or disorder (which is expressed by entropy). Physically, systems are way more complex. All physical systems are open, they always, and permanently, exchange energy, and have multiple levels of order (things are made of parts, which are made of particles, which are made of molecules, which are made of atoms, etc.). So, the entropy of a real system is literally impossible to calculate.

  4. So, a direct consequence of such complexity is that thermodynamics can essentially describe the dissipation of systems, but there is no scientific theory that can describe how systems come into order. The notion of heat death of the universe is stupid, that would be like all the universe becoming atoms perfectly distributed through all space. The universe does not follow such dynamics. Systems are constantly formed, get minimum entropy and then, they start increasing its entropy until dissipation, from big parts (a light bulb or a rock have minimum entropy when they are created), then, they reach the maximal entropy when they break into pieces; each part has at such precise moment, its minimum entropy, until they break, and so, until reaching atomic elements. But each one of such elements, molecules or pieces, can at any moment become integral part of a new system, for example, a new rock, or a new bottle. The moment such new system is created, there is a new minimum value of entropy. The heat death of the universe is an old idea, that don't take the dynamics of real nature in consideration.

Replies to comments:

@causative: "There are no known macroscopic processes in which entropy decreases". Correct, but the observation goes in a different sense: if entropy does increase, it means it was smaller before. More precisely, at the point the system was created. Such point exceeds thermodynamics: the four laws deal with already existing systems, with an initial (smallest) entropy. It is naive to think that entropy only increases in the universe, and that all systems are already there, that no one more will be created and that all are subject to a decrease of entropy. Your observation does not follow (red herring).

@msalters: 'You can't just "add more precision" - Heisenberg forbids it': yes you can (the comment is about speed measuring). For example, you can measure pi=3.1416 or add more precision measuring 3.14159265358989. You don't even need to reach quantum levels. Simple mistake.

@msalters: 'The universe doesn't become atoms - way too much order': "becoming atoms" is just a possibility, I don't know what is the end of the "heat death" because it has no sense. Fallacy: ignoratio elenchi.

  • I do accept your answer but can you please explain to me how entropy connects to the universe as a whole (as a system). Is entropy connected to the "beginning" and possible "end" of the universe or is it just something "we" humans do out of necessity for explanation
    – How why e
    Mar 11 at 2:19
  • 3
    Your point (4) goes against scientific consensus. There are no known macroscopic processes in which entropy decreases. The creation of a light bulb, a rock, or a bottle are not exceptions to this. A bottle has low entropy but a lot of waste heat was involved in its creation, which increased entropy overall when you count all the variables.
    – causative
    Mar 11 at 2:58
  • @causative I agree!
    – How why e
    Mar 11 at 3:21
  • 3
    Point (1) doesn't work. You can't just "add more precision" - Heisenberg forbids it. The microstate you're assuming is not factual; it's just another ideal. And (3) is plain wrong. Any basic course in thermodynamics will explain in great detail different contributions to the entropy of a system. Your "multiple levels of order" is an ideal and does not affect entropy. Yes, the entropy of a real system is quite complex, to the point that we use a logarithm to keep the number manageable. But that's just math, big numbers are just as real as small numbers.
    – MSalters
    Mar 11 at 9:13
  • 1
    Re "Entropy implies an extreme idealization of systems": The "idealization" is only happening to create the concept, much like Newton's laws are best understood by conceptually ignoring friction; but once the concept has been established by separating and analyzing the various interactions at play both theories can be used perfectly well to explain real-world events. Thermodynamics can of course describe, indeed is indispensable to describe real-world systems which obviously are not perfectly closed, like your A/C. Mar 11 at 10:30

Your idea that perfect uniformity could be labeled "orderly" in natural language is perfectly valid; I cannot help but remark that some ideologies would consider a society in which all members are uniform and equal "orderly" and perfect.

In the realm of physics, as in society, perfect uniformity would imply perfect equilibrium. But in both cases, a perfect equilibrium would be the end of history: Nothing can happen any longer! All processes in nature are happening because a state that is not in equilibrium moves closer towards the equilibrium. When we say in casual terms "we produce energy", we actually transform energy (for example, chemical energy into electrical energy in a coal power plant) and then transport an imbalance (for example, in an electrical wire) to a place where the resolution of this imbalance can drive a process that is useful to us (for example, boil water for a cup of tea). If no imbalance is left, no energy flow can happen any longer, no water can be boiled and no tea can be made. That would be the end of history, not only for Englishmen.

The usual terminology, however, labels imbalanced states as order. Anybody who has children will find the analogy of a children's room valuable. Its "natural state" to which it gravitates if left alone is disorder, or chaos: Everything is spread uniformly across the floor. It needs an effort — one must expend energy — to restore order and put stuff back in drawers and shelves, sort the pencils and Lego bricks by color, etc. Quite obviously, that "order" is a state which is highly imbalanced and needs constant effort to maintain, a state which will have a relentless tendency towards the equilibrium we would call "chaos".

A more serious example are living organisms, including ourselves. A constant flow of energy is necessary to sustain the highly specific state we call "being alive"; as soon as that energy flow stops, because we lack oxygen or nutrition, decay sets in. Very quickly, the highly specific state of being alive becomes a very unspecific state of matter distributed through the environment. This flow of energy sustaining us is, at the end of the day, only possible because the sunlight is a stream of low entropy (i.e., highly concentrated, compared to the environment) energy that passes through the system Earth, leaving it as a stream of energy with higher entropy. A small part of this energy stream is routed through the biosphere and is used to sustain the specific states, the "orderly patterns" we call life; islands of low entropy in a sea of chaos.

Since we are in the Philosophy SE, let me add that a typical organism is more than meets the eye. Much of the physical matter of an organism, including ourselves, is exchanged multiple times during its lifetime. What defines us, as individuals, is not so much the 70 kilograms or so of water, carbon and minerals we usually label "self"; many of the atoms making up our bodies right now will be scattered around the world in ten years' time. Instead, what defines us — what we are — is a specific pattern which is self-replicating through time. Our essence is less a physical object than a standing wave in spacetime. Lost skin cells are replaced, blood cells die and are created anew, food and water transition through our bodies in large amounts. Sustaining this pattern, i.e., fighting the relentless onslaught of entropy, needs a continuous flow of energy. This pattern is order; stop the energy flow and it all falls victim to the great equalizer, chaos.


Entropy is physical. The way it's explained may feel idealistic, but we didn't get an industrial revolution because of idealism, we got communism because of the industrial revolution. And the industrial revolution relied heavily on the steam engine whose theoretical underpinnings are thermodynamics, the second law centering entropy.

As has been pointed out, the universe tends towards states of increasing global entropy at local scales, but, and this is a mind-altering set of statements, the universe as a whole has a level of entropy that doesn't change very much at all because it is exceedingly large. Entropy matters for individual stars which themselves are many times the size of our planet which itself is many times the size of you, but each star is surrounded by a much much larger volume of nearly empty space.

In addition, as time goes on across systems and disciplines, we see that small corners of the universe move to ever-more intricate and ordered states. Think of you and I and how much more intricate we are compared to the singular-cellular life that has dominated Earth for billions of years. Think of our computers which began as hulking idiots and whose transistors will be made of individually placed atoms in the near future (everything is near future in the scale of a universe's lifetime). Think of atoms like uranium which are only made in super-novae.

Entropy can be mundane. Think of wiping the counter down with water. You could dry it, but after an hour, the counter will dry itself off. This is not unexpected, but it should be. The boiling point of water is more than 70 Kelvin above room temperature, and this phenomenon will even happen 100 Kelvin below the boiling point. Why does the water "evaporate"? Well, we're in the middle of a discussion about Entropy, so ... a better question might be how does entropy have anything to do with that? The water on your counter, indeed any liquid water, has profoundly less entropy than the water in the air. Air, of course, can accept water molecules into it in a vaguely similar way as liquids can accept dissolved solids and gasses. As a result, while evaporation isn't energetically favorable, it is entropically favorable, and that means that if the air has room for the water (not fully or super saturated with water already), the water will evaporate.

This, of course, brings us back to the heat death of the universe. A post-heat death universe may seem calm, and it is, but that isn't the same as the having low entropy. In the beginning, all of the matter of the universe seems like it was at one place, and this is a characteristically low entropy state - akin to the water all being in your glass. In the end, it's spread out quite a bit, and just like with the water, that represents a significantly higher entropy state. Thankfully, the universe should last about 10^6 times as long as it already has (this explains the statement that the universe's entropy doesn't change much).


It seems to me that observations suggest a duality. A tendency towards order that has a counterbalancing tendency away from it. And in places they meet in the middle, or in distributed pockets among one another.

Gases float around making up the atmosphere. Pushed by winds, and updrafts and downdrafts, and higher and lower pressure pockets.

Mixed in amongst is water vapour. Suspended.

When the temperature is close to, or fluctuating between freezing and not, the water molecules tend to gather together. Some 100,000,000,000,000,000,000 will get together and form a snowflake.

Over and over again, the partially random snowflakes will form similar patterns... with the same underlying framework... a flattened cuboctahedron.

100 billion billion (a sextillion) water molecules... forming repetitive patterns...


Directly related to this geometric pattern...


Crystallization seems like the opposite of entropy to me.

Sure, spill the pencils... they won't magically leap back into the cup.

But spray a bunch of water mist into the air, and you will find patterns emerge from within.

Entropy is likely (and suggested by observation) to be only half the picture. It isn't a hook I would hang my only hat on.

Cause I don't see entropy at play here:

Snowflake 2

Or here:

Snowflake 3

Or here:

Snowflake 4

Remembering that each snowflake is the assembly of 100,000,000,000,000,000,000 water molecules, give or take a few quintillions, or quadrillions.

Snowflake to cuboctahedron

Snowflake to cuboctahedron

  • 3
    An individual snowflake does have low entropy, but the total of what we had before the snowflakes formed had lower entropy than the total after they form. Initially we had above-freezing water droplets and below-freezing air - a hot reservoir and a cold reservoir. Afterwards, the hot and the cold mixed; the droplets cooled down and the air warmed a bit. Mixing hot and cold means entropy is going up. As a side effect, low-entropy snowflakes were produced, but the cost of producing those snowflakes was a greater overall increase in thermodynamic entropy from the mixing of hot and cold.
    – causative
    Mar 11 at 3:31
  • 1
    Another way to put it is that you can decrease entropy (in one part of a system, not overall) if you have a source of thermodynamic free energy, which does thermodynamic work and gets used up in the process. This is why life on Earth can be low-entropy: the Sun is the source of free energy. For a snowflake, the source of free energy was the difference between the relatively warm water droplets and cold air. This energy gets used up when the temperatures mix and equalize, part of it being used to make the snowflakes. In the process entropy increases overall.
    – causative
    Mar 11 at 3:46
  • More than "low entropy" it has "anti-entropy" in the form of observable complex structure. Hyper-order. Otherwise it would be just a big iceball. It ain't. It's a crystal. We could also see that it is part of a larger settling process... like cream rising from milk, but in this case, water vapour settling ouf of the atmosphere, falling to a very thin flat layer on the ground, much different from the minute individual molecules that were bouncing around in the air. A whole lot of order counterbalancing the random chaos. A beautiful dance on the edge of a knife. Mar 11 at 4:16
  • 2
    Observable complex structure is just low entropy, not anti-entropy. It's not that order "balances" entropy. It's that entropy keeps increasing and ordered low-entropy things are often produced as a side effect - but overall the entropy still increases. There's no "balance," entropy always wins when you account for everything.
    – causative
    Mar 11 at 4:54
  • I agree with causative's comments, but on a side note, where are you seeing that a snowflake is a flattened cuboctahedron? A quick search doesnt seem to show anything mentioning that shape in relation to snowflakes.
    – JMac
    Mar 11 at 12:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .