Recently, one of my macro-trader friends (arguably extremely bright) made a very interesting point about Joe Biden. We were debating on why the new U.S. president was about to deescalate geopolitical tensions, and we agreed that, as he was keen to get re-elected at the end of his current term, he would somehow try to emulate Obama’s social bottom line and its Obamacare keystone. For this to happen my friend argued, Biden would have to unlock a substantial amount of money from the budget, and combined with the relatively less belligerent character of Democrat presidents through history (by chance or by choice, you are the judge), he should probably target lower military expenditures, which in the U.S. entails curbing false perils inflation. This has actually already started but at the end of our little debate, one question remained (at least for me): where are the peace dividends going to come in the case of lower U.S./China tension? For Trump’s escalation of tensions with China were less militaristic than commercial — apart maybe from higher weapon sales to Taiwan but weapons exports are a positive for the budget, not a negative. Of course American leaders cherish bogeymen, especially when this serves their friends in the military industry. But i) there is no obvious ways to reduce U.S. presence in the Pacific (mainly South Korea, Japan and Siam), ii) most people in Congress, including the Democrats (and Biden himself) are still sanguine about China and iii) the bulk of high stake military tensions are with Russia, not China.
This leads me to believe that the reasons behind better U.S./China relationships apparently instigated by Biden two weeks ago have to be sought somewhere else more logical: the comeback of “normality”. In France after the admittedly strange Sarkozy moment (2007–2012), the socialist president François Holland’s campaign theme was exactly that: a return to a “normal” presidency, which Biden reminds me very much, as a French person.
Now, you’d be excused to think that, on the boredom scale, “normality” as a subject is hard to beat not to mention how it could lead to such a visibly long article. So I’ll remind you that in politics, normality is generally dubbed “status quo” and it seems that the destruction of the status quo was exactly why so many American people voted for Trump in 2016 and not for Hillary, who was seen all along as the personification of it. You can say a lot of things about Trump, but boring? I don’t think so. He was no revolutionary — he totally hates genuine socialist revolutions — but he antagonised a wide share of the billionaires and DC clique, including many prominent intelligence services, often associated with the “deep state”, and of course MSM. The sheer animosity and outrageous lies of many of them to get back a more “normal” (read: narrative aligned) person on the throne throughout his presidency showed to the world how thoroughly divided these elites really were. Trump was himself an oligarch but one that made clear the oligarchs power monopoly was over which pushed the other bunch of oligarchs to reveal how far they were ready to go to get the supreme job back.
The status quo lead, this time, to the fracturing of the American oligarchy block in two camps fighting bitterly for the job. Plenty of metrics show how divided the whole American political class currently stands. This is for example a measure of the polarisation of Congress:
Normality is the opposite of boring. Prolonged long enough, it bears plenty of spectacular consequences and in this piece I am going to show you why.
Entropy feedback loops
Let’s imagine a (admittedly weird) civilisation which solves all its problems always the same way: by a meeting between its leaders in a boardroom around a wooden table. Every time an issue arises, leaders call for a meeting which first requires to get a special meeting table made by a special and unique table factory, and none other. Once the issue is solved (by a decision), the council dissolves, the table is traditionally burnt and the leaders go back to their mundane occupation. Throughout the long history of this imaginary civilisation, countless meeting tables have been produced, used and destroyed to solve as many countless issues of all sorts.
Now imagine that one day, an unprecedented issue arises: the table factory used to produce the special wooden tables is itself out of order and no more meeting tables can be produced. Arguably a council of the leaders should immediately be called in order to decide how the factory should be fixed, as this is such a central issue, or maybe how the council could proceed in the future. But with no meeting table available, no such meeting can be called and hence no decision made. This is a very thorny issue indeed: any solution to the problem requires the problem to be solved, before being solved.
This is a (negative) feedback loop, and clearly this sort of issue would make this civilisation stop functioning and maybe even disappear (“collapse”) if no solution is found.
There is a number of reasons why this problem arose: it could be the lack of trees used to make the tables, it could be because of an industrial action, a management void, a tricky engineering problem, war damage etc. But one thing is certain: relying exclusively on this specific factory to solve all imaginable problems was a recipe for disaster. The more issues were solved this way, the more likely it was that one day, that civilisation would encounter the fatal blow. This is specifically the problem with linearity’s homogeneity property, i.e. the belief that to solve N times a given problem, all you need to do is apply N times the solution which solved the problem in the first place.
As the feedback loop deals with the limit of homogeneity in the real — generally non-linear — world, it is therefore a function of how many times the table factory was used and to formalise this, we propose the following simple principle. If N is the number of times the solution mechanism was tapped in the past, the probability p of a feedback loop is given by:
where f is a monotonously increasing function whose values are bound between 0 and 1. All this formula says is that the likelihood of encountering a catastrophic feedback loop increases with the number of issues solved in the past by the table factory.
As you have come this far, you are an astute reader who has already realised that equation (1) only holds in a general manner if the set of rules — or principles, to use a physics’ word — used to produce the solutions is fixed and cannot be changed at any point. Any clever leader would make this observation too and propose that, instead of organising a council around that specific wooden table in that specific boardroom, we could just do it while standing up or something totally crazy like that, which is a paradigm shift. The point of this trivial example is to intuitively show you that if all the achievable states that can ever be reached depend on a set of rules that is fixed, there is a link between the number of those states and the probability of a feedback loop. At that point, the initial set of rules can produce no more sane solution.
Equation (1) says that political status quo, which is a fixed set of rules, does not decrease the risk of disaster, but rather the opposite: it increases it. Let’s try to formalise this further.
Entropy pits
In a theoretical paper published in 1999, two physicists, Elliott Lieb and Jakob Yngvason developed an unorthodox approach (later popularised by André Thess’ Thermodynamics for the unsatisfied) to a sometimes confusing physical quantity: thermodynamic entropy. What they showed is that, if you apply, from outside, reproducible actions to a closed physical {System}, the successive states of the {System} can be ranked along an increasing function S of real values respecting certain properties. S is unique (to a multiplicative constant), sub-additive, concave etc. and their result stands for any closed physical system as long as it’s not “too large” (to be able to neglect gravitational forces) or “too small” (so that there is no quantum phenomena involved), as long as the actions are reproducible and carried out from a mechanical {Activating device} placed outside the {System}. The states of the systems have to be equilibrium states, which simply means that, once in a certain state, the {System} remains unchanged until the end of times.
To anyone accustomed to thermodynamics, S should remind them of the entropy state function. And indeed Lieb and Yngvason’s (“L&Y” from there on) S function has precisely the properties you would expect for the usual macroscopic entropy state function used in thermodynamics. Their axioms can be sketched as follows:
After defining an intuitive set of axioms defining what a state change is, they call this relationship between successive states “adiabatic accessibility” and show that State2 is adiabatically accessible from State1 if and only if S(State2) > S(State1).
I am not going to dive into the mathematical details of their outstanding work here but instead try and extract its philosophy. You should focus hard because we are now approaching the heart of the matter. L&Y’s paper is about physical (and chemical) systems. What about other systems, like human ones? Surely they are not that simple. Let’s be clear from the onset, there is no thermodynamic entropy outside the microscopic world, and rest assured, I am not going the route of Nicholas Georgescu-Roegen’s fourth entropy law. Granted, L&Y’s research is quite a startling result (their paper is more than 100 pages long), but still one that can be summed up in a few lines — which is probably why it’s so powerful. Let’s observe two points: first, although they both were physicists, Lieb and Yngavson used a very general reasoning to build S which apply to any state change falling within their axioms. And second, L&Y’s {System} state changes are only allowed along reproducible actions.
So what are the limits to our philosophical transposition attempt? As we want to apply L&Y’s thinking to non-thermodynamic problems, the difference cannot lay in the fixed nature of constraints. The reproducible character of the actions is paramount to the entropy. Opposing human freewill as an argument that L&Y’s research cannot apply to human societies is off topic. To apply L&Y you can only start from a fixed set of principles. There is no other option there.
L&Y’s result also imposes the {System} to be adiabatically segregated from the external {Activating device} and indeed, it’s the whole point if this discussion to exclude all dynamic relationship between the object of humans actions and where decisions take place. Adiabatic means that nothing comes in or out of the system.
As L&Y’s work is axiomatic, their conclusions hold as long as their axioms are respected, even in a human environment as soon as you can find a place with reasonably fixed principles and where there is a clear partition between where decisions are made and where they apply. As long as there is a problem-solving mechanism, be it human or not, which follows a set of reproducible rules, there will be an entropy function increasing from one state to the next. This is why our wooden table civilisation example works.
A last requirement has to be fulfilled: as L&Y entropy only exists for equilibrium states, the set of problems to solve also have to be at equilibrium, meaning that all problems remain unsolved until the end of times in the absence of exogenous perturbations (for example, the broken wooden table factory remains broken until a decision is made). Hence only a small modification in L&Y’s axioms allows us to apply their result to non-thermodynamic subjects:
So the difference is not in the axioms. It’s in what can happen after building S. In the human world, contrary to what happens with physics’ principles, rules are generally fixed by humans themselves. And they are fairly flexible animals. In physics you cannot escape gravity. In the human world, you can organise a meeting room with no wooden table. Even though it’s illegal you can actually commit theft or murder someone. There is a fair chance you get locked up for it, but as long as prison is not a problem for you, you can choose to do it. A smart leader from our wooden table civilisation who decides to hold a council with no meeting table effectively breaches its fixed set of rules. More formally, a paradigm shift is a non-reproducible action, since “reproducible” precisely means “sticking to a fixed set of rules”. This is done by expanding the set of fixed rules, i.e. adding an exogenous rule to the rulebook, one which contradicts it.
Like in physics, non-reproducible actions are the only way to decrease the entropy of a system. In physics, that would be by accepting the {Activating device} to behave randomly, or, more interestingly, the adiabatic partition to become porous to heat and matter so that it stays far from equilibrium. Breaching an older set of rules potentially creates negative entropy, something which I call an entropy pit because it is “somewhere” to store entropy which will allow us to recover our ability to change the state of the {System} further, and without which it would be stuck with is current problems forever. An entropy pit allows for an entropy descent. In thermodynamics, an entropy pit might be the atmosphere surrounding a Diesel engine for example — i.e. the space where it can pour its entropy out. No thermodynamic engine can complete its cycle without evacuating some of its entropy outside at some point (during its exhaustion phase):
To come back to our subject, although Donald Trump was not a revolutionary, by all accounts he marked a breakdown point of American elites’s monopoly on power and narrative. For the first time in a long time, they were bitterly fighting over getting the top job, to the point where they almost seemed ready to kill each other, including some of us and even potentially end all life on earth to achieve this goal through political escalation with the only other nuclear superpower on Earth, Russia. Such divisions are not new in the sense that they certainly come and go periodically. What is noticeable is that they are running at their highest point since 1900 and potentially even since the American civil war. That’s what analysts mean when they say the elite world is polarised or even fractured, and Trump was probably just the latest personification of this situation.
When Donald Trump tried to remove troops from Syria and the Pentagon secretly disobeyed his orders, that was a paradigm breakdown both from Donald Trump’s side and the Pentagon. As he went against the grain on different topics, his appearance on global stage was, in effect, the manifestation of an entropy pit in the same sense that the decision of the smart leader of our meeting-civilisation to organise a meeting with no meeting-table was an entropy pit (note that the Pentagon sabotaging his orders was even probably breaking a hard constitutional rule).
Anyway, to conclude this part, let’s add a last point. As entropy is expressed in bits (which is not a unit, just a real number), it’s context-independent, which I am going to call commensurable. The green new deal proposed by Joe Biden and his team, although morally attractive and necessary, gives a condensed illustration of the consequences of entropy commensurability. According to a Financial Times report, a fully renewable world by 2050 requires 60 times more lithium than current global reserves. If consumption was multiplied by 60, current known reserves would only last about 3 years and lithium is not the only material in this case:
This means that, all things being equal (namely U.S. GDP), solving the looming climate crisis probably requires to get a hand on tremendous amounts of commodities. Some of those are located in the U.S., but most are in foreign countries, which can only mean one thing: more expansionary wars. Finding an entropy pit to advance the climate solution (a rather technical domain) immediately requires finding one in foreign policy and military. The commensurability of entropy means that, near maximum entropy, surprisingly distant problematic like education, military, policing, finance, politics etc. converge into one single unfathomable and dynamic problem (dynamic because in feedback loops’ solutions that change {System} state are influenced by the solution itself). For this reason, the way our economic theories are built and collapsing look like a schoolbook application of L&Y’s axioms.
Because of commensurability, if you choose to ring-fence a certain part of human activities against an entropy descent, you should not be surprised to have a pound of flesh taken elsewhere. We have apparently decided that our financial markets should never ever experience an entropy descent like it did in 2008 (more on this later), which means that any such future occurrence is bound to resurface elsewhere. It should be little surprise therefore that “despite” the massive Western central banks unconventional policies unleashed after Lehman’s default, social and political tensions have soared through the roof. Because of entropy’s commensurability, this “despite” is in fact a “because”.
Admittedly, increasing economic productivity (which contributes to about 95% of GDP growth) creates a negative entropy buffer which could in theory be used to solve our climate nightmare, but i) this reasoning does not account for Jevons paradox and ii) given the size of the issue, this buffer is thoroughly insufficient (remaining below 2°C of warming demands lowering our CO2 emissions by at least 4% every year until 2050, which means reducing world GDP by about as much as it decreased in 2020 every year until then; a decrease of more than 4% of CO2 emissions only happened three times since 1900: in 1933 when the Great Depression went global; in 1945 when both Germany and Japan industrial sectors were wiped out from the surface of the earth; and in 2020).
Energy descent vs. entropy descent
Of course, plenty of specialists argue that it’s first and foremost the lack of energy (as hinted in the FT table above) that articulates initially separated fields. And it’s right. Post-industrial revolution human civilisation has already reached conventional oil peak (in 2006) and there is increasing conviction among energy experts that we could be near a catastrophic energy descent (for a taster, watch Italy, Mexico and this global analysis).
The problem is that energy resources are attached to a unit (for example kWh) which forbids them from being smoothly generalised to more general issues, for example to a field as different as politics. Energy arguably acts as a link between politics and economics, but is not commeasurable to all subjects contrary to entropy which is context-independent and therefore usable indiscriminately. In any case the more primary energy we use, the less there is left, until there is none. Hence it’s trivial to build the same increasing entropy function f as equation (1) by equating it to the amount of resources already burnt (after multiplying it by a constant to turn kWh into bits). In this particular case, the maximum L&Y entropy is reached once there is no more resources. So, really, energy peak is a particular case of a wider entropy feedback loop. In fact, we are tantalisingly close of saying that all negative feedback loops are entropy feedback loops — but it’s a story for another day.
Besides, trying to understand some of our economic and climate issues only from an energy angle runs the risk of “conjuncturising” these issues. After all, if we managed to discover more oil, what’s the issue. Even the climate crisis could be solved with more energy. So what’s the big deal? Surely this is just bad luck and maybe, instead of questioning GDP growth and profit maximisation, we should in fact accelerate GDP growth! This is not even a caricature, it’s exactly the thinking of some very influential people. The problem is, infinite energy resources do not solve all entropy feedback loops.
In a seminal work, spread among many research papers and books that will undoubtedly be considered as revolutionary before long, cliodynamics and cultural evolution specialist Peter Turchin shows how a wide range of cultural, economic and anthropological proxies can help explain the oscillating trajectory of American political stability since the 1770's. He shows how one of his proxies, elite overproduction — an excessive number of highly empowered people competing for too few jobs — has been, in the U.S., a top contributor to political instability and violence, and even partly explains Civil War. To me, there is not much you can do to curb elite-overproduction. Promoting wide-spread secondary and university level education is an objective that should seat at the top of any political agenda. I cannot see anything wrong with constantly reducing industrial production costs either — one of the big drivers of elite over-production — nor encouraging people to get wealthier and wealthier (the American Dream). And yet, there can only be so many billionaires, so many Congress seats, and only one U.S. President. This means that, not withstanding some mitigating measures like lowering immigration (which boosts profits by applying a downward pressure on workers’ wages), any civilisation pursuing this hard-to-counter course of actions — basically progress — long enough will experience trouble. And since those troubles find their origin in the course of actions itself, this is a feedback loop, but not one that more energy resources can solve, which is probably why, even highly industrialised nations whose resources are seemingly unlimited experience troubles and revolutions from time to time. Let’s try and quantify them.
Quantifying a paradigm shift
If ΔS is the (theoretical) entropy step needed to unlock the following state of the {System}, we can amend our initial sketch of what happens when it is at maximum entropy as follows:
This shows that, to overcome the big STOP sign, the set of principles E has to change by an amount proportional to ΔS. The paradigm change is indeed proportionate to the entropy needed to reach a new state, something that can only be achieved, since the {System} is at maximum entropy, by uncovering an entropy pit of at least ΔS.
For example only a paradigm change commensurate to our objective will see us through our climate pickle. All the parameters are in Kaya’s equation but to be sure they are: global economic and energy restraint — that not even accounting for energy peak — “de-growth”, demographic compression, global income redistribution.
Nowhere is this paradigm shift more striking than when it comes to individual well-being. Maximising consumption and GDP is seen, by many countries, as paramount to individual and thus general well-being. This logic is considered purely rationally driven. Like in wartime, who wants to be on the front line? No one. It’s the same thing with climate: who wants to sacrifice even a part of their GDP for the greater good? The only way out is cooperation but viewed within the framework of GDP-maximisation, cooperation is “irrational” since it will eat into the individuals’ share of wealth. Sadly no group can survive a war without cooperation between its members. Hence however unassailable the “rational” arguments look like, they will have to be overcome and mitigated with seemingly non-rational ones, but ones that will assure the survival of the species. This point is actually the subject of a whole new research field called group selection. The amount of disruption needed in that framework is at least equal to the amount of entropy change needed to assure this survival. During your next discussion about climate change where someone arguing that saving the economy should come first in this period where the Chinese are on full throttle, you will now be able to counter any supposedly “rational” argument that individual rationality is only one possible way to think about survival and that solving problems actually involves other factors incompatible with maximum personal well-being.
To gauge how much political change this paradigm shift represents we could start by looking at how much change would be needed to, say, replicate what happened in the U.S. between 1920 and 1945, a period of unrest which solved most of gilded-age inherited social and political problems. By comparing various economic and sociopolitical proxy variations of that time — which can be found in Peter Turchin’s data — to what is needed between today and 2050 we should be able to calibrate an entropy model which will spit out a ΔS. From there we could qualitatively match ΔS(1920–1950) with how much suffering and destruction was endured by the U.S. throughout that period to have a picture of disruptions expected in the world to achieve an entropy pit of ΔS(2020–2050). This too will be for another story because there are other details to explore first.
An apparent paradox before finishing
Deriving entropy from a set of axioms, like L&Y do it, requires the {System} and the {Activating device} to be segregated by an adiabatic partition. Only then can entropy be defined axiomatically. We just showed that, once this is the case and extrapolated long enough, we end up in certain cases with a feedback loop, i.e. a situation where the {Activating device} is engulfed in the {System}’s entropy function. As it’s a circular definition of entropy, this is not acceptable — and it’s a limitation that L&Y are well aware of. On the contrary admitting from the onset feedback loops between the {System} and the external {Activating device} is no sane way to construct an entropy function. Is entropy an illusory concept? It is, if you stick to rigorous axioms. As much as L&Y’s work is admirable, they admit a caveat right in the first few pages of their paper when they say that “it is difficult to avoid some circularity when defining the concept of adiabatic accessibility” axiomatically (p.17).
The only way to break free from this apparent paradox is to observe that their model is linear and that therefore the adiabatic partition is an approximation of reality: as long as the entropy generated by the {Activating device} is negligible compared to the entropy of {Activating device + System}, you are fine. For example, humans transform their ecosystem (the {System}) into stuff that serves their own economy. As long as the scale of this activity is not too large compared to the size of their ecosystem— for example as long as human population remains limited— considering the partition between the human economic milieu and wilderness as adiabatic should not be much of an issue. What we are currently witnessing though, and was thought to be unthinkable, is that human’s activities are actually so large that they influence Earth’s climate. The adiabatic approximation is not valid anymore: humans have been engulfed by an entropy function that they are themselves pushing higher, trapped like someone who would try to run his engine in his hermetically closed garage.
In other words, it’s not so much L&Y’s result that should focus our attention, but rather the point where it breaks down. Our entropy feedback loop, and hence a paradigm shift, is precisely the breakdown point of L&Y’s linear character of adiabatic accessibility, i.e. the moment where you cannot choose your actions anymore, something we will now conclude on.
Wrap up: there is no complete habitat
“Reproducible action” means what it means: you can write it down in non-ambiguous terms — a little bit like an algorithm — so that anyone who would like to replicate your idea would only have to follow your step by step instructions independently of the initial state of the system. Which, in turn, means that any final state can be entirely explained by the initial state of the system and the action performed on it. Hence you know an action is reproducible when you are able answer the question “why is the system in final state B?” for any initial state A. In that case, let’s call B a “decided state” and a milieu where all states have been decided a “complete habitat”. Why this word, “complete”? Expanding a fixed set of principles E by introducing a new rule r contradicting certain parts of E — like our leader organising a meeting with no wooden table — makes an habitat incomplete, since after that, all new states of the {System} will be compatible both with E and r, hence matching the definition of incompleteness in logic.
L&Y’s {Activating device} is a formal mechanism, whose state — i.e. entropy — never changes. How does that work in the human world? Humans perform plenty of changes on their natural environment in order to produce final goods whose physical and chemical properties serve their economy, and of course, the human habitat is meant to remain well isolated from nature. There is no sash window or white medical blouse or Diesel engines in the natural world, but with enough ingenuity and energy, humans make them, theoretically without compromising their own comfortable way of life. As we’ve seen, this is only theoretical but if you extrapolate on a wide enough scale, starting from an initial state, there would be nothing undecided left on the random side of the adiabatic partition: you would be able to explain the origin of everything in non-ambiguous terms, and this origin would be uniquely humane. “Why did it rain today”; “why did the stock market go up (or down) by x percent?”; “why did this bridge collapse?” Etc.
We can now say to Turchin, “Sir, not only are you right, but it seems that any fixed paradigm could have lead to both catastrophic states engulfing America that you discovered and predicted — one of which unfolding right now. As a matter of facts, your indicator of stability provides a good proxy for entropy once reformulated as such: each time the entropy generated by the fixed set of principles reaches a certain level, the probability of a negative feedback loop increases. Entropy peaks at the tops and bottoms of your chart, which, each time, coincides with a troubled period that you identify in “Ages of Discord”:
Human societies can be afflicted by two types of problems: 1) the negative consequences of a given rule (child’s work, slavery, torture etc.) and 2) the consequences of the mere existence of a rule. The latter have no reproducible solution and are the ones Turchin is — although not consciously— referring to. The U.S. is in the midst of such a moment, i.e. at the eve of revolutionary changes in the sense that the fixed set of principles accepted until now will soon be expanded. Trump’s few stabs at the status quo barely scratched the surface, and it seems that Biden will be ignoring this reality altogether, a choice that can only accumulate further entropy. As a matter of fact, Turchin’s three main issues afflicting the U.S. — excessive inequality, elite overproduction and fiscal crisis— are all entropy feedback loops. Let’s review quickly why. Inequality, which flows from profit maximisation, leads to popular discontent, protests, strikes, riots, consumer impoverishment and sometimes revolutions which erode profits but it’s not possible to reducing them without infringing the profit maximisation paradigm. Higher education and enrichment of the elite lead to the fracturing of elites’ monopoly block on power but it’s not acceptable to reduce the number of elites without preventing some people from accessing better jobs and education. Finally, fiscal largess are good for the majority, especially when people struggle financially, but you cannot go to infinity without currency debasement (the fact that it is actually not happening, as noticed by Turchin himself, is quite a puzzling situation).
Now to anyone thinking that habitat completeness would not be much of an issue (there is actually a brilliant fauna-centered video on this thought experiment on the French Youtube) humans tried to decide many times through history and on a large enough scale, everything there is to decide. For example there is a way to look at the 2008 Subprime crisis by starting in 1995 with Clinton’s rewriting of the Community Reinvestment Act which was enacted to pressure banks into lowering their lending criteria in low-income neighborhoods. From there, mortgage-backed debts started to pile up until securitised papers became America’s top export, turning the random GDP data into a mere by-product of the willingness of banks to write more and more debts, which most were more than happy to do. The more debts they issued, the higher the GDP. “Simples!”. We all know how that ended.
When Germany lost WWI it was forced by the Treaty of Versailles to accept having large parts of its territory — among them the all important industrial Ruhr region — occupied by foreign troops. Deciding, from Paris and London — an external and remote position — and with 100% guarantee that there will be zero comeback from the (disarmed) German people, makes this whole experiment another textbook application of L&Y’s result on a large scale. We also know how that ended: the “thinking outside the box” was done by the Nazis twenty years later, who, incidentally, not only stretched the modern times’ humanly acceptable rulebook a fair bit, but also expanded geographically.
The example of WWII also illustrates how entropy feedback loops — in this case a “total war” — leads to non-reproducible final states. The outcome of such wars is “win or die”. No single entity, or, to employ our jargon, no single {Activating device} can decide on its own of its outcome. As total wars divide the world in two, the final result becomes random for everyone, which also illustrates the point made above on the absence of choice at maximum entropy. As no more reproducible action can change the state of the {System}, all further entropy resulting from new reproducible action will leak straight back outside, into the {Activating device} milieu — typically society. And a change of entropy is a change of state. So at maximum entropy the {Activating device} is being imposed a paradigm shift, by definition unacceptable to itself — since it breaks the only admitted rulebook. What’s worse than dying? Dying while not choosing to. A paradigm shift is the negation of a choice, something we encountered when dealing with how individual rationality is wiped out at maximum entropy. This is particularly striking in our wooden table civilisation where the state of the {System} does not allow any further change unless someone first accepts to change the rulebook: this change is not freely decided, it’s imposed by the circumstances. This splitting of society, occurring during total wars, is exactly what we see when looking at the political polarisation currently at work in the U.S..
As an exercise, I let the reader explore how the “Ancien Régime” too respected L&Y’s axioms, hence bringing about a feedback loop which resulted in the French Revolution.
There is another interesting L&Y feedback loop example I am not going to explore here either: sleep (where the {Activating device} is our conscience).
There lays the profound morale of entropy feedback loops. As the universe is fundamentally uncertain it’s futile to try and decide all its possible states: there is no complete milieu, i.e. somewhere where all states come from the same set of fixed principles (and necessarily the same starting point). There is no forcing randomness. A complete universe, due to its unlikelihood, is bound to experience a “statistical tension”. One way to feel this tension would be to imagine the creation of a non biased coin falling always on tail (or heads). For that to happen, you would have to change the laws of physics entirely, which is equivalent to having an {Activating device} placed outside the whole universe (to be able to change the universe’s state). That means that you would also be part of the {System} yourself, since you are part of the universe. And indeed, the amount of energy needed to activate this device would engulf you with it — namely your electricity bills — , whereas you were meant to remain safely on the other side of the adiabatic partition. This is actually general: as the reduction of the number of different equilibrium states in an adiabatic enclosure is not allowed, trying to do so will necessarily negate the existence of any adiabatic partition.
In physics, once entropy has been maximised, as no further change of state is possible, the system reaches thermodynamic equilibrium. This doesn’t happen in dynamic systems. Once entropy has reached its top for a given paradigm, what is admissible can still be changed. Brexit might be one of those recent instances, as globalisation does not seem to be fashionable anymore:
Caitlin Johnstone (purely based on my experience, a “Superforecaster”), who has been talking about pattern breakdown for a while said we should expect more of this in the near future. If I had to make a call on what the L&Y {Activating device} currently driving American society’s actions looks like, this is what I would say: global domination; GDP growth; individualism; profit maximisation as a good proxy to well being; globalisation; corruption; highly educated business and technical elites; nigh infinite freedom (as described by Tocqueville, this is a country whose resources, to any measure and for a long part of its history, were unlimited). As long as this fixed set of constraints is not allowed to unravel— freedom can be a constraint if you stick to it long enough on a finite planet — , we should see further political and social escalations there. According to Turchin, an unraveling is on the way and it’s unlikely that anyone — especially not Biden’s status quo — can stop it. The good news is that, once the U.S. pulls through it, it should be blue sky ahead.
