Energy and Civilization: A History, Vaclav Smil (MIT Press, 2017).
There are two kinds of mathematicians: problem-solvers and theory-builders. They’re what they sound like. A problem-solver gets obsessed with particular, concrete, fiendishly tricky problems, and wants to solve them. Along the way, he may accidentally invent a whole body of theory that sheds light on unrelated areas of mathematics, but that isn’t why he’s doing it. He wants to solve the problem. The archetypal problem-solver is somebody like Paul Erdös.
A theory-builder is just the opposite. A theory-builder is motivated to change or reframe or generalize the whole way that we think about something. He’s driven to find deeper or simpler or more powerful ways of reconceptualizing things that we already know. Along the way, as a kind of side-effect, his new way of thinking about things might accidentally solve some fiendishly difficult problems, but that isn’t why he’s doing it. He wants to build theories. The archetypal theory-builder is somebody like Alexandre Grothendieck.
There’s an analogous distinction in the study of history — some historians micro-focus on particular historical episodes, and document them in excruciating detail. There’s almost no limit to how narrow this kind of study can get — we’ve previously reviewed a book whose entire focus was a few years in the history of South Africa, and another that focused on just a few months in the Donbass. This is analogous to the mathematical “problem-solver”, and like the mathematical version, I love these works not just for the episodes they document, but for the general lessons about the human condition that they turn up as a side-effect.
In fact I have a bit of a bias: it seems to me that in mathematics, both the problem-solver and the theory-builder approaches have been very fruitful, but that in history the theory-builders are almost always kidding themselves. Part of it is that human beings are just more maddeningly complex and intransigent than even the most poorly-behaved topological spaces, and part of it is that it’s harder to muster the necessary dispassion to do good theory-building when your objects of study are people and civilizations (everybody has an axe to grind). Whatever the case, there’s a long horrible practice of people writing sweeping world histories around some grand theme that totally elide important complexity and fail to hold up under any kind of scrutiny.1
Vaclav Smil is a theory-builder, but rarely for historians, he’s pretty good at it. One promising sign is that he has a cult following (Bill Gates once said he waited for new Vaclav Smil books the way he waited for Star Wars movies), but I actually think the reason his approach works is that he’s chosen the most impersonal and non-negotiable of all qualities to build his theory around: energy constraints. For Smil, the history of human civilization is nothing more and nothing less than the history of attempts to balance a single equation: joules of energy input, in the form of food and fuel, versus joules of energy output, in the form of metabolic demands, muscle power to perform labor, and heating.
Smil is almost tautologically correct about this, because for most of human history we’ve surfed the Malthusian wave right on the line where that equation balances. Because of that, and because some of the terms making up each side are fixed, we can learn a lot from seeing how the rest of the sum decomposes. Here’s an example of his method: we can measure the incoming solar flux on the earth’s surface per unit area at various latitudes. Then we can multiply that by the optimal power conversion efficiency of photosynthesis (about 0.5%), which gives us a maximum sustainable rate of plant growth (about 5-10 tons per hectare). Then we can assume that all of it is the best possible firewood (which yields about 18 gigajoules per ton when combusted) and arrive at a maximum power density from wood fires of about 0.5 watts per square meter of harvestable lumber.
On the other side of the equation, we can estimate the heating needs of a large premodern city (including cooking, smelting, and thermal regulation in Winter) and get a figure more like 20 or 30 W/m^2, which leads us to the conclusion that to have a sustainable fuel supply, a premodern city needed to be able to gather wood from an area roughly 50 times greater than the area of the city itself. But wait a minute! The average distance to haul that wood back to the city grows like the square root of the area being harvested, and each additional mile of hauling imposes further energy losses (in the form of feed for draft animals, human labor to maintain roads, etc.), which produces a hard cap on city size, depending on climate and latitude.2 Then we go back and look at the sizes of various ancient cities and…well, would you look at that, the capitals of empires are all clustering pretty close to our calculated maximum size.
You may protest that this is cold, bloodless, and neglects everything that makes history actually “fun.” To which my answer is yes, yes, and yes. This is not history, it’s barely even social science, it’s the sort of analysis an autistic alien robot might do, and yet…it takes my breath away that a little bit of Fermi calculation can reveal so much about how our ancestors lived, and how we do.
Smil’s account begins in prehistory,3 where the energy balance is direct, raw, and existential. Our hominid ancestors had a certain basal metabolic rate, a number of joules of food that they had to consume in order for their hearts to continue beating and their brains to continue working. Every day, they had to find at least that much food on average, but the catch is that hunting or gathering or climbing trees or preparing for those activities itself takes additional energy. So these activities are properly considered investments of energy with some uncertain payoff. And that payoff needs to be not just greater than 1, but enough greater to cover the basal metabolic load (plus some extra for children, the sick, and the feeble).
Some investments take a while to pay off — for example, perhaps one member of your tribe is dedicated to making stone tools for hacking flesh from bones or opening coconuts. This process could take weeks, and during that time the rest of you need to bring in enough bonus joules to feed somebody who isn’t directly contributing. But at the end of it, you have a tool that may serve you for years, and which makes future hunts more efficient, increasing the energy returns of everything you do. Smil, with his autistic alien robot perspective, views all the works of civilization like this. Houses, roads, irrigation ditches, technology, domesticated animals, all represent energy “capital,” crystallized human labor, stored up and converted into a less ephemeral form, an energy bank account that yields regular dividends and can be drawn down in an emergency.
The best section of the book is the part about pre-industrial farming.4 To keep with the financial analogy, traditional agrarian societies are like a family where everybody is working three jobs, the credit cards are maxed out, and somebody has to move the car every few hours to avoid repossession because that’s cheaper than making the payments. Plowing and harvesting are both backbreaking forms of labor with extreme energy requirements, but the real outlays come with trying to circumvent two other limitations on agricultural productivity: nitrogen and water.
As population density grows, you need higher energetic return per area in order to keep everybody alive. This requires you to farm more intensively — for example by growing multiple crops per year. But this in turn exhausts the soil, and suddenly fertilization and irrigation become of vital importance. The bad news is that manure is astonishingly poor in nitrogen compared to modern artificial fertilizers, and so the amount of it that has to be gathered and spread over the fields is correspondingly enormous. It’s still worth it — the energy gained from farming intensively more than pays for the energy cost of collecting and handling huge volumes of human and animal waste, but everybody’s life gets less pleasant, and the system as a whole is much more fragile.
Irrigation, also, makes farming more laborious and more precarious, but capable of supporting larger population densities. Water is heavy, and you have to move a lot of it uphill in order to irrigate a field. The sheer number and ingenuity of labor-saving devices developed to ease the burden of moving water speaks to the energetic imperatives in play. From the ancient Egyptian shaduf, to the Archimedean screw, to the dragon-backbone machine, everybody was keenly interested in increasing irrigation efficiency.
Why go to such trouble for a relatively small percentage improvement? The answer has to do with some counterintuitive arithmetic around productivity and efficiency improvements when you’re operating close to any sort of limit. Imagine you can grow a plant that gives you 1,000 joules of energy when you eat it, but all the direct and indirect inputs into the farming process — the plowing, the sowing, the harvesting, the losses from seeds that don’t take, the extra food required to feed your donkey, the labor required to irrigate, the threshing, the milling, the maintenance of the dirt path that runs next to the field, storage losses from vermin in your granary, etc., etc., imagine all of that when amortized over your whole crop comes out to about 999 joules of energy per plant.
Now imagine that some improved practice or labor-saving device or something makes your farming process 1% more efficient. Barely worth it, right? What’s 1% in the face of this kind of Malthusian crunch? But wait a minute: previously each plant was yielding 1,000 joules and costing 999 joules, for an energy “profit” of 1 joule per plant. After our efficiency upgrade, each plant yields 1,000 joules but only costs ~990 joules, for a profit of 10 joules. So the 1% improvement in efficiency paradoxically increased our “profit” by 1,000%! And now recall that that “profit,” aggregated across your whole society, is the fund out of which you can pay for all the people who matter but who don’t contribute directly to food production — tailors and blacksmiths and priests and soldiers — and is also the fund which pays for future productivity improvements, future technological upgrades or infrastructure which yield yet more efficiency gains, and which slowly dig the whole world out from under the permanent specter of starvation.
The best part of Smil’s book is precisely when he recounts one example after another of the most mundane-seeming inventions and innovations — things like slightly optimizing the curvature of a scythe, or a tweak to the angle at which an ox’s harness attaches to its neck — and the dawning realization you feel that each of these was actually an epochal event, a heroic triumph of mankind over a brutal universe that wants us dead. You think I’m being melodramatic? I have not yet begun to get melodramatic. “Cursed is the ground for thy sake; in sorrow shalt thou eat of it all the days of thy life; Thorns also and thistles shall it bring forth to thee” (Genesis 3:17-18). There’s something downright eschatological at work here — each slightly improved plow design was a rung on the ladder to a completely different mode of existence, a transformation of man’s place in the world more profound than any of the ones we usually think about. In the pantheon of the greatest heroes of humanity, there will be a seat at the head table for all the nameless and forgotten peasants who went: “hang on a minute, what if…”
This brings me to what I think is the single most important concept introduced by this book, but bizarrely, it’s a concept that Smil himself doesn’t seem to grasp. If you thought one step ahead of my toy example above, with the energy balance associated with farming a single plant, you probably realized that the next 1% efficiency improvement wouldn’t have quite as big an effect. In fact, this is a universal rule — gains in efficiency always have diminishing marginal returns, and in fact they also always cap out at a strict horizontal asymptote, because the Second Law of Thermodynamics implies that no form of energy conversion can be perfectly efficient. If we could only grow a fixed number of plants, but needed more than 1,000 joules per plant to supply our burgeoning civilization, we would just be out of luck. The only possible solution would be to say efficiency be damned, and somehow increase the amount of energy input along some other dimension, for example by increasing the area under cultivation.
In fact, in the case of agriculture, both these models — the path of increasing efficiency and the path of just doing more of everything without worrying too much about efficiency — were tried in different parts of the world, and they’re exemplified by two different draft animals: the water buffalo and the horse. In China, India, and the densely settled islands of the Indonesian archipelago, water buffaloes and oxen were the draft animals of choice because they’re extraordinarily efficient in net energy terms. Being ruminants, they can largely just graze on grass and straw (or even underwater plants in the case of water buffalo), and so don’t require “wasting” arable land on growing feed. They have a low center of mass, which means a very high fraction of the animal’s muscle power is converted into useful work, and they have an especially low metabolic cost because an idle ox or water buffalo will just lie down.
All of this is well-suited to the general pattern of traditional farming in East Asia, which often gets described as “agricultural involution,” and which basically amounts to cropping intensification via ever more elaborate cultivation of a limited land area. The path of involution is the extreme case of fighting to eke out every last efficiency return given a fixed budget of resource inputs (in this case, arable land and photosynthetic power from the sun). It succeeds in producing some eye-popping conversion efficiencies, and in practice resulted in crop yields per unit area that were double or more the best yields in premodern European and American agriculture. But efficiency comes at a cost. High-intensity, efficient cultivation consumed an ocean of human labor, and consumed it with tasks that were simultaneously backbreaking and monotonous.
I hesitate to blame this style of farming for Asia’s failure to industrialize early, but it’s just true that classical Asian civilizations had a much higher fraction of their population engaged in grueling farm labor. It’s not just that that means fewer people sitting around able to tinker with machines: my hunch is that it also profoundly distorts and perverts your whole way of looking at the world. A culture that valorizes efficiency is one that almost definitionally is consumed by zero-sum thinking, because the whole point of efficiency is how to make do with less, as opposed to how to expand the frontier of possibilities. It’s also a culture that places a low emphasis on human dignity, since the tradeoff is constantly in the direction of substituting more human labor, even disgusting or boring labor, in a never-ending quest for extractive density. Finally and most speculatively, the path of agricultural involution necessarily results in extremely high population densities, as both the human labor inputs per unit area and the energy output per unit area go through the roof. That’s right, Asian-style farming produces bugmen.
Back to draft animals: the alternative to oxen and water buffalo is horses. A horse is a big animal, it eats a lot, you have to feed it high-quality grain, and it burns a ton of energy just standing around. Horses are very expensive if you’re a premodern farmer, but they’re also very strong. A horse can plow a field much faster than an ox can, and a team of horses can do it even faster than that. In fact, it’s enough faster that the net energy return on horse labor is still positive — that is, it plows so fast that it can plow your original land, plus all the extra land you need to farm in order to feed it. This is very inefficient in terms of land, but if you have the land to spare, then it’s hugely productive per hour of human labor.
Two groups with extremely low population densities and a lot of arable land were European colonists in Australia and North America, and these were also the practitioners of horse-based farming par excellence. The most extreme instance of this pattern was in California at the tail-end of the 19th century, where teams of dozens of horses yoked to combine harvesters could bring in a hectare of wheat in under 45 minutes. The efficiency of this approach was very low, but that didn’t matter, because the 1890s California farmer had something approaching 20 kilowatts of power available to him in that team of horses.
In general, the useful energy you can get out of some process is computed by multiplying together the raw power you can draw on, and the fraction of that energy that can be practically extracted (this second number is commonly referred to as the efficiency ratio). Thus there are two ways to make the final number go up — improve the efficiency ratio, or improve the raw power available to you (ox-based and horse-based farming are examples of these two strategies). The trouble with improving efficiency is that it has diminishing returns, because you can never get your efficiency above 100% (that nasty second law of thermodynamics again). In the end, accessing a massively larger amount of raw power is the only way to really move the needle on the energy at your disposal, and this can be worth it even if you waste a large fraction of the new energy source.
This basic pattern recurs time and again in the history of humanity, and the most famous example is the Industrial Revolution. At the time that steam engines were undergoing rapid iteration and improvement, there was already a powerful and convenient source of inanimate motor energy — waterwheels. The best waterwheels approached 60% conversion efficiency (pretty incredible for a machine with moving parts) and the largest ones produced almost half a megawatt of power. In contrast, the best steam engines had more like 2% efficiency when coal started displacing water power.5 But just as horses were less efficient but more powerful than oxen, there is just so much coal, and it’s so incredibly energy-dense, that coal-powered civilization was a giant step forward in wealth, convenience, and capabilities over water-powered civilization.
I now come to the part of the review where I regretfully inform you that while the first half of this book is a fascinating analysis of premodern life through the lens of energy generation and expenditure, the second half is a dreary assemblage of replacement-tier liberal platitudes so utterly predictable you may as well have been reading The Economist (no wonder Bill Gates likes this author so much). Some of this is downright weird — after maintaining an impersonal academic tone throughout, Smil suddenly starts dropping self-righteous asides ranging from 2005-era complaints about SUVs to bizarre rants about the evils of too many sorts of breakfast cereal, too many recordings of Vivaldi (!??!), and too many trips to the beach “to acquire skin cancer faster.” This is all so stupid I’m tempted to assume it’s self-parodic performance art, except that Vaclav Smil is clearly a man who takes himself very, very seriously. Regardless, all of this could be forgiven, if only it were the biggest problem.
Remember: this man has just written 300 pages that are nothing short of a panegyric to the slow, agonizing, deadly process whereby our species has raised itself from out of the Malthusian muck by mastering ever larger and more powerful energy flows. That mastery has brought us not only cheap and plentiful food, but also powerful and sturdy building materials, safe and rapid ground transport, the ability for ordinary men and women to soar through the air as our ancestors dreamed to do. Not to mention refrigeration. Air-conditioning! I can go on listing things, but I don’t need to, because Smil lists them all for us, and then proceeds to spit on the civilization that has made them possible and berate us for our lack of efficiency. But the entire point of this book thus far (granted, perhaps not the point the author intended to make) is that efficiency is overrated, and that it’s better just to have more energy at your disposal. More, MORE, MORE. But Smil’s embrace of this obvious conclusion is only for the distant past, his sympathy for humanity’s energy-starved condition ends abruptly somewhere in the middle of the 20th century.
And it’s not just Smil, it’s the entire political and cultural establishment here in the rapidly-dying West. Let me bring your attention to two important charts from the wonderful website “WTF Happened In 1971?”:
Since Smil has dropped the mantle of “autistic alien robot,” let’s be the ones to pick it up. If you saw charts like that for any historical civilization, you would immediately conclude that some kind of catastrophe had occurred at the moment where energy usage fell off trend. That line going up represents literally everything getting cheaper in the sense that matters — more affordable per hour worked by the average laborer. That line going up represents entirely new technological possibilities coming into being — air travel, personal automobile ownership, air-conditioning. Imagine where we could be if that line had kept going up!
Every sane civilization in history has sought to increase the energy inputs under it control, only stagnating loser civilizations focus on eking a bit more sustenance out of a fixed pile of energy. The goal should not be “efficiency”, the goal should be energy that is too cheap to meter, energy overflowing, so much energy that we have no idea what to do with it. What are the odds that some future civilization studying the collapse of 21st century America will look back and conclude that the 1970s are when the “collapse” occurred, and that everything since then was just coasting on residual momentum? Maybe this would be obvious to us were it not for the all-consuming, full-spectrum propaganda campaign aimed at convincing us that “efficiency” is the goal, up is down, poverty is wealth, and death is life.
Now I’m a Christian, so I accept the notion that poverty is sometimes wealth and I embrace as salutary the practice of asceticism, but I also think if you’re going to eat bugs you should do it because John the Baptist did it, not because Bill Gates did. More seriously, note that all traditional religious practices of self-denial are oriented towards the personal spiritual improvement of the practitioner. In contrast, the Neo-Malthusian embrace of self-denial is a horrid little mixture of self-hatred, puritanical malice towards others, and cowering propitiation of a vengeful deity (“Mother Earth”). In other words, the Cult of Efficiency’s own doctrines bear an uncanny resemblance to the slanderous folk-Nietzschean critique of Christianity. What can I say? It’s all just projection, man.
It may actually be giving these people too much credit to accuse them merely of preaching a false religion. The more cynical take is that they don’t believe any of their claptrap about running out of natural resources, it’s all just a big exercise in managed decline.6 Energy has never been more abundant, there are zettajoules of it just lying around in crustal thermal energy alone, but our exhausted ruling class is too lazy and vain and stupid to gather it, and finds it easier to bleat about efficiency. But efficiency has diminishing returns, efficiency is zero-sum thinking, efficiency is a dying old woman scrutinizing every cent in her checking account, not a hungry young civilization with free energy raining down on it like manna from the heavens. Efficiency produced the poverty and stagnation of premodern East Asian agricultural involution. Efficiency will not allow us to inherit the stars.
There are two roads, the road of efficiency and the road of plenty, but these roads actually diverged before modernity, before industrialization, even before agriculture. Go all the way back to the beginning, back to the hominids desperately trying to balance the books on their energy input and energy output, trying to gather enough food to match their basal metabolic load. Even then, our ancestors were already walking the road of plenty — our large brains are like using horses as draft animals, like burning coal. Our brains are astonishingly hungry for energy. They’re highly inefficient, but it doesn’t matter, because they increased our capacities, allowed us to grow and dominate and expand our control of other energy flows and thereby more than paid for themselves.
Not everybody walked that path, though. On the Indonesian island of Flores, a band of human-like beings was marooned and isolated, and the result was the stuff of horror movies and H.P. Lovecraft stories. The road of efficiency is a different way to deal with the harsh equations of energy balance, and Homo floresiensis began to reduce their basal metabolic load first by becoming smaller and then by giving up their brains. After a few thousand years of evolution, we see skeletons that are stunted, deformed, and with skulls that have room for brains no larger than those of a chimpanzee. How wonderful. Just think how energy efficient they must have been.
Please note that there’s a distinction between the kind of “theory-building” history I’m talking about and historical surveys. A historical survey is, at its heart, still a history of particular places and events, just one that happens to be grand in scope (hence necessarily less detailed). Theory-building doesn’t start from events, it starts from theory and attempts to recontextualize history in terms of it. Think Karl Marx, not Edward Gibbon.
The math is actually more brutal for Northern latitudes than this simple calculation suggests, because for a variety of uses you actually need charcoal rather than raw lumber, and charcoaling carries its own punitive conversion losses.
Actually, his account begins a little earlier than prehistory, with a discussion of the energy constraints on evolution itself. Bipedalism might have evolved because it results in reduced energy costs for locomotion, not for anything to do with tool use or whatever. Hominids “paid for” the extreme metabolic requirements of our brains with a significant shortening of our guts and loss of muscle mass, a development that necessitated “pre-digestion” via cooking. Humans are also adapted for extreme levels of heat dissipation (sweating, and lack of fur), which plausibly allowed us to occupy a new evolutionary niche as diurnal hunters. Probably all of these changes fed on and encouraged each other in a relatively rapid virtuous cycle.
Interestingly, Smil shares James Scott’s contrarian view on the origins of farming as expressed in Against the Grain. Briefly, sedentarism preceded agriculture, rather than vice versa. Hunter-gatherers fortunate enough to live in especially fertile places at the intersections of several distinct biomes didn’t have to move around as much, so they didn’t. Farming emerged long after these peoples settled down and built houses, and there was a long period during which foraging, hunting, and shifting agriculture all overlapped.
In fact, the gap was so great initially, coal extraction began for reasons basically unrelated to burning it in engines. The first major use of coal was in British iron-smelting — charcoal-based smelting uses vast amounts of wood, and Britain was already somewhat deforested from the demands of construction and ship-building. Iron-workers tried shifting to burning coke, and this was what created demand for steam-engines to pump out the flooded lower levels of coal mines. Since the steam engines were already right there at the mine, feeding them with coal was a no-brainer, and this was what kicked off the cycle of iterative refinement and process improvement that eventually resulted in the steam engines being competitive with waterwheels, etc.
I learned of a heart-rending example of this the other day: early pacemakers made in the 1970s were powered by plutonium decay which gave them extraordinary power density and meant they could last an entire human lifespan. Thanks to anti-nuclear hysteria, modern pacemakers come with normal batteries, and need to be replaced after a few decades, which means you might be lucky enough to have open-heart surgery in your 90s. Yes, this is a synecdoche for our entire society. Also did you notice the pun in the first sentence of this footnote?
This reminds me of John Michael Greer's idea of "catabolic collapse". See, in particular, his comments about the 1970s: https://www.resilience.org/stories/2011-01-20/onset-catabolic-collapse/
Sorry, I know I'm late to the party here; only recently discovered this splendid (and more importantly this Wodehouse-referencing...) blog. Hope nobody minds!
[TL;DR: the utility landscape is bloody complicated and whilst it's true that the high peaks represent improving access to resources rather than improving efficiency of resources we already have access to and the troughs represent 'improving-efficiency-only traps', it's far from obvious that the best next step for our particularly civilisation, for where we happen to be on the utility landscape right now, is a resource-improving rather than an efficiency-improving step.]
I'm skeptical of Psmith's criticism of efficiency-improving over resource-improving (whether that resource is energy or anything else); obviously in the long run resource-seeking must win out and efficiency-seeking must stagnate, for all the reasons described, but it's not clear to me that we're at a point in the utility-landscape where *specifically the best next step* is definitely a resource-seeking one rather than an efficiency-seeking one.
Right now, there are obviously gains to be had from both efficiency-improving and resource-improving and even if energy is theoretically unlimited (or limited only at a Kardashev-scale that's so far beyond our needs we may as well call it unlimited) it *is* limited in the practical sense that we don't have the political, social, and to a lesser extent technological, mechanisms to extract vastly more of it right this moment, whereas there are efficiency-improvements currently being implemented successfully: so, whilst we structure society, develop technology, etc. to extract more energy, why should we eschew the available efficiency gains owing to the abstract principle that resource-improving is better.
Steam engines only took over from water wheels after a period of development and refinement (made possible in part by their synergy with waterlogged coal mines, as Psmith fascinatingly relates); the technology (and the social gumption) to improve our energy resource extraction appears to need a similar period of development and refinement (and if you can measure this in terms of eg. trends in the cost-across-time of offshore wind or whatever, we do appear to be progressing through this necessary period of refinement). However - we need to make changes to our energy usage *now*, if we're to avert climate catastrophe, and the resource-improving interventions aren't entirely ready to take the strain; it seems crazy to criticise the efficiency-improving interventions that may be the only thing that staves off disaster for long enough to allow the resource-improving interventions to mature.
(By analogy: if a town's energy needs had risen by 1% -say, owing to a gold-rush-style immigration boom- literally the same afternoon that the steam engine was invented, it would be madness for the town's burghers to say "well, we *could* improve our water wheels to eke out a 1% gain to cover our increased needs and then nobody has to go hungry... but probably in a decade's time steam engines will be better than water wheels so in the meantime let's just do nothing")