1Part Four, in four paragraphs
So this part stands on its own, even if you have not read the last one.
Part Four was Harari's most famous argument. Farming, he says, was history's biggest fraud. It started independently in at least six places within a few thousand years, so nobody invented it and nobody spread it. And it did not make life better. It made human bodies break — slipped discs, arthritis, worn teeth. It narrowed the diet to one crop, which is what makes famine possible. It put people in dense permanent settlements next to their own waste and their own livestock, which is what makes epidemics possible. And it gave them a granary that could not be carried and was worth stealing, which is what makes war worth it. The reason nobody stopped is the luxury trap. Every step was small, voluntary and sensible — stay a bit longer, scatter some seed, build a hut, have another child. And the population grew to fit each step before the next was taken, so going back stopped meaning a harder life and started meaning most of the children die. The exit was blocked by arithmetic, not by force. And underneath it, the distinction I said was worth the price of the book. Evolutionary success and individual wellbeing are different measurements. Wheat won. Cattle won. Our species went from perhaps eight million to two hundred and fifty million. Every single one of those is true, and none of them tells you whether the person holding the plough was all right.
Farming produced a surplus — food beyond what the growers eat. Everything in this part comes out of that one fact.
Chapter 6 asks what the surplus bought, and then asks the question that follows: once you have a city of fifty thousand, how do you keep them cooperating? Personal trust ran out at a hundred and fifty back in Part Two. So something else has to hold the structure up.
Chapter 7 is about a problem so boring that it changed the world. A kingdom produces more numbers than a brain can hold. Somebody had to solve that, and the solution was writing — invented not for poetry but for counting barley.
One chapter is about the story. The other is about the filing cabinet. You need both, and almost nobody writes about the second one.
2Whose grain built the pyramid
A surplus does not just feed more farmers. It feeds people who are not farmers at all — and those are the people history is about.
Start with the logic, because it is simple and it explains almost everything that follows. A forager band feeds itself. That is the whole operation. No spare food means no spare people, which means everybody does roughly the same work, which means there is nobody available to do anything else. A farming village can produce more food than the farmers eat. And that extra food can feed people who do not grow food. Kings. Priests. Soldiers. Scribes. Artists. Officials. Tax collectors.
Those people build pyramids, write laws, fight wars, paint ceilings and invent religions. And every hour of it is paid for by grain taken off somebody who is bent over in a field.
History is something that very few people have been doing while everyone else was ploughing fields and carrying water buckets.
What he means Almost every name you learned in a history lesson belongs to the small minority who ate somebody else's wheat. The people who grew it left no names, no writing and no monuments, because they had no surplus of their own to spend on any of that.
That sentence permanently changed how I read a history book, and I do not think I can undo it. When you see a photograph of a great monument, you are looking at a bank statement.
And farming installed a new feeling
NThere is a mental change here too, and it is easy to skim past. Foragers mostly thought about now. This week's food, this season's movement, where the herd went. Farmers had to think about next year. Save seed for the spring — do not eat it, however hungry you are in February. Store grain for winter. Watch the sky. Worry about locusts, about the river, about the men from the next valley. Harari argues that farming installed anxiety about the future as a permanent feature of the human mind. And then he makes the connection that matters: that anxiety is exactly what the elite lived on. A frightened person hoards. A hoard can be counted, taxed and taken.
3Two documents, three and a half thousand years apart
The comparison at the heart of the chapter, and it is very hard to argue with.
ow the real question. Once you have a city of fifty thousand people, or a kingdom of a million, how do you keep them cooperating? Not by personal relationships — that ceiling was about a hundred and fifty, and we hit it in Part Two. Not by force alone. To control a million people by force you would need soldiers watching everyone, and then soldiers watching the soldiers, and you have not solved anything, you have only moved the problem. You do it with a story. Harari calls these imagined orders — shared beliefs about how society ought to be arranged, believed by enough people that everybody behaves as though it is simply how the world is. And then he proves it with two documents, written exactly 3,552 years apart.
A set of laws for around a million people. It claims to come from the gods, and it says it is describing eternal justice.
Here is the justice it describes.
Society divides into superior people, commoners and slaves. Men rank above women. And human lives have different prices.
The clearest example: if a superior man kills a commoner's daughter, the superior man does not die. Instead the killer's own daughter is put to death. She has done nothing. She is not the point. She is a unit of value belonging to her father, and the punishment is a transfer of loss between two men.
That is written down, formally, as justice.
It also claims eternal principles.
We hold these truths to be self-evident, that all men are created equal, that they are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness.
Two documents. Both certain. Both universal. Both eternal.
And they flatly contradict each other.
Harari's point is not that Hammurabi was wrong and Jefferson was right. His point is that neither is objectively true, because there is no objective fact about human worth sitting out there in the universe waiting to be discovered.
Both are imagined orders. Very different ones — one produces slavery and one produces democracy, and that difference is enormous — but both are stories that human beings agreed to believe.
And then he does the thing that annoys people
EHe translates the Declaration into strict biological language, to show what is actually there and what has been added: All humans evolved differently. Each is born with certain mutable characteristics, among these are life and the pursuit of pleasure. His reasoning. Biology recognises no equality — evolution produces difference, that is its entire mechanism. Biology recognises no rights — it recognises organs, abilities and characteristics. Biology recognises no creation — it recognises birth and mutation. There is no gene for liberty and no molecule of justice. Open a human being up and you will not find any rights inside. I want to be precise about what he is and is not saying, because he is misread here constantly, usually by people who have not read him. He is not saying human rights are worthless. He is saying they are inter-subjective — real because we collectively maintain them, not real the way carbon is real. Which means they are fragile. Which means they have to be actively defended, because nothing in nature will do it for us. That is the opposite of a cynical position. It is a demand.
4Three ways a story stays standing
The most useful part of the chapter, and the third one reaches inside you.
One: never admit it is imagined
very imagined order in history has claimed to be natural, divine, self-evident or scientifically proven. Hammurabi said the gods wrote it. Jefferson said it was self-evident. Empires said it was civilisation. The twentieth century said it was science. Nobody has ever founded anything by announcing: here is a story we made up, which I would like everyone to pretend is true. The moment people see the story as a story, it starts to lose its grip.
You can carry this one around and use it.
Whenever a system tells you that its arrangement is simply the natural order of things — that it is how people are, that it is common sense, that it is obvious — that claim is doing work.
It is not necessarily a lie. But ask who benefits from it not being examined, and you will usually find whoever is currently at the top of the arrangement.
This works on nations, on religions, on companies, on families, and on the way your own industry describes itself.
Two: teach it from birth, so it shapes what you want
This is the disturbing one, because it does not stop at controlling your behaviour. It manufactures your desires. Harari's example is romantic consumerism — the modern belief that a full life means accumulating experiences and things. Travel widely. Try new food. Buy the thing that expresses who you really are. Treat yourself; you have earned it. That feels like the most personal, most authentic urge you have. It feels like it is coming from inside you. It was installed.
A medieval European peasant did not dream of finding themselves in Thailand. The idea would not have been unaffordable to him — it would have been unintelligible.
An ancient Egyptian nobleman with a fortune spent it on a tomb, not a holiday, because his imagined order said the afterlife was the real one and this life was the corridor.
A Roman aristocrat spent it on public buildings with his name on them, because status after death meant being remembered by the city.
Same species. Same brain. Completely different deepest desires — because a different story had been installed before they were old enough to examine it.
When you feel a genuine, personal longing for a new phone or a trip somewhere, you are feeling your society's myth from the inside. And from the inside it is indistinguishable from your own soul.
I did not enjoy reading that. I still have not got round it.
Three: it lives in millions of heads at once
OThis is why you cannot escape it on your own, and it needs a proper distinction.
5Objective, subjective, and the third thing
The most useful three-way split in the book. Learn this one properly.
bjective: exists whether or not anybody believes in it. Radioactivity. Gravity. Not knowing about radiation will not protect you from it.
Subjective: exists in one person's mind. A childhood imaginary friend. It stops existing when that person stops believing.
Inter-subjective: exists in the shared imagination of a large group. Money. Nations. Laws. Companies. Gods. Human rights. Caste. Cricket records.
Why it matters: this is the category almost everything important to human beings sits in, and we have no everyday word for it. Because we have no word for it, we keep sorting these things into the wrong bin — either treating them as objective facts of nature, or dismissing them as mere opinion. They are neither.
An inter-subjective thing cannot be dissolved by one person disbelieving it. You can decide, personally and sincerely, that the rupee is meaningless. The bank will still take your house. You can decide that a border is imaginary. It is imaginary, and the guard will still stop you. To actually change money, you would have to change millions of minds at once — and to organise millions of minds you would need an organisation, and an organisation is itself an imagined order.
Which gives the chapter its closing thought, and it is genuinely unsettling. There is no way out of imagined orders. Knock down the walls of one prison and you march out into the exercise yard of a bigger one. Escape a religion into nationalism. Escape nationalism into liberal individualism. Escape that into something else. Every escape is a transfer.
Everybody in this argument — Harari, his admirers, his critics — slides between two claims as though they were one.
Claim A: these things are not found in nature. True, and Harari demonstrates it well.
Claim B: therefore no one of them is better than another. This does not follow, and it is never argued for. It is smuggled in by the structure of the comparison.
Put Hammurabi and Jefferson side by side, say "both are stories", and stop — and the reader concludes they are equivalent. But you can accept completely that rights are not written into physics and still say that a society where a man's innocent daughter is executed for his crime is worse than one where she is not. Worse by consequences. Worse by suffering. Worse by what people choose when they are allowed to choose.
"Imagined" and "arbitrary" are not the same word. Harari lets them blur, and the blur is what makes the chapter feel daring. I think it is the weakest joint in the book, and it is load-bearing.
6The most boring problem in history
A kingdom drowns in numbers. Your brain, which is superb at faces, is hopeless at numbers.
Imagined orders have a practical problem, and it is so dull that it is easy to miss how much it changed. Numbers. You are running a kingdom in Mesopotamia. You need to know how much barley each household owes. Who has paid and who has not. How many soldiers you have and where they are. What is in the royal storehouses. Who owns which field, and what happens when he dies and three sons start arguing. That is a mountain of data, and a human brain cannot hold it. Harari gives three reasons, and the third is the interesting one.
It fills up. You can hold a few hundred people, a few hundred plants, a few thousand facts. Not the grain accounts of a hundred thousand households.
It dies. When the man who remembers the property records dies, the records die with him. Everything a society knows has to be re-taught to the next generation or it is gone.
It is the wrong design. This is the one that matters. Your brain evolved to store faces, relationships, plant locations, animal behaviour and stories, and it is extraordinary at all of those. Try to recall a face you saw once ten years ago and you might manage it. Try to recall a ten-digit number you saw an hour ago and you almost certainly cannot.
The forager brain was never asked to hold numbers, so it never got good at holding them. And administration is almost entirely numbers.
7The first name in history may be an accountant
Somewhere around 3000 BC, somebody solved the problem — and what they wrote will surprise you.
Around 3500 to 3000 BC, in Sumer in southern Mesopotamia, people started making marks on clay tablets. And here is the detail that stopped me. Writing was not invented for poetry. Not for epics, not for love, not for prayers or prophecy. It was invented for accounting. The earliest surviving tablets are dull to the point of comedy. Quantities of barley. Numbers of sheep. Payments of tax. Lists of goods in a warehouse. One of the oldest readable texts we have says, roughly:
Meaning, most likely, that over thirty-seven months, 29,086 measures of barley were received, and Kushim was either the official responsible or the name of the office. And here is the thing people take a moment to absorb. "Kushim" may be the first name of a human being ever recorded in history. Not a king. Not a warrior. Not a god. Quite possibly a clerk. Harari lets that sit, and he is right to. It says something true about what history actually runs on. Before anybody wrote down a god or a battle, somebody wrote down the grain.
Partial scripts, and the cords that ran an empire
Early writing was a partial script — it could only record one narrow kind of information.
Partial script: a writing system that can only record specific kinds of data. Early Sumerian tablets could describe quantities of barley and nothing else. You could not have written a prayer, a poem or an insult with them, because the signs did not exist.
Full script: a system that can express anything a human being can think. This arrived later, around 2500 BC. Once you have it you can write laws, myths, love letters, philosophy and lies.
Why it matters: writing was not one invention. It was a slow expansion in what could be captured — and for the first thousand years, the only things worth capturing were owed.
The most remarkable partial script was the Andean quipu — bundles of coloured cords tied with knots, where the colour, the position and the type of knot each carried meaning. No marks. No clay. No paper. Just string.
And with it, the Inca administered an empire of somewhere between ten and twelve million people across the Andes. Taxes, censuses, storehouses, labour obligations, the movement of goods over mountain roads.
When the Spanish arrived they could not read quipu at all. At first they had no choice but to rely on local specialists who could — which meant the specialists could have told them anything, and on some occasions did.
So the Spanish gradually replaced the cords with Latin letters and Arabic numerals.
The quipu stopped being read. The specialists grew old and died. And the knowledge of how to read them died at the same time, because it had never been written in anything else.
Today only a fraction of surviving quipu can be interpreted at all. An entire empire's records became decoration.
That is not just a loss of information. It is a demonstration of something this whole series keeps running into: the archive that survives is the archive of whoever won.
8What writing did to the inside of our heads
The most interesting claim in the chapter, and the origin of a frustration you have felt this month.
Once information lives outside the brain, you have a problem that never existed before: finding it again. A brain retrieves things by free association. You smell something and a memory arrives sideways, connected by mood or coincidence or a song. That is a wonderful way to think and a catastrophic way to run a tax office. An archive needs something else entirely. Drawers. Shelves. Categories. Indexes. Catalogue numbers. Everything must go in exactly one box, and everybody must agree which box. That is bureaucracy, and it requires a specific kind of mind — one that chops the world into artificial compartments and keeps them separate.
Reality does not come in drawers.
A sick child is at the same time a medical case, a family crisis, a school absence, a transport problem and a money problem. It is one situation.
A bureaucracy has to file that child under one heading, and send you to a different counter for each of the others, and each counter can only see its own drawer.
Everybody has felt this. Being bounced between departments. Being told your problem is not this office's category. Filling the same details onto four forms because four systems cannot see each other.
That frustration is five thousand years old. It is the price of being able to store more than one brain can hold, and there has never been a version of it that does not have this cost.
And the story runs directly into your pocket.
Numbers eventually got their own script. The digits we use — 0 to 9, usually called Arabic numerals although they came originally from India — turned out to be an extraordinarily efficient way to write quantity, and mathematical notation grew on top of them.
Today enormous parts of the world are described in a language most people cannot read. Statistics. Equations. Models. Interest rates. Probabilities. Economics, physics, engineering and medicine all run on it.
And the newest script is binary — everything reduced to ones and zeros, which is what your phone is thinking in right now.
Harari's observation is unsettling and I think correct: more and more of the decisions that shape human lives are being made in a script almost no human reads directly. We invented writing to hold information we could not remember. We are ending up with writing we cannot read.
Everybody treats the written record as a record of what happened. Historians know better in principle, and then write as though it were true anyway, because it is the only material there is.
Look at what Chapter 7 has just told us. Writing was invented to count things somebody wanted counted. The earliest archives are tax records, storehouse inventories and property claims.
Which means the archive is not a window onto the past. It is a record of what an authority considered worth counting — and everything it did not care about is simply not there.
We have the barley receipts of Sumer. We do not have what the people who grew the barley thought about the man collecting it, because nobody with a stylus considered that a quantity.
So the deepest bias in history is not that the record is incomplete. It is that the record was created by the act of administration, which means the past arrives to us pre-sorted into the categories of the people who were taxing it.
Harari shows us the machinery and does not turn it on his own sources. Neither, mostly, does anybody else.
9Where I think he is wrong in these two chapters
One of these chapters is nearly bulletproof. The other has the weakest joint in the book.
1. "Imagined" is doing the work of "arbitrary"
Imade this the hidden assumption above because it is the most important thing in this part, but it belongs here too. Harari establishes brilliantly that imagined orders are not found in nature. He then structures the chapter so that the reader concludes they are therefore all equivalent — and he never argues for that step, because it cannot be argued for. You can hold both of these at once: that human rights are a story we tell, and that it is a much better story than the one Hammurabi told. Most moral philosophy since the eighteenth century is about how you do that. Harari does not engage with any of it. He gestures at the gap and walks past.
2. Surplus does not automatically produce kings
Harari's version. Surplus food feeds people who do not farm. Those people become an elite. The elite needs a story to justify itself. Therefore states, kings and imagined orders.
The problem. Several societies had substantial food surpluses and stayed stubbornly flat for a very long time. Some built large communal structures with no evidence of a ruling class, no palaces, and no rich burials. Others seem to have alternated — hierarchical in one season, when large groups gathered, and egalitarian in another.
What that suggests. Surplus is a precondition for hierarchy, not a cause of it. It makes kings possible. It does not make them inevitable, and a number of societies appear to have noticed the possibility and refused it.
Where it stands. This has become one of the liveliest arguments in archaeology, and the picture is now considerably messier than the tidy sequence of farming, surplus, cities, kings that most of us were taught and that Harari repeats.
Why it matters. If hierarchy is inevitable once you have food, then complaining about it is like complaining about gravity. If it is a choice that some societies made and others declined, then it is politics — and politics can go differently.
3. Writing-for-accounting is true in Mesopotamia and not everywhere
The claim that writing was invented for administration is well supported in Sumer. But writing arose independently in several places, and the Mesopotamian pattern does not obviously hold in all of them. Early Chinese writing appears heavily on oracle bones — questions put to ancestors about harvests, illness, war, and whether a particular day was auspicious. That is much closer to religion than to bookkeeping. Harari generalises from the case that best fits his argument. By this point in the series that is a familiar move, and it is worth naming every time.
4. "Kushim is the first name in history" is a lovely line and genuinely uncertain
Scholars disagree about whether Kushim is a personal name or the title of an office, and about whether the tablet really is the earliest. It is a plausible reading presented as a fact, and it is presented that way because it makes a much better sentence.
5. And where he is simply right
The three mechanisms — deny it is imagined, install it before people can question it, keep it in millions of heads — are excellent, portable, and I have not found a counter-example. The three-way split between objective, subjective and inter-subjective is the most useful distinction in the book, and I think it should be taught in schools. And Chapter 7 is nearly bulletproof. The argument that complexity forced the invention of external memory, and that external memory forced categories, and that categories are why every institution feels the way institutions feel, is tight from end to end.
Surplus food buys people who do not grow food — kings, priests, soldiers, scribes — and those are the people history is about. Every monument is stored peasant labour.
Above a hundred and fifty people, only a story can hold a society together. Nations, religions, laws, money, companies. All of them are imagined orders.
An imagined order survives three ways: it never admits it is imagined, it is installed in you before you can examine it, and it lives in millions of heads at once so you cannot leave it alone.
Objective, subjective, inter-subjective. Gravity, your dream, and money. Almost everything humans argue about sits in the third box, and we have no ordinary word for it.
Writing was invented to count barley. The first name in recorded history may belong to a clerk — and because the archive was built by administration, the past reaches us already sorted into the categories of the people who were taxing it.
10Words from this part
Everything I stopped to explain, gathered up.
| Word | What it means |
|---|---|
| Surplus | Food produced beyond what the growers eat. It buys everybody who is not a farmer. |
| Imaginedorder | A shared story about how society should be arranged, believed by enough people that everybody acts as if it were simply true. |
| Objective | True whether or not anybody believes it. Gravity. |
| Subjective | True inside one person's head. A private grudge. |
| Inter-Subjective | True because a large group jointly believes it. Money, nations, laws, gods, caste. |
| Codeofhammurabi | Babylonian laws, about 1776 BC, claiming divine authority for a society of superiors, commoners and slaves. |
| Cuneiform | The wedge-shaped writing of Mesopotamia, pressed into clay with a reed. |
| Partialscript | A writing system that can only record certain kinds of information — quantities, but not sentences. |
| Full Script | A system that can express anything a person can think. From around 2500 BC. |
| Quipu | The Inca system of knotted, coloured cords, used to run an empire of 10–12 million people. Mostly unreadable today. |
| Bureaucracy | Running an organisation by written record and fixed category. The unavoidable cost of storing more than one brain can hold. |
11The timeline so far
Carried forward, with these two chapters added.
| When | What |
|---|---|
| 9500 BC | Göbekli Tepe. Farming begins in the Middle East. |
| 3500 BC | First cities in Sumer. Writing appears — as accounting. |
| 3000 BC | The Kushim tablet. Possibly the first recorded human name. |
| 2500 BC | Full script: writing that can express anything. |
| 1776 BC | The Code of Hammurabi. |
| 1000 BC | Money, empire and universal religion begin to appear. Part 6 starts here. |
| 1776 AD | The American Declaration of Independence — the same claim to eternal justice, the opposite content. |