Chapter 2: The apprentice and the slave-seller

The lake is calm again. The pharaoh has his pendant. The boating party has resumed. The Hyksos-period scribe whose hand we are not going to be able to name is, presumably, brushing palace dust off the roll and considering what to copy next.

I want to spend this chapter on what was already on his desk before he started — the four thousand years of copying that reached him with the Sneferu tale already grinning, and the parallel four thousand years of similar copying that, in places he could not have pointed to, in scripts he could not have read, was reaching other scribes with other things to grin at. The argument I am going to land is that the joke was already polished before any of these people had a roll in front of them. The joke was already engineered. It was already in stock. Writing did not make the joke. Writing was the first thing slow enough to keep it around long enough that, four thousand years downstream, we can still read it. The joke is older than the medium. The medium had to catch up.

The first case is, of course, on a tablet.

The wet clay

The Sumerian proverb you met in Ch1 is from a collection edited by Bendt Alster as Sumerian Proverbs Collection 1; the fart-of-the-young-woman is item 1.12 (Alster 1997). The collection sits inside a curriculum that included, by current scholarly consensus, around two dozen proverb collections in varying states of preservation, and the curriculum sits inside an institution that Assyriologists conventionally call the eduba, "tablet house," though specialists periodically remind one another that the term is a moving target across the third and second millennia BCE and that a real eduba in Nippur in 1900 BCE was not necessarily what eduba meant in Ur three centuries earlier (Veldhuis 1997; Robson 2001). What it meant in 1900 BCE Nippur is the period this section opens in, because that is where the specific tablets that preserve our specific proverb were produced.

I want to camp on a scene. I am going to be honest that the scene is reconstructed. Every physical detail in it is something the tablets directly attest, and the reconstruction sits at the seam between the surviving objects and what an apprentice's morning was made of. I want the seam to be visible.

It is the dry season in Nippur sometime in the early second millennium BCE. The school is a courtyard or a side-room of a private house, the kind of architecture preserved best at sites like House F at Nippur, which a Penn excavation in 1951 recovered with its floor paved in discarded school tablets the way an English schoolroom's floor used to be paved with chalk dust (Robson 2001). The sun is up. A teacher — in the eduba literature he gets the title ummia, "expert" — sits with a small stack of working tablets and a finished exemplar he has propped against a brick. He has thirty or so apprentices in front of him, ranging from boys learning their first cuneiform signs to senior students who can already write a contract for the local livestock market and are now copying out the proverbs because the proverbs are what one copies in the second year of training.

The boy I want to focus on is twelve years old, give or take, and on his second year. We do not know his name. We know that some of his peers' names survive — the Sumerian scribal habit of writing one's name and the date in a colophon at the bottom of a finished tablet means apprentices' names are buried in the secondary literature in considerable density — but for the boy on this particular morning, copying the fart proverb, the name has not survived. The seat has. So has the lentil.

The lentil is what apprentices used for single-line writing exercises. It is a circular tablet, hand-sized — small enough to sit in the palm of a twelve-year-old, which is the relevant ergonomic fact. The clay is unfired and still slightly damp. The surface, where the boy's thumb rests on the underside, is matte and fingerprint-pocked, the colour of riverbed silt — the same Tigris-Euphrates alluvium that paved the schoolroom and the canals the boy would in due course be paid to count the harvest from. The teacher has written one line in his own hand on the obverse — the model — and the apprentice's job is to copy that line, in his own hand, on the reverse. Or, on slightly larger tablets in the same format, the model goes on the upper half and the imitation goes on the lower half, with a horizontal divider scratched between them so a corrector can see where the boy's work begins. The model on this morning's lentil — and we know this from the dozens of duplicates that have survived — is fourteen cuneiform signs across two clauses. Translated by Alster: Something which has never occurred since time immemorial: a young woman did not fart in her husband's embrace.

The apprentice, I would like to think, looks down at his lentil and almost smiles. He has been waiting for this one. The teacher knows he has been waiting for it — every cohort waits for it, every cohort gets it eventually. The teacher does not look up. He is going to let the boy laugh at it, but not yet, and not first.

I want to dwell on what is happening to the boy's hand. The Sumerian for "fart" — Alster reads the verb as šer, conjugated as a non-finite form for the construction — is not a word a literate Sumerian-speaker would have heard often in any written register. It is not impossible: the Sumerians were not delicate, and the Mesopotamian medical literature, whose physicians cheerfully diagnose by the smell of breath and the colour of stool, has no problem with the body. But in a school context, copied onto a tablet that will be turned in to an ummia who may smack it with his palm if the wedges run crooked, the word is sitting in front of the apprentice in a register he is being asked, very seriously, to render. He cuts the wedges. The reed in his hand is held the way the ummia taught him in his first month — thumb braced along one face of the shaft, three fingers curled around the rest, the cut end angled some thirty degrees off vertical so the wedge takes a clean impression and does not drag the clay behind it. The grip is muscle now. He is no longer thinking about it. He is thinking about the punchline. The corrector — the ummia himself or a senior student called the adda eduba, "elder of the tablet house" — comes round behind the apprentice's right shoulder while the clay is still wet enough that a misformed sign can be erased with a wet thumb and re-impressed before the surface dries. The apprentice has, on the obverse, copied the second sign one wedge short. The corrector reaches over with his own stylus and re-impresses it while the boy is still holding the tablet. The corrector's stylus comes in at a flatter angle than the boy's, the cut end closer to the surface — a geometry settled into over decades of doing this, by which the ummia re-impresses a sign without lifting the clay any deeper than the original wedge's depth. The correction sits flush. Above the line, in a smaller hand left over from another apprentice's session, the Akkadian gloss for the harder Sumerian verb has not been wiped clear: the translation is sitting there, in the boy's eyeline, for the learner who is no longer being raised on Sumerian at home. He says nothing. The boy is nine wedges into the next clause when the corrector's hand withdraws, and the boy realises only at that point that the laugh he was holding has not gone away.

This is the scene I want to give Chapter 2 its first impact moment on, because the scene contains, in compressed form, almost everything the chapter is going to claim. The clay is real. The misformed sign on wet clay is something we have, on tablets, every kind of evidence for: extant exercise tablets routinely show corrections in a different hand, in fresh clay, with the corrector's stylus angled differently from the apprentice's (Robson 2001). The lentil format is real. The named apprentices in colophons are real. The bilingual gloss — the small-handed Akkadian translation written above the harder Sumerian lines for the student who, by 1700 BCE, is no longer learning Sumerian as a mother tongue but as a prestige second language — is real. The proverb is real. And the boy's grin is, on the available evidence, the only thing in the scene I cannot prove.

I would like to say, gently, that I think the grin is real too. I think the grin is real in the way the wet thumb on the misformed sign is real. We do not have the grin on a tablet. We have the next twenty centuries of curriculum that included this proverb, and the next twenty centuries of similar proverbs adults kept in the curriculum despite no shortage of grammatically interesting model sentences they could have replaced them with, and the named colophons of the apprentices who copied the curriculum down, and we have what seventy years of Mesopotamian-pedagogy scholarship has now told us: that the eduba curriculum was not designed by ascetics. It was designed by people who had been twelve once. The proverbs that survived are the proverbs that worked.

I want to spend a moment on what worked meant. There are a few things going on simultaneously.

The first is mechanical. A boy who is going to spend several hours a day for the next decade pressing reed wedges into clay needs to remember thousands of sign combinations. The proverbs are mnemonic. Each one is a complete grammatical unit, ten to twenty signs long, with internal rhythm — paired clauses, parallel verbs, a rhetorical figure that flips the second clause into commentary on the first. They are good vehicles for sign practice because the boy will copy each one many times in a year, and after three or four repetitions of something which has never occurred since time immemorial, the formulaic opening is committed and the back half — the punchline slot — is the variable, the part the boy is paying attention to. Comedy has, since approximately forever, been engineered around exactly this distribution of attention: a setup the listener tracks loosely, a punchline the listener tracks precisely. The eduba curriculum exploits this, presumably by accident, presumably by good teaching, presumably both.

The second is moral. Sumerian proverbs are largely sententious — the proper care of livestock, the proper conduct toward the gods, the importance of not lying to a tax collector — and in any sufficiently long collection of sententious material, the entries that anchor the curriculum are the entries the apprentice cannot pretend not to have read. A young Sumerian schoolboy can pretend not to remember the prescription on the proper treatment of one's elders. He cannot pretend not to remember the line about the impossibility of a young woman not farting in her husband's embrace. The proverb sticks because the proverb is funny, and the curriculum sticks because the proverb sticks. The mechanism by which a body of learnable material installs itself in the heads of children is a mechanism pedagogy has known about for as long as there have been children to teach. The Sumerians had figured it out by the year 1900 BCE. Anyone who has ever sat in a classroom with a memorable joke in it knows the bench-press version of the same insight without having to be told.

The third is that the joke is, and was, a joke. This is the point at which modern scholarship occasionally bends backwards to insist that ancient texts must have meant something other than what they obviously meant. An older line in Mesopotamian studies treated the proverb as a coded marriage maxim ("women are universally undignified — a useful lesson for the husband-in-training"); a newer line treats it as a bawdy aside an ummia slipped into the curriculum to leaven a dry day; a third line, perhaps the cleanest, treats it as a joke about jokes — a self-aware paradoxical-formula deployed within a paradoxical-formula tradition that filled out an otherwise well-behaved proverb collection with deliberate blasts of body humor in scattered slots, to keep the apprentices laughing every fifth or sixth tablet. Alster's edition, which is the edition this section is leaning on, takes the third line. The proverb is not coded. The proverb is not slipped in. The proverb is curriculum. It is comic curriculum. The Sumerians had a comic curriculum in the year 1900 BCE.

I want to flag, since this is where the chapter is at its most argumentative, that I am not arguing that all ancient material was secretly comic. Most of it was not. Most of any culture's writing in any era is not comic. What I am arguing is that the comic register was a known register in the eduba, and that the proverbs that anchor the eduba curriculum include several whose only obvious purpose is to make a twelve-year-old laugh. That is a fact about the curriculum. The fact about humor, which the chapter is trying to get to, is downstream of it. If the curriculum used humor as a load-bearing element, then humor was already, in 1900 BCE, polished and reproducible and teachable to the next generation. Polished, reproducible, teachable: the same three properties that the printed joke book, and the recorded silent film, and the algorithmic meme will rest on. The eduba is the first edition of the apparatus.

The other thing I keep not getting over is the chain.

Sumerian was, by the time the apprentice on the morning above is copying his proverb, already a dying mother tongue. Mesopotamian scholarship places the rough end of native Sumerian-speaking populations somewhere in the late third millennium BCE — which is to say, at least two centuries before our boy. By 1900 BCE the language he is being asked to render is a language his grandparents almost certainly did not speak at home, and the language he is being trained to write because the prestige tradition of Mesopotamian scribal literacy is going to demand it of him for as long as he wants to stay in his profession. Akkadian — the Semitic language now spoken in the streets of Nippur — is what he speaks at lunch. Sumerian is what he writes.

That is itself a startling thing for the curriculum to contain a fart joke at all. The whole apparatus of eduba learning, by the early second millennium BCE, is already an exercise in copying the prestige texts of a culture nobody was native to anymore. It is the kind of curriculum an outsider would expect to be reverent — Latin in the European Renaissance, Sanskrit under colonial India, Tibetan in modern Western Buddhist seminaries — and the easy guess is that the older the prestige language gets, the more solemn the things one is asked to copy in it. The Sumerian school did not get that memo. It kept the fart joke in. It kept dozens of similar items in. The corrector's hand on the wet clay, in the year 1900 BCE, was correcting a sign for a verb that no longer existed in the apprentice's spoken language, in a sentence whose punchline the apprentice would have understood faster than he understood the rest of the grammar, because he had been twelve and married couples have always been married couples.

The chain runs forward from there as well. Sumerian remained a curriculum language in Mesopotamian scribal schools for the better part of fifteen hundred years after our apprentice's morning. The Old Babylonian curriculum migrated, with modifications, into the Middle Babylonian and then the Neo-Assyrian curriculum; the Library of Ashurbanipal at Nineveh, founded around 650 BCE, contains tablets of Sumerian proverbs being copied by Assyrian apprentices roughly twelve hundred years downstream of the morning we just imagined (Robson 2014). The proverb survived. Other things did not. The eduba lost some of its Old Babylonian compositions to attrition; the Akkadian gloss thickened over the centuries until in some Middle Babylonian copies the Akkadian is the running text and the Sumerian is the headword for a learner's reference; the political horizons changed three or four times. The fart proverb did not change. It was not changed. The Akkadian gloss says the same thing the Sumerian says. The corrector's hand kept correcting the same sign on wet clay across no fewer than thirty generations of apprentices.

I once tried to count the generations. The number depends on how you draw the boundaries of "the curriculum," but the order of magnitude is between thirty and sixty. Each generation contains some apprentices who lost the joke and some apprentices who got it. Over thirty to sixty generations the loss rate is below the gain rate, because the proverb survives. Over the same period the eduba loses nine-tenths of its Old Babylonian inventory — most of the proverb collections are now badly fragmentary; many are duplicated only in school exercise scraps; a few are gone entirely except for incipits. The fart proverb is in better shape than most of the curriculum it once sat inside. The shape it is in is the shape that comes from having been laughed at, copied, laughed at, copied. The chain of survival is a chain of grins.

I want to stop on that line because I want to mean it. There is a sentimental version of the chain of survival is a chain of grins that is too easy. The sentimental version is that the past was a brotherhood of grinning twelve-year-olds. It was not. The past was, generally, brutal — the eduba texts on the matter of beatings are candid, and the Mesopotamian polities the apprentices were being trained to administer ran on slavery, conscription, debt bondage, and the kind of casual violence that would today be called a war crime. There were grinning twelve-year-olds. There were also twelve-year-olds being beaten with reed canes for crooked wedges. The grin is not a defence brief for the world the apprentice sat inside. The grin is a piece of evidence about what survived. What survived is not the world; what survived is the grin. The grin is a small piece of data about an inner life, and it is more data than we have for almost anyone else in 1900 BCE.

By the end of the morning the boy has fourteen wedges per clause, the corrected sign is dry on the obverse, the laugh has come and gone and come back, and the corrector — who is also smiling, somewhere underneath the ummia's standard frown — is on his way to the next apprentice. The lentil tablet goes into a basket. The basket goes into a corner. Some of the lentils go home with the apprentices for re-use as drinking-vessel tops or sling stones. Some of them, against the run of probability, end up swept into a foundation deposit when the schoolroom is renovated, and lie there for three thousand seven hundred years until a Penn excavator scrapes the dust off House F's flooring with a small brush and finds them.

It is from those lentils, and the larger tablets that came up with them, and the dozens of similar finds at Ur, Sippar, Tell Haddad, and a half-dozen other sites, that Bendt Alster collated the proverb collection that contains, as item 1.12, a Sumerian sentence about the impossibility of a young woman not farting in her husband's embrace. Alster published the edition in 1997. He died in 2009. The proverb is older than his career and older than any career; it is older than England; it is older than the alphabet; it is older than the verb to laugh in any language any twelve-year-old in any modern classroom would recognise. It is also a joke. The joke is the same joke a twelve-year-old in Nippur in 1900 BCE was asked to copy. The continuity is not an idea. It is a sentence on a small piece of clay, and the small piece of clay survived because somebody, at the end of every link in the chain, decided to keep it.

What the scribe was holding

I want to come back briefly, now, to the scribe Ch1 introduced — the Egyptian one, in the late Hyksos period, with the Sneferu story in front of him. Ch1 watched him copy. I want to watch what he was copying from.

The Westcar Papyrus, viewed in Berlin's reserve collection today as Papyrus Berlin 3033, is one papyrus roll, about five feet long, written in hieratic — the cursive Egyptian script that scribes wrote in the way the Egyptian elite spoke in: faster than hieroglyphs, less ceremonial, more economical with the reed pen. The hand on the roll is a single scribe's. Lichtheim says so, Lepper agrees, and the textual paleography places the copy somewhere in the late Second Intermediate Period — that is, somewhere on the far side of 1600 BCE. The story it tells, however, is older. The Sneferu episode is set in the Old Kingdom court four or five centuries earlier than that, and the linguistic features of the surface text — particularly the verbal forms in the framing narrative — suggest the copy is itself based on a Middle Kingdom original, perhaps Twelfth Dynasty (Lichtheim 1973, pp. 215–217; Lepper 2008). The chain of copying that reached the Berlin scribe with this scroll in hand was, in other words, already at least three centuries deep by the time he picked up his brush.

It has gone considerably deeper since. The Hyksos copy was set down. Whatever was on the Berlin scribe's desk when he was finished — the ratty old original he had been copying from — went elsewhere; we cannot now name it. His own roll went, after some interval, into a tomb or a temple library or a wealthy private collection, and lay buried in Egyptian sand until an English traveller named Henry Westcar came through Egypt in 1823–1824 and acquired it. We cannot now reconstruct the acquisition beyond the fact that Henry Westcar was the kind of British traveller of the 1820s who acquired papyri without exactly worrying about the legal apparatus of acquisition. Westcar gave the roll to the Egyptologist Karl Richard Lepsius in 1839. Lepsius transferred it to Berlin in the late 1850s. Adolf Erman published the first scholarly edition, Die Märchen des Papyrus Westcar, in 1890. Verena Lepper's 2008 monograph remains the most recent linguistic edition. Each of those names is also a name in the chain.

I want to pause on the chain because the chain is the point.

The Westcar roll, when it reaches us, is incomplete. The first part is missing. The text begins mid-line, inside what is clearly already a frame narrative. The frame is this: King Khufu, second pharaoh of the Fourth Dynasty (the builder, in due course, of the Great Pyramid), is sitting in audience and is bored. His sons are entertaining him with tales. Each tale is about a wonder worked by a magician at the court of one of his predecessors. The first tale — the part that is missing in our copy — appears to have been told by Prince Dedef-Hor, and was set in the court of King Djoser, the Third-Dynasty builder of the Step Pyramid; the resemblance to the surviving fragments of related cycles suggests it was a tale of a wax crocodile come alive. The second tale — the one we have — is told by Prince Khafre, set at the court of his grandfather Sneferu, and concerns the bored pharaoh's lake. The third, told by Prince Bauefre, is set at the court of King Nebka, and is a different wax-crocodile-come-alive tale. The fourth, told by the youngest son Hardedef, is about a magician named Djedi who can reattach severed heads to bodies and who lives, at the time of the telling, in Khufu's own kingdom. Khufu sends for him. The fifth tale, which collapses into a prophecy about the founding of the Fifth Dynasty, ends our text where the papyrus does — abruptly, mid-prophecy.

It is a frame narrative, in other words, in the structural sense the term means in literary criticism: stories nested inside a story whose audience is itself part of the storyworld. Khufu is the in-story listener. We are the out-of-story listeners. The narrator — the Berlin scribe, copying his older copy — has put us in audience with Khufu, who is in audience with his sons, who are telling tales about their grandfather and great-grandfather. The Sneferu story is the second of the tales. It is told to entertain a king. The king it is told to is the great-grandson of the king in the story, and the story's first move, before any of the action begins, is to inform the great-grandson that the ancestor's mental state at the start was, more or less, Sunday afternoon, palace, no plans.

Ch1 made the case that this is a deliberate piece of storytelling — that the deadpan procedural register was the joke. I am going to take that as proven and move on to the question Ch1 did not answer, which is who decided to copy this? The honest answer is several people, in a chain stretching from the Twelfth Dynasty to the Berlin scribe to Lichtheim to me, and not one of those people had any obligation to keep the line about the lake. They could have left it out. The scribal habit of dropping individual passages between copies of long literary papyri is documented across the Egyptian corpus; the Tale of Sinuhe, the Eloquent Peasant, the Shipwrecked Sailor, all show variants between manuscripts where individual passages have been compressed, expanded, or simply lost. None of the magicians' tales is sacred text. The scribes were not under instructions to keep them word for word. They were under instructions to copy them, and copy them they did, and the variants that survive show that the scribes were exercising judgement about what to keep. They kept the bored pharaoh and the fishing-net rowers. They kept the lost pendant. They kept the negotiation. They kept the line where the woman refuses the substitute.

The scribes were people. The scribes had been twelve-year-olds. The scribes — I am willing to bet money — laughed at the line about the pendant the same way I laugh at it now, and copied it because they laughed at it. The Sumerian boy on his lentil and the Hyksos scribe on his hieratic roll are doing the same thing on different days, in different mediums, in different writing systems, in different cities, in different millennia. They are keeping the joke.

The thing that is unusual about the Westcar Papyrus — among Egyptian literary papyri — is its concentration of what we would now call narrative comedy. Most surviving Egyptian literary texts are reverent: hymns, royal autobiographies, Maxims, the Instruction of so-and-so to his son, prayers to the Aten, hymns to Hapi. The Westcar tales are a different register. They are stories told for pleasure, by sons to a father, with magic in them and ridiculous-prescription humor and at least one sequence that lands as a small comedic riff on the third item of a list. (The magician Djedi, with the severed-head trick, refuses to perform it on a human prisoner — not on a man, your Majesty; not on the noble cattle — and substitutes a goose, then a duck, then an ox; the joke is the substitution, and it is the joke of a writer who knows comedy lives in the third item of the list, not the first. Lichtheim 1973, p. 219, has the substitution sequence.)

The Westcar Papyrus is, for our purposes, the third witness. It is not the first writing of the Sneferu tale. It is somebody's copy of somebody's copy. Behind it sits a Middle Kingdom version. Behind that sits — necessarily — an oral tradition of magician-tales told at court for centuries, of which the Middle Kingdom version is one written-down crystallisation. The boundary between the tale was first told and the tale was first written down is unrecoverable from the evidence we have. What we do have is the latter half of the chain: a line of named hands and unnamed hands going from the Twelfth Dynasty to a Hyksos-era table somewhere in northern Egypt to the Berlin Neues Museum to my desk in 2026.

This kind of chain is, in the literature on textual transmission, called a recension if the scribes are working from a known archetype, or a tradition if they are not. The Westcar tradition is the latter. It is a tradition of magician-tales in which the Sneferu story is one node, and it survives at all because, at every link, somebody kept it. The number of links is staggering and, since most of the links are lost, also unreconstructible. What we can say is that the surviving copy is enough to establish that the comedic structure was already finished by the time it reached the Hyksos scribe. The deadpan register was already in place. The bored-rich-guy energy was already running. The pendant negotiation was already the longest beat in the scene. The Hyksos scribe found those elements in the older copy and kept them. He did not invent them. He found them. He kept them because they worked.

I have been saying this in different forms for a few pages, and I want to consolidate what the consolidation is for. The case I am building, item by item, is that comedy was, in pre-text antiquity, a finished thing. Not a finished thing in the sense that no new jokes were ever added to the curriculum — they were, dozens of times in every century — but in the sense that the recurring comedic structures, the deadpan register, the paradoxical-formula opener, the procedural prescription on undignified content, the surprise-via-specificity move, were already engineered and recognisable to scribes across at least two of the great riverine civilisations. The Sumerians had them. The Egyptians had them. They had them, in some cases, in identical structural forms — both the Sumerian things which have never happened formula and the Westcar's let there be brought to me twenty nets prescription are deadpan registers stretched over absurd content, and both are a fully formed comedic instrument, not an experimental one. The scribes of 1900 BCE Nippur and 1600 BCE Avaris were already practitioners, not pioneers. The instruments had been tuned by audiences before them and were now being passed forward by training and transcription.

I want to close this section with a remark about who the people in this chain were. They were not common people. The Sumerian apprentice was being trained, on the day of the lentil, to enter the literate-bureaucratic minority of an early second-millennium BCE Mesopotamian state — the people who counted the grain shipments, drafted the contracts, witnessed the marriages, and read the omens. The Egyptian scribe copying the Westcar roll was a member of the analogous Egyptian elite. The percentage of either of these populations that could read and write at all was, by current estimates, in the low single digits. The proverbs and the magician-tales, when they were written down, were being written down for an audience that could read them. That was a small audience. Below it sat ninety-eight percent of the population, who were illiterate, who were also presumably laughing at jokes none of which we now have because none of them got written down.

I want to be clear that the case the chapter is building is conservative. The cases I am walking are the cases for whom we have written witnesses. They are necessarily a sliver of what the comedic life of the relevant cultures actually looked like. The fart proverb is not the first joke; it is the first joke a literate adult thought worth keeping. The Sneferu tale is not the funniest tale told in pharaonic Egypt; it is one of the few funny tales whose wear-and-tear of survival and re-copying carried it down to us. What we have is the residue. What was alive in the rooms of the cultures these residues come from was, almost certainly, several orders of magnitude more.

That residue, however, is enough. The case slate is moving now from Mesopotamia to Egypt to China, and the next case is one in which we can name not just the joke and the scribe but the comedian. He is a man named You Meng. He worked for a king named Zhuang of Chu sometime around 600 BCE. He is the earliest named comedian in any literature I know of, and he survives because Sima Qian, the great Han historian writing five centuries later, decided he was important enough to include in a chapter of the Shiji whose entire purpose was to record people like him.

You Meng walks in

The chapter of the Shiji I want to bring in next is Chapter 126. The Shiji is the Records of the Grand Historian — Sima Qian's massive history of China, completed around 94 BCE, running from the Yellow Emperor down to Sima Qian's own time. The work is one hundred and thirty chapters long. The first twelve are basic annals; the next ten are tables; the next eight are treatises; chapters thirty-one through ninety-five are the hereditary houses; and chapters sixty-one through one hundred and thirty are the liezhuan — "arrayed traditions," or, more loosely, "biographies." The biographies are organised, at their best, by category. Chapter 124 is Biographies of the Wandering Knights. Chapter 125 is Biographies of the Imperial Favourites. Chapter 127 is Biographies of the Diviners. Chapter 126 is Guji liezhuan. The literal translation is Biographies of the Jesters; less literal renderings include Biographies of the Wits and Humorists. Guji (滑稽) names a particular kind of skilled performer who deployed indirection, wordplay, and impersonation to deliver moral or political commentary at a court, under the cover of being amusing.

Sima Qian, in other words, wrote a history of comedy in 94 BCE.

I would like to say that again because I have not yet recovered from saying it the first time. The earliest known historian of any literate civilisation to dedicate a chapter of his major work to the lives and methods of comedians is the same historian who is, by general consensus, the founder of Chinese historiography, the model for two thousand years of subsequent Chinese historical writing, and one of the most important prose stylists in any language that has had two thousand years to argue about prose styling. He thought comedians were worth a chapter. The chapter is not an aside. It sits, in the original organisation of the Shiji, sandwiched between the imperial favourites and the diviners — between, that is, the people who ran the inside of the palace and the people who ran the inside of cosmology. The jesters got the chapter between them.

I want to walk one of the biographies in Chapter 126, because the biography in question is the source case for one of the recurring comedic figures the rest of this book is going to keep meeting. The figure is the comedian-as-impersonator-of-power. The biography is the biography of You Meng.

You Meng — yōu (優), the role-noun for actor or entertainer, plus Méng (孟), his name — was a court entertainer in the state of Chu in the seventh century BCE. Chu was one of the larger states of the Spring and Autumn period; its capital, at this point in its history, was Ying, in what is now Hubei province along the Yangtze. The king on the throne during You Meng's tenure was Zhuang of Chu, a real historical figure who reigned from roughly 613 to 591 BCE, who is one of the so-called Five Hegemons of the Spring and Autumn era, and who is in the Chinese historical tradition associated with the saying bú fēi zé yǐ, yī fēi chōng tiān — if it does not fly, fine; once it flies, it pierces the sky — which he is supposed to have said about himself in his early years on the throne, when he was being underestimated. Zhuang, by reputation, was not a man easily fooled.

You Meng worked, for an extended period, for Zhuang's chief minister, Sun Shu'ao. Sun Shu'ao, like the king he served, is not a fictional character. He was the chancellor of Chu under Zhuang, credited in the historical tradition with substantial agricultural and irrigation reforms, and remembered in later Chinese ethics as a model of the incorruptible official. He died in office. The death is the event the comedy hangs on, because Sun Shu'ao, having served the king of Chu for many years with a reputation for fairness, with no taste for personal enrichment, and with an unshakable refusal to accept the kinds of gifts and properties that other officials of his rank made standard practice, left his family more or less destitute. His son, after his father's death, was reduced to selling firewood for a living.

You Meng, who had known Sun Shu'ao well, found out about this. He went home and started preparing.

The preparation took, in Sima Qian's account, more than a year. You Meng studied. He studied Sun Shu'ao's voice — the cadence, the choice of words, the particular tics of phrasing that any educated Chu listener at the king's court would have recognised within a sentence or two. He studied Sun Shu'ao's gait. He studied the way Sun Shu'ao held his head, the way he gestured with his sleeve when he was making a point, the way he sat when he was being addressed by the king. He studied for long enough that, when he was ready, his impersonation was indistinguishable from the original. The verb the Shiji uses, in the standard English of William Dolby and John Scott (1974), is direct: he learned to be Sun Shu'ao. Not to mimic him, not to imitate him, but to be him. The preparation has the shape of a method actor's apprenticeship. It also has the shape of a comedian's apprenticeship, because the acting and the comedy are one job in this case, and the difference between them is whether the audience figures it out before or after they laugh.

When You Meng was ready, he put on Sun Shu'ao's robes — by which Sima Qian means the actual robes; the widow gave them to him when he asked, knowing what he was doing — and his cap, and walked, on a feast day, into the king's banquet.

The room was full. King Zhuang was at the head of the table. The court was in attendance, including, presumably, the present chief minister and any number of senior advisors who had been at the same banquets when the original Sun Shu'ao was alive. The hall, by the standards of a Chu royal banquet of the Spring and Autumn period, was not a small room — bronzes set out, lacquered tables in their ranks, oil-lamps along the walls and the smoke from them rising. The court had been there for a while. Food had been served. The conversation had loosened. You Meng stepped through the threshold, and from the threshold to the seat the dead minister had occupied was, on the standard plan of a Chu palace audience hall, a walk of some thirty paces. He took them slowly. He walked the way the dead man walked — the slight forward set of the shoulders, the deliberate placement of the foot, the cadence Sun Shu'ao had used in the same room when the same king was watching — and the men nearest the door looked up first, and then the men in the second row of tables, and by the time he had reached the seat the room had begun to quiet. He took his place. He set his hands. He bowed. The bow was the dead man's bow.

What we are told about the next moment is that the king, momentarily, believed he was looking at his dead minister returned to him. Whether this is the king's full and considered belief or a startled half-second is a question Sima Qian does not pretend to answer, and reasonable readers can take it either way. What we are told he did next is rise to his feet and offer Sun Shu'ao the chief ministry of Chu, on the spot, in front of the room. If the king thought the man at the table was the dead Sun Shu'ao, the offer is the natural one: take back your post. If the king understood, by the second sentence, that he was looking at You Meng, the offer is the natural one as well: take the impersonation seriously enough to play it through. Either reading lands the same beat. The king made the offer.

You Meng demurred. The reading I want to flag here is that the demurral is the moment at which the game tips into its endgame: You Meng is now openly performing, and the king is now openly the audience, and the rest of the court has, depending on their seat and their reading speed, caught up. You Meng said he would need to consult his wife on whether to accept the post. He withdrew. He came back several days later — or, in some retellings, a moment later, depending on the rhetorical compression of the version — and announced that he had a song he would like to sing, on the subject of whether or not to accept the chief ministry of Chu.

The song, paraphrased here from the standard English of Sima Qian (Dolby and Scott 1974; reviewed in Otto 2001), runs more or less like this. I have considered, my wife and I, whether to accept the chief ministry. Sun Shu'ao served the king of Chu his whole life with absolute honesty. He took no bribes. He kept no concubines on state property. He died with debts. After his death his son, having no inheritance, was reduced to selling firewood from a cart in the marketplace. Knowing this, my wife and I have decided that to be a faithful chief minister of Chu is to make a man's son a firewood-seller. We respectfully decline.

That is the joke. That is the whole joke. It is also, by the conventions of any literate culture in any era, the cleanest example of the form one could ask for: an entertainer wearing a dead man's clothes, in the dead man's voice, in front of the king the dead man served, naming exactly the thing the king has not been doing about the dead man's family. The wrapper is comic — the impersonation is uncannily good; the demurral is the comedian's slow build; the song is the punchline — and what is inside the wrapper is, unmistakably, an accusation. The king is being told, in song, in the voice of his dead chief minister, that the king has failed in the smallest of his obligations to a man who served him faithfully.

The king, Sima Qian says, understood the joke. The verb in the original is laughter. The verb is also shame. The two are not separate beats in the Shiji's account; they are the same beat, registered as a single stroke. The king laughed and the king felt the rebuke and the king, the next item in the historical sequence, granted a fief to the impoverished son of Sun Shu'ao. Zhuang of Chu was not a man easily fooled. He was a man fooled, and called out, and corrected, and in correcting he used the standard mechanism of a Spring and Autumn court: a fief, with attached agricultural revenue, granted to repair an injustice the court had been quiet about. The mechanism worked. The son survived.

I want to camp here, because the structure of what just happened deserves careful naming.

The comedian wore the dead minister's clothes. The comedian impersonated the dead minister's voice and gait. The impersonation was good enough that the king — for a moment, or a half-second, or a quarter of a second, depending on how cynical one wants to be about the king's recognition — was seeing his dead minister returned to him, with the whole content of the impersonation now read in the impersonated man's voice. The comedian then stepped through that opening, in the impersonated voice, to land a moral point about the king's behaviour. The point was not delivered as commentary. Commentary at a Chinese court of the seventh century BCE was a high-stakes proposition, and there are biographies elsewhere in the Shiji, including in chapter 126 itself, of officials who got it wrong and died for it. The point was delivered as a song the impersonated dead minister sang at the banquet. The king laughed. The king felt the point. The king made the correction.

I want to call this, by the name it is going to keep through the rest of the book, performed satire. Not satirical text — performed satire. The defining property is that the comedy is a vehicle for a moral or political claim that, said straight, would be unsayable, and the comedy works because the wrapper is sufficiently entertaining that the claim arrives under cover. The audience laughs first and registers the rebuke after; or, in better cases, registers both at once. You Meng's banquet is the source case. The structural move — comedian-as-impersonator-of-power, landing a moral point under entertainment cover — has a long history that this book is going to keep walking into. By close family resemblance, it shows up in the work of a Mexican comedian called Cantinflas in 1940, on a courtroom stand. By closer resemblance again, it shows up in the political-impressionist video clips that circulate as memes in the algorithmic-age decade. Each of those is You Meng in different clothes.

For the record — because I want the proper noun in the reader's mouth before this chapter ends — Sima Qian wrote a Biographies of the Jesters in the 1st century BCE. The chapter is Chapter 126 of the Shiji. The Chinese name is Guji liezhuan. You Meng has his entry in it. So do several other named entertainers — the early Han fool Dongfang Shuo, the dwarf jester known in the Shiji by the title You Zhan, the actor Wu of Linling. I am not going to walk those entries here, because You Meng is the source case I need to plant in the reader's hands. But the genre is real. There was a Biographies of the Jesters, and it was about comedians who did jobs structurally identical to the jobs comedians do at modern political networks. The comedian put on the powerful man's clothes. The comedian impersonated the powerful man's voice. The comedian, in that voice, said the thing the powerful man would not let anyone else say. The audience laughed. The audience also felt the point. The point landed. In Sima Qian's case, the point landed hard enough that the dead man's son got a fief.

I will come back to this chapter, by name, later in the book. The proper noun Biographies of the Jesters is the hook. I am laying it now so that, when it returns, the reader will recognise it without having to be told.

The slave-seller

The fourth case is a 4th-century Greek joke book.

Ch1 named it: Philogelos, "The Laughter-Lover," 264 jokes, attributed to two compilers named Hierocles and Philagrius, of whom we know almost nothing. The standard view in classical scholarship is that they were probably Late Antique Greek-speaking grammarians or schoolmasters of some kind, but the textual tradition does not give us biographical detail. What it gives us is the joke book. It gives us the joke book in a manuscript tradition that runs through Byzantine codices, surfaces in the Western Renaissance as a cento of jokes attributed to Hierocles lifted whole into Latin and then European vernaculars, and gets its modern critical edition — the one the rest of this book is leaning on — from Roger Dawe in 2000, with English translations by Barry Baldwin (1983) and William Berg (2008). The Berg edition I keep on my desk, because it is open access and because Berg's English is salty in places where the Greek is also salty.

The 264 jokes break down, by stock character, into categories. Roughly a hundred and ten are jokes about the scholastikos, the standard Greek joke-book name for the absent-minded intellectual or pedant — an "egghead" or "scholar," cognate via the Latin scholasticus with the medieval academic — who is the joke book's central recurring fool. Twenty-odd are about a misanthrope and an envious man. About thirty are about regional stereotypes — the Abderites, the Sidonians, the Cumaeans — most of which have not aged into anything any contemporary reader can pick up the lay of without footnotes. A few are about a man with bad breath. A few are about his wife. A handful are about a man with a hernia. A scattered remainder are about a misogynist, a drunk, a glutton, a coward, and a man who keeps borrowing things and not returning them — the last of whom has, in the 4th century CE, exactly the same joke surface he has in the modern Anglophone office. The structural distribution is one a contemporary stand-up could reproduce by intuition. There are about a dozen recurring stock characters; the joke book runs maybe twenty jokes deep on the most popular ones; the punchlines compress into one or two lines apiece; and the book reads, to a reader in 2026, the way a medium-quality late-night writer's room would read if you flattened the room's collective output into a numbered scroll and left it on a shelf for sixteen centuries.

Ch1 mentioned the British comedian Jim Bowen retiring several of Philogelos's bits, untouched, on a 2008 Edinburgh Fringe stage. I am not going to retell that scene, but I want to take from it the fact that the bits, when delivered by a working comedian to a working audience in modern English in modern Scotland, got laughs in the same beat the original audience appears to have laughed. Bowen's act in 2008 was, in this respect, a small live-fire experiment in the universality of the comedic structure of the Philogelos: drop sixteen centuries of cultural distance, give a comedian the lines, see what happens. What happened was that they laughed. They laughed before they were told the bits were old. Some of them, when they were told afterwards, laughed harder.

I want to walk one bit. The bit is Joke #18 in the standard numeration of the manuscript order. Berg's English (which I am quoting verbatim, attributing to Berg's translation, not to the original Greek):

A man complains to a scholastikos that the slave he sold him has died. The scholastikos answers: "By the gods! When he was with me, he never did anything like that."

I want to spend the rest of this chapter on that punchline, because that punchline is the load-bearing demonstration of the chapter's case.

I want to start by registering, plainly, what the joke is about. It is about a slave. The slave is a human being who has been sold, by one person, to another, as property. The slave has died, after the sale. The buyer has come back to the seller with the complaint. The setting is a fourth-century Mediterranean economy in which human chattel was a commodity, and the readership is a fourth-century Greek-speaking literate audience that bought slaves, sold slaves, had slaves in the household when they read the joke, and would have read the joke, in the typical case, in the company of slaves who could not refuse to be present. There is no version of this joke that is not also a joke about somebody's death, which is to say a joke about the fact that the seller's property has, by the operation of being a person, become not-property by dying, and the seller's response to the news is to refuse to register the change of category.

I want to camp on the moral content for a moment, because the chapter does not honour the joke if the chapter does not honour the moral content.

The slave-seller in the joke is a contemptible person. The joke trades on his contemptibility — the joke's whole comedic structure is built on the seller's failure to register that the slave was a person who could die. The seller is treating death as a behaviour the slave personally chose to engage in: never did anything like that is the locution of a master complaining about a piece of misbehaviour, an act of disobedience, a category of conduct. He never did anything like that. He didn't behave that way around me. Why did he start now? The answer the joke does not let the seller hear is: because he was a person, who had a body, and the body broke down. The seller has, on the available evidence, never permitted himself to think about his slave that way and is not about to start. The buyer, presumably, says nothing more in the original setup; the punchline does its work and the buyer is left holding it. The buyer is now in the position of the joke's audience: hearing the seller's response, registering that the seller is being absurd, registering that the absurdity is on a moral subject the joke has not made absurd by joking about.

I want to be clear that the Philogelos is not on the seller's side. The scholastikos is the joke book's recurring fool. The joke book runs at his expense. The reader of the joke book is being invited to laugh at the seller, not with him, and the joke would not be a joke if the seller's response had been read as reasonable. The fourth-century Greek-speaking reader knew the seller was being a fool. The joke was funny because the seller was being a fool.

It is also funny because the seller's foolishness has a particular shape, and the shape is the shape of the deadpan.

The deadpan, as a comedic move, is what happens when a speaker refuses to register the absurdity of what is plainly the case. The standard analytical name for this in 20th-century criticism is deadpan understatement or deadpan refusal. In its purest form the deadpan is verbal: the speaker, on being confronted with a fact that ought to elicit some sort of recognition, response, adjustment, or alarm, declines to do any of those things, and proceeds in the register he was using before the fact arrived. The slave-seller in Philogelos #18 is a perfect deadpan. The fact in the room is that the man he sold has died. The response that would register that fact would be: I am sorry; let me refund you; what happened? The response that would register half of that fact would be: That happens. People die. The response that would register none of it would be silence. The response the scholastikos gives is in a fifth category: he registers the fact as a piece of behaviour he is now obliged to defend his own honour against. By the gods! When he was with me, he never did anything like that. The slave's death is being framed as a slight to the seller's own competence as a slave-haver. The seller is indignantly refusing to acknowledge that the slave was the kind of thing that could have died at all.

The deadpan refusal-to-register is, structurally, the comedic move that survives 1,600 years. I have been laying the structural claim across the whole chapter; now I am pointing at the specimen. The slave-seller in Philogelos #18 is the archetype. He refuses to register the change of category. The refusal is the joke. The refusal is funny because the refusal is absurd, and the refusal is absurd because reality has changed under the seller, and the seller is determined to keep walking through the room as if it had not.

This is the move the rest of the book is going to keep finding, in clothes that look completely different from a Late Antique Greek slave-market. It returns, briefly, in an 18th-century English jest-book lifting a Joe Miller in which a witness undoes a speaker. It returns, conclusively, in a Mark Twain frog with a bellyful of quail-shot in 1865 Calaveras County, where Smiley's last line in the story — Why blame my cats if he don't weigh five pound! — is the slave-seller's by the gods in American English. It returns in a 1969 British television sketch about a customer trying to return a parrot to a shop that does not concede the parrot is dead. It returns, last, in a two-panel American webcomic from 2013 in which a dog in a hat sits at a kitchen table in a kitchen that is on fire and says, mildly, this is fine. The webcomic is wordless except for that line. The dog refuses to register the fire. The fire is the joke. The shopkeeper refuses to register the parrot's death. The dead parrot is the joke. Smiley refuses to register the substitution of the frog. The frog full of quail-shot is the joke. The slave-seller refuses to register the death of the slave. The dead slave is the joke.

Each of the four refuses to register a different thing. Each of the four refuses to register a thing whose registering would force a real moral or material adjustment. Each of the four does the refusing in a register so calmly continuous with the speaker's prior register that the listener — reader, audience, viewer — is left holding the gap. The speaker has not changed. The world has changed. The world's change is what the audience is supposed to register, and the speaker's failure to register it is the engine of the comedy. The deadpan is, in some structurally precise sense, a refusal contagion: the speaker refuses, and the audience laughs because the audience is being asked to register what the speaker won't.

I want to come back to the slave-seller, because the chapter is meant to land on his voice.

The slave-seller, as a moral being, is small. He is small because the slave was a person and the seller is treating the slave as something other than a person, and the seller's smallness consists exactly in his unwillingness to revisit that treatment in the face of the slave's death. The seller's smallness is not specific to the fourth century. It is, alas, a recurring kind of human smallness — the kind that, on encountering an inconvenient fact, declines to revise its categories. The book is going to come back to that kind of smallness more than once, and not only in jokes. There are versions of the slave-seller's voice that are not funny. There are versions of the slave-seller's voice that are running countries. The chapter is going to honour that distinction by not flinching from it now.

The joke is not on the slave's side. The joke is also not on the seller's side. The joke is on the side of the reader, who is invited to register, in a single beat, that the seller is the kind of person who will not register what is plainly the case. The reader's laugh is the recognition that the world has just continued without the seller catching up to it. The recognition is the same recognition the audience of Dead Parrot gets when the customer says this parrot is dead. The recognition is the same recognition the reader of "The Celebrated Jumping Frog of Calaveras County" gets when Smiley pries open the frog's mouth and the quail-shot pours out. The recognition is the same recognition the viewer of the burning-kitchen webcomic gets when the dog says this is fine. The world has changed. The seller has not. The buyer has, by being told the seller's response, been deputised as the audience. The audience laughs.

I want to give the chapter its last sustained note now, because the slave-seller's voice is the note.

The slave-seller is a man who, on being told a categorical fact about the slave he just sold, declines on principle to believe it. He declines without apology. He declines with the indignation that comes from being asked to revise a worldview he has not got the time, the energy, or the moral apparatus to revise. He declines, in Greek, in a register so flat and so confident and so untouched by the fact in the room that, twelve centuries later, a Byzantine scribe will copy his voice into a manuscript, and four centuries after that an early-modern compiler will copy it again into a Latin cento, and three centuries after that a British library will hold one of those copies on a shelf with a call number, and a few decades after that Roger Dawe will produce his critical edition in Munich, and a quarter-century after Dawe a British comedian will pull the bit, intact, and walk it onto the Edinburgh stage and get a laugh on the first beat from people who did not know they were hearing 1,600 years of refusal-to-register.

The slave-seller's voice is the chapter's last sound because the slave-seller's voice is the source the rest of the book is going to keep finding. He is in 1865 Calaveras County. He is in 1739 London. He is in 1969 BBC. He is in 2013 in a wordless webcomic. He is, as a structural matter, a constant. His clothes change. His refusal does not.

By the gods, he says, when he was with me he never did anything like that.

The chapter ends with him saying it. I want him to say it, and to mean it, and to be wrong, and to be funny, and to be 1,600 years old, all at once. He is. He was. Nothing about him has changed since the fourth century. Several hundred million people, between Hierocles and Philagrius and you, have laughed at him. Several hundred million more, on the chapter's available trend lines, will. The joke survived because it was good. The seller survived because the joke survived. The chain of survival, here as everywhere, is a chain of grins.

By the gods. When he was with me. He never did anything like that.