The Book of the Dead Is a User Manual

The Book of the Dead is not scripture. It is documentation: a modular, customised, commercially produced set of procedures for getting a specific dead person through a specific sequence of hazards, and the people who wrote it, sold it, and were buried with it treated it exactly that way. We call it a "book" because a German philologist needed a title in 1842. The Egyptians called it something plainer and more accurate: ru nu peret em heru, the utterances for going forth by day.¹ Utterances. Plural. Things you say, in order, at the right moment, to make a mechanism work. I have spent a lot of my working life reading and writing runbooks, and the more of these papyri I read in translation the less they look like a holy text and the more they look like the folder that ships with a piece of equipment.

That is not a put-down. A good runbook is one of the most humane documents a civilisation produces. It assumes you will be alone, frightened, and under time pressure when you need it, and it is written to work anyway. Read the Book of the Dead with that assumption in mind and every strange feature of it, the red-ink instructions, the blank slots for names, the spells that exist only to repair the failure of other spells, the price tiers, the successive editions, stops looking mystical and starts looking like engineering.

Rubrics: the instructions are in red

Open a decent facsimile of a New Kingdom papyrus and the first thing that separates it from anything we would call scripture is the colour coding. Most of the text is black. Certain lines are red. The red lines are not the sacred content; they are the metadata. They give the spell its title — one is a spell against dying a second time, another a spell for turning into a swallow, to name two in my own words — and, more tellingly, they tell you what to do with it: words to be spoken by so-and-so; to be recited over a figure of such-and-such made of such-and-such material; do this and the outcome is such-and-such.² Egyptologists call these rubrics, from the Latin for red, and the borrowing is exact, because a rubric in a mediaeval service book did the same job: it told the priest what to do while the black text told him what to say.

Look at what a rubric contains and you have the fields of a procedure record. A name for the step. The actor: the words are to be spoken by the Osiris so-and-so. The trigger or context. The materials, sometimes to a startling level of specificity: green stone, gold mounting, a certain oil, a figure of clay, a strip of linen. The expected result. And, on some spells, a provenance note that reads like a changelog entry: this spell was found in the time of King Menkaure, under the feet of a statue of the god, by the king's son Hordjedef, and described — I paraphrase Faulkner and Allen here rather than quote them — as a genuine secret that no eye had seen and no one had looked upon.³ Whether or not any of that is true, and I do not think anyone believes it is literally true, it tells you what the copyists valued: an old, authenticated version, with a chain of custody. That is not how people talk about revelation. It is how people talk about a trusted source file.

Even the spells that look most like theology are shaped like reference material. Spell 17, one of the longest and oldest in the corpus, is not a prayer but a text with running commentary: a line of the old utterance, then a question asking what it means, then one or more glosses, sometimes several competing ones stacked up and introduced by a tag that amounts to: or else it means this.² That is exactly the texture of an annotated standard, where the original wording is preserved because it is authoritative and the annotations accumulate because nobody dares delete the ambiguity. Anyone who has maintained a long-lived configuration file will recognise the genre.

Named slots and the blank for the customer

The single most manual-like feature of the Book of the Dead is the one most often skated over in museum labels: the owner is a variable. Throughout the text the beneficiary is the Osiris so-and-so, justified, and the so-and-so is whoever bought the roll. In some surviving papyri the name was written by the same hand as the rest; in others the scribe left gaps and someone else filled the name in later, in a visibly different hand and ink, and occasionally forgot a gap altogether so that the slot is still blank three thousand years on.⁴ These were, at least sometimes, stock rolls: written ahead of demand, personalised at point of sale.

Once you see the name as a field, the rest of the personalisation follows the same logic. Vignettes could be adapted to show the owner's face, dress and profession, or not, depending on what you paid. Spells could be added or dropped. Titles could be padded. But the underlying text was templated, and it was templated because it was produced in quantity by workshops, not composed fresh for each burial by a priest in ecstasy. The scribe was doing what a technical writer does: instantiating a standard document for a particular customer.

The owner is a variable. These were, at least sometimes, stock rolls: written ahead of demand, personalised at point of sale.

Editions: from Thebes to Sais

The Book of the Dead never existed as one book. It is a corpus of somewhere near two hundred spells, from which any given papyrus drew a selection, in an order that in the New Kingdom varied from copy to copy. Egyptologists distinguish this looser New Kingdom material, the "Theban recension" in Édouard Naville's 1886 phrase, from the later "Saite recension" that took shape around the Twenty-sixth Dynasty and remained standard through the Ptolemaic period.⁵ The Saite version fixed the sequence: the same spells in the same order, so predictably that when Richard Lepsius published the Ptolemaic papyrus of Iufankh in Turin in 1842 and numbered its spells 1 to 165, that numbering could be projected backwards onto every earlier manuscript and still, with additions up to about 192, serves as the field's citation system today.¹

What is that, if not version control? A long period of divergent local forks; then a consolidation, in which someone, or some priestly college, decided which spells were canonical and in what order; then a stable release that later copyists reproduced with only minor drift. The Saite standardisation is sometimes described as an archaising, conservative move, and it was, but conservatism in documentation is not stagnation. It is what you do when a system has to be reproduced accurately by many hands across many centuries and you cannot afford each scribe reinventing the sequence.⁶

The editions also track the medium and the script. Early New Kingdom examples turn up on linen shrouds and coffins before the papyrus roll becomes the norm. The New Kingdom rolls are mostly in cursive hieroglyphs; from the Third Intermediate Period on, hieratic becomes common; by the Ptolemaic period you find versions in hieratic and, occasionally, demotic, alongside abbreviated successor texts such as the Book of Breathings.⁴ Same procedures, ported to whatever writing system the market and the scribal schools were currently using. Nobody thought the words lost their power in a new hand. That is a documentation attitude, not a scriptural one; readers of the Rosetta decree, published in three scripts precisely so it would work for three readerships, will recognise it.

Quality tiers and price points

The finest surviving copies, the papyrus of Ani in the British Museum, the roll of the architect Kha found intact in his tomb at Deir el-Medina, the papyrus of Hunefer, are luxury goods: many metres of good papyrus, full-colour vignettes, meticulous script.⁷ But they are the top of a range. Below them are rolls with vignettes in outline only, or with vignettes but botched text, or with text only. Below those are short excerpts on a strip of linen, a spell or two on a coffin, a single spell 30B engraved on a heart scarab because that was all the family could afford. And a great many burials, of course, had none of it.

The village of the royal tomb-workmen at Deir el-Medina, whose ostraca and papyri give us the only substantial price data from pharaonic Egypt, shows draughtsmen and scribes producing funerary papyri for their neighbours and, presumably, for the market. Jac. Janssen's compilation of Ramesside commodity prices includes funerary goods, coffins, and at least one attested price for what the scribes called a roll for coming forth by day, and although I would not lean hard on any single figure, the order of magnitude is clear: a proper roll was a serious purchase, months of a workman's wages, comparable with a good coffin rather than with a loaf.⁸ There was a market, the market had tiers, and the tiers were reflected in the product's completeness. Vendors do not do that with revelation. They do it with a manual, where the deluxe binder has every chapter and the pocket edition has the pages you cannot do without.

Error-tolerant design

Here is the feature that convinced me. Good procedural documentation assumes failure. It does not just say "do X"; it says "if X does not work, do Y." The Book of the Dead is full of Y.

Take the judgment. Spell 125 is the famous scene: the deceased enters the Hall of the Two Truths, greets Osiris and forty-two assessor gods by name, and makes a long series of declarations: in my own words rather than Faulkner's, that he has done no wrong, robbed nobody, killed nobody, and never tampered with a balance. Older books call it the Negative Confession; the better term is a declaration of innocence, because nothing is confessed.² This is the deceased asserting conformity to ma'at, the order against which everything in Egypt was measured, in the format of a checklist. The list is not a moral examination; it is a compliance attestation, drafted in advance and recited from the roll.

But the Egyptians did not trust the attestation alone, because attestations can be contradicted, and the thing that could contradict you was your own heart, weighed on the balance against the feather of ma'at with the Devourer waiting. So the same document carries spell 30B, whose whole purpose is to stop the heart from testifying. In my paraphrase of Faulkner's text, the dead man addresses the heart he had from his mother and instructs it not to rise as a witness against him, not to take the other side in the tribunal.³ And the rubric on 30B tells you how to make it stick: recite it over a scarab of green stone, mount it in gold, place it at the throat of the dead. Thousands of heart scarabs survive with exactly this text carved on the base. That is not piety; that is a redundant safety on the most dangerous step in the process. Belt, and braces, and a spell over the braces.

The pattern repeats. Spell 6, the shabti spell, provides for the case where the dead person is called on to do agricultural corvée in the afterlife: a figurine will answer for you, and the text to be inscribed on the figurine is given.² There are spells for not having your head cut off, for not being eaten by snakes, for not eating excrement, for having air, for having water, for not being turned upside down, for knowing the names of the gatekeepers of each portal so that the gate will open. Every one of them is a contingency handler. Together they form a coverage matrix over the known failure modes of being dead. It is not comprehensive, no runbook is, but the intent is unmistakably to enumerate the things that go wrong and pre-supply the fix for each.

Spell 30B is not piety; it is a redundant safety on the most dangerous step in the process.

Even the closing formulae are written for reliability rather than reverence. Several rubrics end with efficacy claims which, paraphrased, assert that the thing has been proved true a million times over, or that the reciter will not be turned back at any gate, or promise that the spell works if recited over the correct object and that the reciter and the family will benefit.² A theologian reads those as magical thinking. An engineer reads them as the acceptance criteria.

Where it came from: three generations of one document

The lineage of the Book of the Dead is itself an argument for the manual reading, because it is a lineage of copies, adaptations and re-releases rather than of revelation. The earliest funerary corpus, the Pyramid Texts, appears carved on the walls of the pyramid of Unas at the end of the Fifth Dynasty, around 2350 BC, and in the royal pyramids of the Sixth Dynasty after him.⁹ Their concern is the king's ascent to the sky. In the First Intermediate Period and Middle Kingdom, texts drawing on and expanding that material appear painted inside the wooden coffins of officials and their families, the Coffin Texts, with a much stronger Osirian flavour and a new interest in the perils of the underworld route.⁹ Then, in the early New Kingdom, the utterances for going forth by day appear on shrouds and coffins and shortly afterwards on papyrus, containing spells that can be traced directly back through the Coffin Texts, some of them to the Pyramid Texts, alongside a large amount of new material.

The textbook narrative for this sequence, which you will still meet in survey courses and museum labels, is called the "democratisation of the afterlife": the royal prerogative of the Pyramid Texts trickled down to nobles in the Coffin Texts and then to anyone who could pay in the Book of the Dead, so that by the New Kingdom every Egyptian could hope to become an Osiris. The phrase goes back to James Henry Breasted's lectures of 1912 and it has an obvious appeal, since it makes Egyptian religious history look like a slow march toward spiritual equality.¹⁰

I do not think it survives contact with the evidence, and I am not alone. Harold Hays argued in 2011, in an essay bluntly titled "The Death of the Democratisation of the Afterlife," that the story confuses the appearance of written texts with the appearance of the beliefs behind them; Old Kingdom non-royals already expected an afterlife, already used offering formulae, and were already called "Osiris" in some contexts, and what changes over time is which social groups had access to which forms of display, not who was theologically eligible for eternity.¹¹ Mark Smith reached a compatible conclusion from a different direction, showing that the language of becoming Osiris is not restricted to kings before the Middle Kingdom, and that the notion of a top-down "democratisation" is largely an artefact of the accidents of survival and of what was, at each period, customary to write down.¹² The older view is not dead; you can still find scholars who think the term captures something real about who could commission what. But it is contested, and the direction of the argument favours the sceptics.

Read as documentation, the sequence makes better sense without the theology of trickle-down. What we see across Pyramid Texts, Coffin Texts and Book of the Dead is the same procedural corpus being repackaged for successive media and markets: from a one-off installation carved into a royal monument, to a bespoke painted job inside an elite coffin, to a portable, templated, purchasable roll. The customer base widens because the format gets cheaper and the production gets more industrial, not because the gods changed their admissions policy. That is how manuals spread. It is also, incidentally, why the corpus survives the state that made it: when the Amarna experiment pushed the Osirian afterlife out of official favour for a generation, the workshops seem simply to have paused, and when Osiris came back so did the rolls.

What is lost, and what is gained, by reading it this way

The obvious objection is that I am flattening something numinous into something mundane, and that the people who were buried with these rolls certainly believed in the gods, the balance and the Devourer. They did. But believing that a system is real is not the same as writing scripture about it. Scripture, in the sense the word usually carries, is a text you read to know what is true or how to live. The Book of the Dead is a text you use, once, at a moment of maximum need, in order to make a mechanism behave. Nobody meditated on it; it was rolled up and put between the legs of the mummy or inside a hollow statue of Osiris-Sokar and left there to do its job. Its authority came from its efficacy claims, its provenance rubrics and its faithful copying, not from any claim to reveal the mind of a god. Egypt had hymns and wisdom texts and temple liturgies for that. This was a different genre, and the Egyptians knew it was different, and they named it accordingly.

What is gained is that features which look like superstition when you expect scripture look like competence when you expect a manual. The redundancy of 125 and 30B. The red-ink field structure. The name slots. The successive recensions converging on a fixed order. The tiered products for tiered budgets. The commentary layer in spell 17 preserving three incompatible readings of one line because you do not delete ambiguity from an authoritative source, you annotate it. Every one of these is a decision that someone in the documentation business would make today. The Egyptians made it first, on papyrus, for a user who could not ask questions and would only ever run the procedure once. That is the hardest kind of manual to write. On the evidence, they were very good at it.