Six of six · Comparative mythology
Old religious texts were copied by hand for centuries, and each generation of copyists
adjusted what it disagreed with. What they did not think to adjust stayed exactly where
it was.
Which means the awkward leftovers — the lines that no longer fit the beliefs of the people
preserving them — are very often the oldest and most reliable part of the whole thing.
When a story is written down, copied, added to and copied again over centuries, the result is a single object that reads as though one person made it. It was not. Under the surface there are earlier versions, later insertions, and material that came in from somewhere else entirely.
Usually that has to be argued for. Occasionally the seams are visible to anybody who looks, which makes it a good place to learn what looking involves.
The Epic of Gilgamesh is the best-preserved long poem from the ancient Near East, and it survives in several versions spread across roughly a thousand years.
The three buttons take you down through the text. The first shows it as it comes to us, and each one after that adds what is known about how it got that way.
Tablet I. Gilgamesh, king of Uruk, oppresses his people. The gods create Enkidu to match him.
Tablets II to VI. The two become friends, kill the forest guardian Humbaba, and refuse the goddess Ishtar.
Tablets VII and VIII. Enkidu sickens and dies. Gilgamesh will not leave the body.
Tablets IX to X. Gilgamesh goes looking for the one man who never died.
Tablet XI. Utnapishtim describes the flood: the boat, the animals, the birds sent out to find land, the sacrifice afterwards.
Tablet XII. Enkidu goes down to the underworld and describes what he finds there.
Nothing there is speculation. There are physical tablets carrying the Sumerian poems, physical tablets carrying the flood story about Atrahasis, and a final tablet in which a character who has been dead for four tablets is walking around.
The Gilgamesh case works because different versions were written on different tablets and buried in different places. The same thing happens with texts that stayed in continuous use, and there the evidence takes a different form: manuscripts that disagree.
Deuteronomy 32 contains a short passage about how the world's peoples were divided up. It exists in two forms, and the difference between them is a single phrase. What follows is a description of the passage rather than any translation of it, and the point at issue is a matter of manuscript evidence rather than of interpretation.
When the Most High divided up the nations and separated humankind, he fixed the boundaries of the peoples.
He fixed them according to the number of the sons of Israel.
He fixed them according to the number of the sons of God.
For the LORD's own portion is his people, and Jacob is his allotted share.
Reasonable people read this differently, and the scholarship on it is extensive and not settled. What is not in dispute is the manuscript situation: two ancient witnesses carry one phrase, later manuscripts carry another, and the older reading is the more difficult one, which is generally taken as a sign of age rather than of error.
Anybody who has been described by an opponent knows the answer is not nothing. A hostile account gets the emphasis wrong and the motives wrong, and it is often extremely well informed about the details, because it had to be in order to land.
The same thing happens on a scale of centuries. You cannot tell people to stop doing something without saying what it is, and you cannot refute a belief without first stating it. So the surviving record of a practice is very often produced by the people who wanted it stopped, and is frequently the only record there is.
Three of them below. Each opens on what was actually found, and the second step is what follows from it.
Kuntillet Ajrud is a waystation in the Sinai desert. Travellers passing through in about the eighth century BC wrote blessings on the plaster and on the large storage jars, in the ordinary way that people wrote blessings.
Several of them invoke Yahweh, identified by locality as gods were then, and add three further words: and his asherah.
Separately, the Hebrew Bible instructs its readers, repeatedly and at length, to cut down and burn asherahs.
A French excavation on the Syrian coast in 1928 turned up the archives of a city called Ugarit, written some centuries before most of the Hebrew Bible and in a closely related language.
The tablets describe a storm god, Baal, and give him a set of standing epithets that recur wherever he appears. One of them calls him the one who rides on the clouds. The same phrase turns up in the Psalms, applied to Yahweh.
Manichaeism was for several centuries a genuinely intercontinental religion, running from the western Mediterranean to China, and its own scriptures were destroyed with unusual thoroughness.
For most of history, what survived came through the people suppressing it. Christian bishops quoted passages in order to refute them. Muslim heresiographers catalogued the beliefs in order to classify the error precisely.
The previous section dealt with traditions arguing with themselves. This one is harder. There are whole religions for which no follower ever wrote a surviving account, and everything known about them comes from somebody who arrived from outside.
That is not the same as the record being worthless. It does mean the record has a shape, and the shape belongs to the writer rather than to the subject.
A large share of everything known about Gaulish religion comes from Julius Caesar's account of his own campaigns, written for readers in Rome whose support he needed.
He describes the druids at some length: how long the training took, the authority they held in disputes, the rites he presents as barbaric. He then lists the gods. The Gauls, he says, worship Mercury above all, and after him Apollo, Mars, Jupiter and Minerva.
Bernardino de Sahagún was a Spanish friar who spent decades working with Nahua elders and with trained Nahua assistants to compile an enormous encyclopaedia of Aztec life, in Nahuatl and Spanish, covering the calendar, the rhetoric, the trades and the religion in great detail.
He was entirely open about why. A physician has to know the disease before he can treat it, and missionaries could not root out practices they were unable to recognise.
Most of what is known about Norse mythology comes from Snorri Sturluson, writing roughly two centuries after Iceland converted to Christianity. He was a Christian, a lawyer and a formidable political operator, and his handbook exists so that poets could go on using the traditional imagery without losing the key to it.
His prologue explains that the gods were never gods. They were men, refugees out of Troy, whose descendants travelled north and were taken for divine by the people they found there.
Three writers, three purposes, and the same structural fact underneath. Each was working against the thing he was recording, and each is now indispensable, which is why the discipline spends so much of its time working out what a source was for.
Everything so far has worked downwards through one tradition at a time. There is a second method that works sideways, and it is the reason this subject is a subject rather than a collection of resemblances.
Similar stories in different places prove nothing on their own. People everywhere have floods, tricksters and dead relatives, and inferring contact from that is how the field embarrassed itself for most of the nineteenth century.
What does prove something is a name that behaves like ordinary vocabulary. Take a word nobody argues about. Latin pater turns up as French père, Italian padre and Spanish padre. Those are not four similar words. They are one word, with each language's own regular changes applied to it, and those are the same changes that language applies to every other word it inherited.
So if a divine name shows up across a family of languages, having been put through exactly those same changes, it did not travel on its own. It rode the language, along with the words for father and water and two.
One last thing, and it is the reason the subject is worth a life rather than an afternoon.
Everything above has involved standing in front of a tradition and working out what is underneath it. That is a modern activity with journals and departments attached. It is also, in a rougher form, something the traditions were already doing.
People in every one of these places found monuments they had not built, in a style nobody around them could manage, left by somebody. And they did what anybody does with unreadable evidence. They explained it.
Irish tradition describes a sequence of peoples arriving on the island, each displacing the last. The Tuatha Dé Danann are defeated by the final wave and withdraw, not overseas but downwards, into the mounds. They are still there, they are called the folk of the hills, and it is well understood that you do not annoy them.
An Old English poem stands in front of a ruined city of worked stone, with towers and roofs and baths that ran hot, and describes it as the work of giants, fallen now, its builders long in the earth.
The menehune are a small people who were on the islands before the Hawaiians. They work only after dark and only in one night, so a fishpond or a watercourse that took them longer than that was abandoned unfinished, and several are pointed out as such.
Aymara and Quechua traditions describe the chullpas: an earlier people who lived before the present sun rose, in dimmer light, and who were destroyed when it did. Their name is also the name given to the stone tower tombs that stand across the high plateau.
Four traditions on four continents, none of them in contact with the others, each arriving at a prior people who were smaller or larger or worked at night, and who are gone but not entirely gone. The explanations are wrong in the way that unaided explanations are wrong, and the observation underneath them is completely correct. Somebody was here before, and left something nobody could account for.
Which is the whole of the subject in miniature. A myth is what a people says about a past it cannot reach, assembled out of whatever happens to be lying around, and it preserves far more than it means to.
Everyone who has ever tried to read the ground has been doing this. These are just the earliest drafts.