The Tower of Babel and the Mesopotamian Texts

Summary

No Mesopotamian text tells the story of Genesis 11.  The building details of Genesis 11:3-4 are right for southern Mesopotamia, down to the idiom, and that is a real point for the text's knowledge of Shinar.  But the one Babylonian tablet once offered as a tower story is about something else.  The Sumerian spell offered as a memory of one language has no tower, no city and no scattering.  The texts closest to Genesis in thought are royal inscriptions of the eighth and sixth centuries BC, and the world legends that look most like Genesis were written down after missionaries arrived.  No text from anywhere corroborates the event independently.  This article is a child of The Tower of Babel, which sets out what the text says and how it has been read;  its siblings ask where the tower was and when it was built.

The Spell of Nudimmud is genuinely disputed.  In the Sumerian epic Enmerkar and the Lord of Aratta, a spell speaks of all lands addressing Enlil in a single language and of Enki changing the speech in their mouths.  Samuel Noah Kramer read it in 1968 as a memory of the past, "beyond all doubt".¹  The Oxford ETCSL translation and Herman Vanstiphout read it as a wish for the future, and Vanstiphout concedes the other reading:  "Grammar and philology could bear such a reading."²  Checked against the standard rule for its verb, the line's form states a fact rather than a wish, which leans to Kramer.  On either reading the spell has no tower, no city, no building and no scattering.  And Kramer made Genesis the later, reworked story, "a product of the Hebrew religious imagination and moralistic temperament."¹

The building scene has Mesopotamian antecedents.  In Enuma Elish VI the gods mould bricks for a year and raise the peak of Esagil, and Claus Westermann notes that the Hebrew "let us brick bricks" has an exact Akkadian equivalent in the epic.³⁴  E. A. Speiser concluded in 1956 that "each item in the biblical narrative proves to have its appropriate Mesopotamian antecedent".⁵  A. R. George showed in 2005 that the epic, and the ziggurat at Babylon it celebrates, existed in the second millennium BC.⁶  "Its top in the heavens" is a stock boast of Mesopotamian builders, already used by the Assyrian king Tiglath-pileser I.

No Babylonian tower story exists.  K 3657, published in 1876 as a possible one, is about the people of Babylon "burdened with hard labour".⁴  Herodotus describes the tower of Babylon and tells no story of its building.

The closest parallels are late and imperial.  Sargon II in the eighth century BC settled deportees of many tongues in his new capital and, in the literal Akkadian, made them one mouth.⁷  Nebuchadnezzar II in the sixth century finished Etemenanki "so that it vied with the heavens", with "all the teeming people of the world" conscripted for the work.⁸⁸  These texts show where a first-millennium audience would have heard the story as satire on empire.  They do not date its writing, because the boast of a top in heaven is older than both kings.

The world parallels are borrowed or partial.  James Frazer sorted the legends in 1918 into three groups.  The stories with both a tower and a confusion of tongues were recorded after contact with missionaries, and the Cholula version has bitumen, which "is not found anywhere near Cholula".⁹  The stories from before contact lack either the tower or the confusion.

Verdict.  Genesis 11 knows Shinar, and no text corroborates its story.  The young-earth literature has no published answer to the point that the closest parallels are first-millennium.

Part 1: The Mesopotamian texts

No Mesopotamian text tells the story of Genesis 11.  Three kinds of text come near it:  a Sumerian spell about one language, a Babylonian myth in which gods mould bricks to build Babylon, and a long run of royal inscriptions that boast of towers whose tops reach heaven.  The building details of Genesis 11:3-4 match them closely.  The texts that come closest in thought, a city where many peoples are made one and a tower finished with the labour of the whole world, are Assyrian and Babylonian royal texts of the eighth and sixth centuries BC.

The Spell of Nudimmud

In the Sumerian epic Enmerkar and the Lord of Aratta, the king of Uruk sends a messenger to the lord of Aratta with a spell to recite.  The spell is lines 134-155.  The Oxford ETCSL translation reads it as a wish for the future:  "At such a time, may the lands of Cubur and Hamazi, the many-tongued, and Sumer ... and Akkad ... and the Martu land ... may they all address Enlil together in a single language!"  Then Enki "shall change the speech in their mouths, as many as he had placed there, and so the speech of mankind is truly one."¹⁰

Samuel Noah Kramer read the same lines as a memory of the past.  In 1943 he had the passage without the verb of line 146 and guessed it.  In 1968 a school exercise tablet in the Ashmolean Museum supplied the verb:  "it turns out to be 'spoke' rather than 'gave praise' as I had surmised".¹  His conclusion:  "Our new piece, therefore, puts it beyond all doubt that the Sumerians believed that there was a time when all mankind spoke one and the same language, and that it was Enki, the Sumerian god of wisdom, who confounded their speech."¹

Bendt Alster argued the reverse in 1973, and Wenham reports his reading:  the text "is really looking forward to a time when all mankind would speak the same language, the Sumerian language."¹¹  Herman Vanstiphout, who edited the epic for the Society of Biblical Literature in 2003, states where the field now stands:  "Many of us now agree that the burden of this spell is that in an ideal world of peace and bliss, which Enmerkar is striving for, only Sumerian will be spoken."²  He reads "the whole world of well-ruled people" as "the imagined or ideological state of Ur III Sumer".²  Then he concedes the point that keeps the dispute open:  "An alternative reading, which puts it in the past, not in the (immediate) future, is not impossible.  Grammar and philology could bear such a reading.  Yet the style of the episode and certainly its meaning and function in the story argue strongly for my interpretation.  Putting it in the past cannot be of any use in the development of the narrative."²

So the dispute is real.  The grammar leans to Kramer without settling it.  Kramer was sure of the past:  the only thing he left "uncertain for the present" was the nuance of one infix, -da-, in the verb of line 146, ḫe₂-en-na-da-ab-dug₄.¹  Its prefix ḫe₂- can mark either a wish or a statement of certainty, and the stem of the verb tells which.  Gábor Zólyomi, one of the ETCSL editors, gives the rule in his grammar of 2017.  For certainty "the verb uses as a rule the preterite tense", and for a wish "Transitive action verbs prefixed with a deontic /ḫa/- use the present-future tense."¹²  For "to speak" the preterite stem is dug₄ and the present-future stem is e.¹²  Line 146 has dug₄.  Zólyomi glosses the same prefix on the same verb in a letter of the 21st century BC as "I did tell him".¹²  On that rule the line states that the lands did speak to Enlil in one tongue, which is Kramer's reading, and both future renderings, ETCSL's "may they" and Vanstiphout's "will be able to", need the rule to bend.  Three cautions apply.  The rule holds "as a rule", not always.  The application to this line is this article's own, not a published reading.  And ETCSL, which Zólyomi helped edit, still prints the wish.  Vanstiphout grants that the grammar could bear the past.  The future reading rests on what the spell is for in the story:  a messenger recites it to make a rival king submit.

What the spell lacks matters more than its tense.  On either reading it has no tower, no city, no building and no scattering.  Its gods are Enlil and Enki.  Kramer guessed that the cause of the change was "Enki's jealousy of Enlil".¹  And Kramer did not treat the spell as an independent witness that confirms Genesis.  He treated Genesis as the later, reworked version:  the biblical explanation "is quite different from our Sumerian version, and is undoubtedly a product of the Hebrew religious imagination and moralistic temperament."¹  A young-earth writer who quotes Kramer's "beyond all doubt" is quoting a scholar who drew the opposite conclusion about which story came first.  Wenham's summary is the fair one:  "on either view the tower of Babel story can be seen as an interesting comment on earlier tradition."¹¹  Longman agrees that "we should not make too much of this connection."¹³  The article on ancient monotheism shows why the spell is no evidence for an original belief in one god either.

Bricks for Babylon in Enuma Elish

The Babylonian creation epic ends with the gods building Babylon for Marduk.  Tablet VI, in W. G. Lambert's edition:  "'Build Babylon, the task you have sought.  Let bricks for it be moulded, and raise the shrine!'  The Anunnaki wielded the pick.  For one year they made the needed bricks.  When the second year arrived, They raised the peak of Esagil, a replica of the Apsû.  They built the lofty temple tower of the Apsû".⁴

E. A. Speiser in 1956 set Genesis 11:3-4 beside these lines.  Both have a solemn moulding of bricks, and Westermann notes that the Hebrew "let us brick bricks" has an exact Akkadian equivalent in the epic's libittašu iltabnu.³  Both raise a head:  line 62 raises "the peak", in Akkadian the head, of Esagil, and Genesis 11:4 has a tower whose head is in the heavens.  Speiser concluded that "each item in the biblical narrative proves to have its appropriate Mesopotamian antecedent" and that the source was a text, not a building,⁵ partly because "the earliest known references to Etemenanki do not antedate the seventh century B.C."⁵  He also saw that the Hebrew writer had to explain the bricks to his readers:  the detail "bothered the Hebrew writer sufficiently to cause him to include a clause for the benefit of his readers".⁵

Speiser's dating point is out of date.  A. R. George showed in 2005 that a Middle Assyrian manuscript of the epic proves it existed in the second millennium BC, and that the ziggurat of Tablet VI line 63 is Etemenanki.⁶  Babylon had a famous tower in its literature by the late second millennium at the latest.

The polemic reading rests on a reversal.  In Enuma Elish gods build Babylon to honour a god.  In Genesis men build it to make a name for themselves, and God comes down to see it.

K 3657 and the missing Babylonian story

In 1876 George Smith and W. St. Chad Boscawen published a broken tablet, K 3657, as a possible Babylonian account of the tower.  Lambert:  "Since King in 1902 showed that this was a misconception, the text has been very generally neglected."  The tablet is about the people of Babylon, "burdened with hard labour", whose lament keeps a god from sleeping.⁴  Skinner had already written in 1910 that it was "shown to contain nothing of the sort".¹⁴  Westermann's survey found only the Sumerian spell:  "The motif of the confusion of languages occurs in a Sumerian myth about the gods without any connection with the building of a tower."³

Herodotus, writing about 430 BC, describes the tower:  "a solid tower has been built, two hundred and twenty yards long and broad; a second tower rises from this and from it yet another, until at last there are eight."¹⁵  He tells no story of its building, no confusion of tongues and no scattering.  The Greek tellings of the Babel story all come through Jewish writers, as The Tower of Babel shows.

"Its top in the heavens": the building inscriptions

The phrase in Genesis 11:4 is a standard boast of Mesopotamian builders.  Nabopolassar, Nebuchadnezzar's father, says Marduk commanded him to rebuild Etemenanki "to ground its bottom on the breast of the 'Great Place' (i.e. the netherworld), to make its top vie with the heavens".  George's comment:  "This is a common cliché: many great buildings are said to bridge the gap between the lowermost and uppermost levels of the cosmos."⁶  Skinner noted in 1910 that "similar language is used by Tiglath-pileser I. of the towers of the temple of Anu and Ramman", an Assyrian king centuries older than Nabopolassar.¹⁴

The tower's own name says the same thing.  The Sumerian E-temen-anki means "House, Foundation Platform of Heaven and Underworld";  the tower at Borsippa was "House that Controls the Seven Me's of Heaven and Underworld."⁸

Nebuchadnezzar II's foundation cylinder for the tower, which George translated in 2011, is the closest single text to Genesis 11:3-4.  His father had raised it "in bitumen and baked brick to a height of thirty cubits on the outside, but which he did not finish (to) the top—I set to work on finishing E-temen-anki (to) the top so that it vied with the heavens."⁸  The labour came from everywhere:  "all the teeming people of the world, whom my lord Marduk had consigned into my possession, I conscripted for labour in the building of E-temen-anki and imposed corvée-duty on them."⁸  The stele that shows him before a seven-storey tower says the same:  "I mobilized [all] countries everywhere".⁸  Baked brick, bitumen, an unfinished top, a top that rivals heaven, and the peoples of the world at one building site in Babylon:  most of the building scene of Genesis 11:3-4 is in a sixth-century BC royal text.

"One mouth": Sargon II

Two centuries before Nebuchadnezzar, Sargon II of Assyria built a new capital, Dur-Sharrukin, and filled it with deportees.  Grant Frame's 2021 edition:  "People from the four (quarters of the world), (speaking) foreign language(s) (and of) diverse speech ... whom I had carried off as booty by the power of my staff at the command of the god Aššur, my lord, I made act in concert and (re)settled (them) inside it".  Then Sargon "commissioned native Assyrians, masters of every craft, as overseers (and) commanders to instruct (the settlers) in correct behavior (and how) to revere god and king."⁷  The Akkadian behind "I made act in concert" is pâ ištēn ušaškin, literally "made (them) one mouth".  K. Lawson Younger translated it literally and took it to mean "the Assyrians instructed some deportees in Assyrian or some common language";  Frame follows the standard dictionary, for which the phrase means "to make act in concert".⁷

This is the text on which Uehlinger built the imperial reading described in The Tower of Babel.  On either translation its sense is political:  an empire turns many peoples, and perhaps many tongues, into one at its new city.  It shows where the vocabulary of Genesis 11:1 would have sounded political to an eighth-century hearer.  It is not evidence about whether the event happened.

Parallels from the rest of the world

Popular books list tower-and-tongues legends from Mexico, Burma, Assam and the Pacific.  James Frazer, an anthropologist with no wish to defend Genesis, collected them in 1918 and sorted them himself.⁹  They fall into three groups.

1. Climbing to heaven, with no confusion of language. The A-Louyi of the Upper Zambesi, the Bambala of the Congo, the Mkulwe of East Africa, the Ashanti, and the Anal Kuki of Assam each tell of people who try to reach the sky. None has a dispersal by language.⁹
2. A tower and a confusion of tongues. Frazer judged each of these a borrowing. In Diego Durán's version from Cholula, written in 1579, giants build a tower with "clay and a very sticky bitumen" and the lord of the heights says "Come, let us confound them". Frazer: "bitumen seems never to have been used by the Mexicans for such a purpose and is not found anywhere near Cholula." He quotes Edward Tylor that the confusion of tongues had "very probably been learnt from the missionaries".⁹ Of a Toltec tradition with a flood, a closed chest, a high tower and a confused tongue: "the coincidences with the Biblical narratives of the flood, the ark, the tower of Babel, and the confusion of tongues seem too numerous to be accidental."⁹ Of the Karens of Burma, whose story traces the tribe to Adam: "a tribe who display a peculiar aptitude for borrowing Christian legends and disguising them with a thin coat of local colour."⁹ Of a story from the Admiralty Islands: "There can be little doubt that this story is merely an echo of missionary teaching."⁹
3. Many languages, with no tower. "Not a few peoples have attempted to explain the diversities of human speech without reference to a Tower of Babel or similar structures."⁹ The Greek writer Hyginus has Hermes divide speech; the Wa-Sania of East Africa and the Kachcha Nagas of Assam have their own stories.

So the parallels that look most like Genesis are the ones written down after missionaries arrived, and the ones from before contact lack either the tower or the confusion.  The bitumen at Cholula gives the borrowing away.  Skinner pointed to a different link, "the unmistakable resemblance between the Mexican pyramids and the Babylonian zikkurats",¹⁴ and Habermehl argues that after the dispersion "men went on to build ziggurats and temple complexes wherever they settled", patterned after the one begun at Babel.¹⁶  A likeness of buildings is a separate argument from a shared story.  It has to carry a design across thousands of years and oceans with no examples in between on most routes, and the abstract does not say how.

What the texts show

• The Mesopotamian details of Genesis 11:3-4 are right, down to the idiom. Speiser, Westermann and George all show it. That is a real point for the text's knowledge of Shinar.
• No Mesopotamian text tells the story. The one tablet once offered is about something else. The Sumerian spell has one language and a god who changes speech, its tense is disputed, and it has nothing else of Genesis 11.
• The closest parallels in thought are Sargon II in the eighth century BC and Nebuchadnezzar II in the sixth. They fit a first-millennium audience hearing the story as satire on empire.
• The boast of a top in heaven is older than those kings. It is in Tiglath-pileser I and in Enuma Elish, which existed in the second millennium. So the vocabulary alone does not prove a late date of writing.
• No text from anywhere corroborates the event independently. The young-earth literature read for this article has no published answer to the point that the closest parallels are first-millennium. An answer would need an earlier text with the same motifs, or an argument that Genesis 11 was written before them.

Objections and responses

"The story is a satire on Nebuchadnezzar's tower, written in the Exile."  The closest parallels in thought are sixth-century, and a sixth-century audience would have heard the story that way (Part 1).  But the tower was not new then.  George shows that the 91 m tower already stood under Esarhaddon in the seventh century, that its smaller precursors go back into the second millennium, and that Enuma Elish, which existed in the second millennium, already celebrates a ziggurat at Babylon.⁶⁶  A writer of any date from the late second millennium on could have known a great tower at Babylon.  The parallels show who would have heard the story as political.  They do not date its writing, and they say nothing about whether the event happened.

"The Sumerians remembered Babel in the Spell of Nudimmud."  Kramer read the spell as a memory of one past language changed by a god, and his reading is not impossible:  Vanstiphout concedes the grammar could bear it.²  But most editors read it as a wish for the future, it has no tower, no city and no scattering, and Kramer himself made Genesis the later and derivative story.¹  Osgood's identification of Nimrod with Enmerkar, the king of the epic, is a separate claim:  "whom I will identify with the personages of Cush and Nimrod (Sumerian MeshkiagKASHer and Enmerkar)".¹⁷  No text names Enmerkar as Nimrod, and the epic does not mention a dispersion.

"World legends of a tower and confused tongues confirm Babel."  Frazer, Tylor and Skinner traced the ones that look most like Genesis to missionary teaching after contact.  The Cholula version has bitumen, which "is not found anywhere near Cholula".⁹  The stories recorded before contact lack either the tower or the confusion (Part 1).

Issues and unanswered questions

• The grammar of the Spell of Nudimmud. The rule for the Sumerian modal prefix in line 146 has been checked in Zólyomi's grammar, and it favours a statement over a wish. Whether any Sumerologist has applied it to this line in print is not known: the later studies by Bendt Alster (1973), Thorkild Jacobsen (1992), Vanstiphout (1994) and Jacob Klein (2000) were not read, and one of them may answer it. Other manuscripts of line 146 were not checked for a present-future form of the verb.

Issues that bear on more than one of these articles, among them the imperial reading, the printed pages of George 2005 and the young-earth books not read, are listed in the parent, The Tower of Babel.

Verdict

Genesis 11 knows Shinar, and no text from Shinar or anywhere else tells its story.  The building details of Genesis 11:3-4 are right down to the idiom:  the year of brick-making and the raised head of Esagil in Enuma Elish, the Akkadian twin of "let us brick bricks", and the builders' boast of a top that vies with heaven.  Speiser, Westermann and George all show it, and that is a real point for the text.

There is no Babylonian story of a tower and the confusion of tongues.  K 3657 is about the toil of Babylon's people.  The Sumerian spell has one language and a god who changes speech, its tense is disputed, and it has no tower, no city and no scattering.  The texts closest to Genesis in thought are the royal inscriptions of Sargon II and Nebuchadnezzar II, in the eighth and sixth centuries BC.  They show how a first-millennium audience would have heard the story.  They do not show when it was written or whether it happened, since the boast of a top in heaven is older than both kings.  The world legends most like Genesis were written down after missionaries arrived, and the older ones lack the tower or the confusion.

The young-earth literature read for this article has published no answer to the first-millennium date of the closest parallels.  It would need an earlier text with the same motifs, or an argument that Genesis 11 was written before them.

Companion pages

• The literature: every source consulted for the Tower of Babel articles, what it shows, and which way it cuts.
• Quotations: the key quotations behind the Tower of Babel articles, each with its locator.
• The Tower of Babel: what Genesis 11:1-9 says, how it has been read, and the verdict on every test.
• Where Was the Tower of Babel?: the bricks, bitumen and temple platforms of Shinar, the level-count test, and the six proposed sites.
• When Was the Tower of Babel?: the window Peleg's life allows on each text, where AiG's 2242 BC comes from, and what the record shows at each date.

Sources