Middle Assyrian Kingdom

Two records, one state

Drafted from Frahm, Assyria: The Rise and Fall of the World's First Empire (2023)formation

There are two Assyrian empires, and they are the same city, the same god, the same royal title and very nearly the same institutions. This one runs from 1363 to 1056. The one everybody means by the word runs from 911 to 609, and begins a hundred and forty years after this one ends.

Whether that makes them one state or two is a real question and not a bookkeeping one. Some historians of Mesopotamia treat Babylon as a single continuous object across seventeen centuries, on the reasonable ground that the city, the gods and the language persist through every change of dynasty. The other way of counting asks what the state was doing, and by that test the interruption matters: in the years between these two Assyrias the country lost every province, stopped campaigning and stopped writing.

Nothing passed between them in the ordinary sense. Nobody claimed the earlier kingdom's legitimacy, because there was nobody to claim it against; nobody conquered anybody. What happened is that the same place, under the same god, started doing the same things again after a long silence — and there is no ordinary word for that relation, which is why it keeps having to be described rather than named.

Out from under Mitanni

Ashur in 1400 was a city-state paying tribute to Mitanni. Seen from the other side, Ashur was one of the dependencies that made Mitanni a great power, and the Mitannian king treated its ruler as a subject.

Ashur-uballit I, who came to the throne around 1363, ended that. He took advantage of the wars in which Suppiluliuma I was dismantling Mitanni from the west to take the eastern half, and then he did something more consequential than any conquest. He wrote to Egypt.

The letter is in the Amarna archive. Ashur-uballit addresses Akhenaten as a brother, which is the term great kings used to each other, asks for gold, and complains about the treatment of his envoys. It provoked a furious letter from Burna-Buriash of Babylon to the same pharaoh, pointing out that the Assyrians were his own vassals and had no business writing to anybody, and asking Egypt not to receive them.

Egypt received them. That exchange is the moment Assyria enters the club, and it is preserved because a heretic pharaoh's abandoned capital happened to keep its foreign correspondence and be excavated in 1887. Almost everything known about this century depends on accidents of that kind.

The ginā'u

Drafted from Postgate, Bronze Age Bureaucracy: Writing and the Practice of Government in Assyria (2013)expansion

Every province of the Middle Assyrian kingdom owed a fixed recurring delivery to the temple of Ashur: grain, oil, honey, fruit, in specified quantities, on a schedule. It was called the ginā'u, the regular offering. A central office recorded what each province owed, what each province sent, when it arrived and who brought it, and chased what was late.

Those records survive. Postgate works through them, and through nine other government archives from Ashur and the provinces, and what comes out is a description of how the state worked that no narrative of kings could supply.

It is a tax in the strict sense — assessed in advance, scheduled, delivered, audited, arrears pursued — and it is functioning in the thirteenth century BC. The next tax system anywhere that can be described at the same resolution is the Neo-Assyrian provincial apparatus, six hundred years later, and after that the Neo-Babylonian temple accounts, eight hundred.

What writing did

Postgate's larger argument, and the reason his book is called what it is, is that the writing was not a record of the administration but a part of it. Middle Assyrian officials wrote to each other constantly, sealed and archived what they wrote, and structured their responsibilities around documents that could be produced later. A governor's obligation was defined by what the tablet said. The practice of government and the practice of writing were the same practice.

The army worked the same way. The ilku was land held from the crown against an obligation of service, and versions of that arrangement turn up in most premodern states — usually known from a historian describing it in general terms. Here the men owed, the equipment owed and the substitutions permitted are itemised in archives kept for the purpose.

Deportation, early

The other institution this kingdom is running three centuries ahead of schedule is the resettlement of conquered populations. Deportation as a system — populations moved, mixed and resettled as military colonists and as labour, deliberately and at enormous scale — is what the later Assyrian empire is most notorious for, and it is usually dated to Tiglath-Pileser III's reorganisation after 745.

The practice starts here. Tukulti-Ninurta I claims tens of thousands of people moved out of Babylonia and the east, and the administrative texts show deportees settled on state land and issued rations. The scale is smaller and the intent is recognisably the same.

So the machinery usually credited to the Neo-Assyrian empire — provinces with appointed governors, an assessed tax, service land, systematic resettlement, government by archive — is substantially present six hundred years earlier in the same city. What 745 did was not invent it, but apply it to about three and a half times as much ground.

Hanigalbat

Drafted from Frahm, Assyria: The Rise and Fall of the World's First Empire (2023)expansion

Adad-nirari I, and then his son Shalmaneser I, took what was left of Mitanni and turned it into Assyrian provinces under Assyrian governors. The country the Assyrians called Hanigalbat — northern Mesopotamia between the Tigris and the Euphrates, the Habur triangle, the Mitannian heartland — stopped being a kingdom around 1260 and became administered territory.

That is the end of Mitanni, and it is worth noticing who ended it. Suppiluliuma had broken the kingdom from the west two generations earlier and taken its Syrian empire; what he left standing was a rump. The Assyrians finished it, quietly, over about thirty years, and turned it into a province.

What that leaves of Mitanni

Almost nothing of its own. A great power of two centuries with no capital that has ever been located, and a history reconstructed entirely from other people's documents — Hittite treaties, Egyptian letters, and now Assyrian provincial administration. Adding the polity that ended it does not make it legible. It makes the illegibility attributable.

Facing west

The strategic consequence is the one the Hittite record's fifth chapter is about. With Hanigalbat absorbed, Assyria stood on the Euphrates facing Hatti directly, and by the 1230s Tukulti-Ninurta I would defeat a Hittite army and boast in an inscription about the number of prisoners taken.

That is why Hattusili III wanted a treaty with Egypt in 1259. The southern frontier had been a war for two centuries and was, by then, the less dangerous one. The famous peace between Hatti and Egypt exists because a third power had made the Hittite position in northern Mesopotamia untenable, and that third power is not a party to it.

Three great powers, two treaties, a marriage and forty years of correspondence, and not one of the three ever succeeded, inherited or governed either of the others. They determined each other completely and are connected by nothing that an account of successions would register.

The murder in the new capital

Drafted from Frahm, Assyria: The Rise and Fall of the World's First Empire (2023)peak

Tukulti-Ninurta I reigned thirty-six years and did two things that no Assyrian king had done. He conquered Babylonia, and he left Ashur.

In about 1225 he took Babylon, deported the Kassite king Kashtiliash IV in chains, and carried the statue of Marduk north. It is the first time Assyria held Babylonia and it lasted roughly seven years. The relationship between the two states that Sennacherib would poison for good in 689 — sacking the city, demolishing its temples, diverting a canal over the site, and setting off a chain that ended with his grandsons at war and, eventually, with Nabopolassar — begins with this campaign. An Assyrian epic was written about it, and a Babylonian text was written accusing Tukulti-Ninurta of sacrilege.

Kar-Tukulti-Ninurta

Around 1233 he built a new capital directly across the Tigris from Ashur and moved into it, naming it after himself: Port Tukulti-Ninurta.

A king of Ashur leaving Ashur is not a change of address. The city and the god have the same name; the royal title is iššiak Aššur, the god's steward, and the stewardship is of that place. Sargon II tried the same manoeuvre at Khorsabad seven hundred years later. Both new capitals were abandoned within a generation of their builder's death, and both builders died violently.

In about 1207 Tukulti-Ninurta was besieged in his own new city by a conspiracy led by his son, and killed.

The rule and the outcome

Assyria had a rule for the succession: the king designated an heir. Seven hundred years later Esarhaddon would give that rule the most elaborate instrument the ancient world produced — an oath sworn by every subject and vassal of the empire — and the two sons it was written to protect fought a four-year civil war anyway.

Rules and outcomes are different objects, and Mesopotamia keeps demonstrating it in both directions. Babylon under the Neo-Babylonian kings had three coups in five transitions. Ur III and Old Babylon had eleven and five flawless father-to-son successions between them and never wrote down a rule at all. Here the rule is clear and the greatest reign in the kingdom's history ends with the king killed by his own son.

Four records, four different relationships between the rule and what happened, and the field says the same word or nothing. That is the field working correctly and it is also the limit of what a coded vocabulary can carry, which is why the chapters exist.

After

The empire came apart within a decade. Babylonia was lost, the west was lost, the new capital was abandoned, and Assyria did not reach the same extent again for a century — until Tiglath-Pileser I, the greatest of these kings, who reached the Mediterranean, cut cedars in Lebanon, and hunted elephants in Syria, all of which he recorded at length.

The sea-horse

Drafted from Luckenbill, Ancient Records of Assyria and Babylonia, Volume I (1926)peak

Tiglath-Pileser I ruled at the end of this state's life and was the furthest it ever reached. His inscriptions are also the point at which Assyrian royal writing becomes recognisably itself — the annual campaign, the count, the boast — and they are strange in a way the later versions are not.

To the Upper Sea

The claim is geographical and it is enormous: from beyond the Lower Zab to the Upper Sea of the setting sun, meaning the Mediterranean.

The Nairi lands north in the mountains were campaigned in three times, and thirty kings of Nairi were brought in submission, hostages taken, yoke-broken horses received as gifts, tribute and toll imposed. Then Lebanon, for cedar for the temple of Anu and Adad at Ashur — the logs cut and carried away, as every Mesopotamian king who could reach that far had done and would do.

Then the coast. Byblos, Sidon and Arvad paid tribute. And Tiglath-Pileser went to sea in Arvadite ships, from Arvad to Samuri, a journey he measures in the overland unit — three double-hours by land — which is how a man from the Tigris records a distance he has just travelled by water for the first time.

And in the middle of the sea he killed a nahiru, which he explains for his readers: which they call a "sea-horse."

An Assyrian king went whaling. He thought it worth putting in the same inscription as the thirty kings and the cedar, and he thought his audience at Ashur would need the animal glossed.

The count

The hunting records are kept with the same apparatus as the campaigns, and the numbers are the point of them.

Four wild bulls, mighty and of monstrous size, killed in the desert in the country of Mitanni near Araziki opposite the land of Hatti, with the bow, the iron spear and sharp darts; their hides and horns brought to Ashur. Ten bull elephants killed in the country of Harran and along the Habur, four more caught alive; the hides, the tusks and the live elephants brought to Ashur.

Then: one hundred and twenty lions killed on foot, and eight hundred more from the chariot.

Nine hundred and twenty lions is not a plausible personal tally and it is not meant to be read as an afternoon's work. It is a reign-total, kept the way the conquests are kept, because in this system killing lions and taking cities are the same activity — the king imposing order on what is outside it — and both belong in the ledger.

The elephants are worth pausing on for a different reason. There were elephants on the Habur in the twelfth century BC, and Assyrian kings hunted them there. They were gone within a few hundred years, and the hunting is part of why.

What it was the peak of

Everything in these texts is about reach: the sea in the west, the mountains in the north, animals brought back alive to be looked at in the capital, timber carried the length of the Fertile Crescent for a temple roof.

None of it was held. There is no Assyrian province on the Mediterranean at the end of this reign and no garrison at Arvad — there is tribute, a hunt, and a record. What Tiglath-Pileser I established was not a frontier but a claim about what an Assyrian king was for, and three centuries later a much larger state would go back over exactly this ground and this time keep it.

He died in 1076, and what happened to Assyria immediately afterwards was not a defeat.

A reminder to the governors

Drafted from Luckenbill, Ancient Records of Assyria and Babylonia, Volume I (1926)contraction

The reign after the furthest reach is documented by almost nothing, and the almost-nothing is informative.

Assur-bel-kala, Tiglath-Pileser's son, came to the throne — Luckenbill's guess is probably after a long struggle, which is the kind of sentence an editor writes when the succession has left no record and the next reign has to be explained somehow.

Where the text is

His main surviving inscription is not on a stele, a temple wall or an obelisk. It is written on the back of the torso of a female statue, now in the British Museum.

And Luckenbill's reading of who it was for is the whole chapter: the text seems to have been intended for his city governors, as a reminder of their allegiance to him.

It opens the way any Assyrian royal inscription opens, with the titulary — king of the universe, the mighty king, king of Assyria — and the ancestry, his father and his grandfather each given the same titles. Then a damaged passage about statues and about the curses placed on them. Then the standard closing formula against anyone who alters the inscription or the name: may the god Hanish and the gods of the land of Amurru smite him in open battle.

What has changed

Compare it with what his father had been writing a few decades earlier.

Tiglath-Pileser's inscriptions count thirty submitting kings, cedar cut in Lebanon, a sea crossing, a whale, nine hundred and twenty lions. His son's inscription counts nothing. It has no campaign, no tribute list, no building programme of any scale. What it has is a genealogy and a curse, addressed to the men who administered the cities of Assyria.

A king does not remind his own governors of their allegiance when it is not in question. The document is a small piece of evidence, and it is nearly the only piece, that the thing beginning to fail was not the frontier but the relationship between the centre and the men holding the provinces for it.

The shape of the evidence itself

There is a second lesson in it, and it is about what survives.

The material for this reign is a statue torso, a stele that may or may not be his, and two tablet fragments that were catalogued as a religious text until a scholar in 1917 noticed that one of them was historical. That is the whole dossier for the son of the greatest king the Middle Assyrian state produced.

What the record is about to do is not stop because Assyria was destroyed. It is about to stop because the apparatus that produced records — the scribes, the provincial administration, the campaigns worth commemorating — was contracting faster than the territory was. The silence comes first.

The blank spaces

Drafted from Luckenbill, Ancient Records of Assyria and Babylonia, Volume I (1926)aside?What aside means

There is an object in the British Museum, said to have come from Kuyunjik, that nobody has been able to assign to a king. It is called the Broken Obelisk, and it has been furnishing scholars with arguments, in Luckenbill's phrase, ever since its inscription became known.

The problem

The obelisk carries a long introduction recording the conquests and hunting exploits of an Assyrian king — in the third person. At the end there is a building inscription, and that one is in the first person.

Two voices, one monument. The obvious reading has been that the introduction belongs to Tiglath-Pileser I, whose campaigns and hunts it resembles, and the first-person section to some later king who put the thing up.

Which later king depends on evidence that keeps moving. When eponym lists and the annals of Adad-nirari II came to light, it seemed almost certain to Luckenbill that the obelisk in its final form belonged to that reign: the building section mentions a canal dug by Ashur-dan and restored thirty years later, which fits Ashur-dan II and his son Adad-nirari II better than any other pair, and it mentions the rebuilding of the city of Apku, which Adad-nirari's own annals describe in detail.

Julius Lewy argued instead that the third-person narrative was Tukulti-Ninurta I, three centuries earlier — plausible, because Adad-nirari II's own son was named after that king. Luckenbill records the argument, says that until eponym lists turn up containing the names in the obelisk any solution must be received with caution, and then notes that the traces of a royal name in one damaged line appear to be a contemporary of Tiglath-Pileser I, in which case Lewy's argument falls.

Where the numbers should be

Luckenbill's own view is the one worth keeping, and it rests on a physical detail.

He believes the monument was inscribed with the annals and hunting exploits of Tiglath-Pileser I, and that it was never finished — because blank spaces are left in the stone where the numbers of slain beasts were to stand. Then, long afterwards, Adad-nirari II acquired merit by setting the unfinished thing up and adding a short inscription of his own to the end.

A monument cut for one king, abandoned with the kill-counts unengraved, and raised generations later by a different king who signed it.

Why it belongs here

Nothing states the condition of Assyria between Tiglath-Pileser I and its recovery more exactly than that.

The state that could commission the stone could not finish it. Whatever interrupted the work — the reign that ended, the scribe who was not replaced, the campaign whose totals never arrived — the object sat with holes in it where the achievements went. And when Assyria came back, a century and a half later, the first thing one of its kings did with the abandoned monument was stand it up and put his name on it.

He was not being fraudulent. He was doing exactly what Ur-Nammu did when he built his ziggurat around Naram-Sin's: claiming the sequence. What makes this one harder to read is that the object he claimed had never been completed, so the argument about who it belongs to is an argument nobody in antiquity had settled either.

The silence

Drafted from Frahm, Assyria: The Rise and Fall of the World's First Empire (2023)end

Nobody conquered Assyria. That is the difficulty with the ending, and there is no agent anywhere to name for it.

What happened, over the decades after Tiglath-Pileser I's death in 1076, is that Aramaean groups moved across the country between the Assyrian core and the Euphrates and stayed. There is no battle anybody names, no capital taken, no king deposed. The provinces west of the Tigris stop appearing in the record; the Habur is gone; the routes to the Mediterranean are gone. By the death of Ashur-bel-kala in 1056 Assyria held not much beyond the triangle of Ashur, Nineveh and Arbela — and then, for something like a hundred and forty years, it stops writing.

Not entirely: there are king lists, there are eponym lists, there are scraps. But the archives Postgate works from end, and the royal inscriptions become brief and few, and the century and a half between this kingdom and the empire that follows it is the darkest stretch in Mesopotamian history after the Bronze Age collapse itself.

None of the ordinary ways of describing how a state ends quite fits what happened here. The state did not break into pieces, which rules out the plainest reading of a collapse. It was not absorbed into anybody else's kingdom, since it was still there in 911, with the same god, the same title, the same city and, when the documents resume, recognisably the same institutions. It was not conquered, because a conquest needs a conqueror and none is named anywhere in the record. What actually happened sits closest to a state shrinking back onto its own core and going quiet — the least wrong of the available descriptions, and no more than that.

The same state, later

In 911 Adad-nirari II began the reconquest, and the empire everybody means by Assyria starts there.

Nothing connects the two in the ordinary sense. Nobody claimed the earlier kingdom's legitimacy, because there was no rival claim to answer; nobody conquered anybody. What resumed in 911 was the same city, the same god, the same royal title, the same provincial machinery and, as far as anyone can tell, the same families.

The same thing is true of the two Babylons, separated by a thousand years instead of a hundred and forty, and of the two Elams. In each case the ordinary vocabulary fails: it is not a succession, not a conquest, not a restoration by anybody's claim, and not a new state either. It is a place going quiet and then starting again.

Why that is hard to say

Because the relation has no date. There is no morning on which Assyria resumed being Assyria — only a stretch of about a hundred and forty years during which it was quieter, at the end of which somebody starts campaigning and writing it down again.

And because the obvious test is the wrong one. Sharing a capital proves nothing: plenty of unrelated states have governed from the same city after a gap — the Umayyads and the Burids at Damascus, the Idrisids and the Marinids at Fez, Han and Tang at Chang'an. This case does not even share a capital, since the earlier kingdom ruled from Ashur and the later empire from Nineveh and Nimrud.

The test that works is whether the sources treat the two as periods of one continuous object. On that test Assyria is one thing with a long silence in the middle, and the silence is the only reason it needs explaining at all.

The shapeless mass

Drafted from Layard, Nineveh and its Remains (1849)afterlife

The Middle Assyrian kingdom has two afterlives. One is the empire that resumed its work three centuries later and is what everybody means by Assyria. The other is what happened to its capital.

Ashur

Ashur was the city, the god and the country, and the kingdom's kings called themselves the god's steward of that place. It was where the archives were kept, where the assessed tax was audited, where the king was crowned and buried.

In 1840 Austen Henry Layard rode out to look at it, though he did not know that was what he was looking at. He calls it Qalat Sherqat and describes it as a vast, shapeless mass, covered with grass, showing scarcely any trace of the work of man except where the winter rains had cut ravines down its almost perpendicular sides and laid its contents open.

He and his companion spent the better part of a day going over the heaps of earth and brick. There was an Arab tradition of strange figures carved in black stone somewhere in the ruins; they looked and did not find them. What they did find was a few fragments of pottery and some inscribed bricks — enough, Layard noted, to prove that the mound had been built by the same people who built Nimrud.

That is the whole of it. The capital of a kingdom that had run a provincial tax system, a service-land army and a state archive in the thirteenth century BC was a grassy hill with potsherds in the runoff, and the visitors could establish nothing about it beyond that somebody Assyrian had made it.

The paperwork

What survived instead was the writing, and it survived because clay does.

The archives Postgate works from — the tax records, the land grants, the inventories of what each province owed — came out of the ground at Ashur when the German expedition dug there from 1903. So did the Middle Assyrian Laws, the fullest legal text of the Late Bronze Age, and the palace edicts regulating the royal household.

The result is an unusual inversion. This kingdom is invisible in the standard histories, is not the Assyria of the prophets or the museums, has almost no famous objects, and is the best-documented administrative state of its century anywhere in the world.

The other afterlife

The same machinery did start again, under the ninth-century kings everyone means by Assyria, for the reasons already given.

What it started again with is the point worth closing on. The empire of Ashurbanipal is usually credited with inventing the provincial governors, the assessed tax, the land held for service, the resettlement of conquered populations, the whole apparatus of government by archive. It inherited all of it. What it added was the scale, and the scale is the reason one of the two Assyrias is famous and the other is a mound somebody walked over in 1840 without recognising it.