The Murašû Archive

The archive was found in one place, all at once. The first American excavations at the ruins of Nippur, in central Babylonia, ran intermittently from 1888 to 1900; the third of four campaigns opened in April 1893 under John Henry Haynes, working a mound in the north-western part of the site called Camp Hill. Late in May the workmen struck a small buried room:

At the end of May workmen began to find cuneiform tablets on the floor of a small room about six meters below the surface of the mound. They cleared the room between May 27 and June 3. They recovered from it about 330 intact tablets, 400 or more damaged tablets and fragments, and 20 small clay tags with seal impressions.

Stolper, Entrepreneurs and Empire, p. 1^[1]

A search of the adjoining ground the following week turned up nothing more. Haynes sent the tablets to Istanbul, where Hermann Vollrat Hilprecht examined them in 1894 and, in 1898, identified them as the late-Babylonian records of a business house he named "Murašû Sons of Nippur", after the ancestor of the firm's chief members. From that identification the group took the name it still carries.^[1] What one room had preserved was the working file of a single family firm, and Stolper set out at the start of his edition why so confined a find matters so much:

The Murašû Archive is still the largest single source of written evidence on conditions in Babylonia at a time when the Persian Empire's political form had matured, its hold on its provinces was established, and its effects on Babylonian society had begun to tell.

Stolper, Entrepreneurs and Empire, p. 1^[1]

The tablets are dated documents, and they cluster in a narrow band of history. The earliest, BE 9 2, was drawn up on 17 March 454 BCE, in the tenth year of Artaxerxes I; the latest, BE 9 1, on 1 November 404, under Artaxerxes II, a span of about half a century.^[2] The find scattered on excavation and sale: the main collections went to Istanbul and to the University Museum in Philadelphia, with a group in the Hilprecht collection at Jena, one tablet at Berkeley and four in the British Museum. The distribution of what was published skewed the picture. The first edition (Hilprecht and Clay's BE 9, 1898) drew on the reign of Artaxerxes I; the two that followed (Clay's BE 10 of 1904 and PBS 2/1 of 1912) tried to publish everything from the reign of Darius II, so that, in Stolper's words, "These editorial choices skewed the chronological distribution of the published evidence in favor of the Archive's later years."^[3] Counting joins and antique breakage, Stolper reckoned the known Murašû texts and fragments at 879, and judged Hilprecht and Haynes's original estimate of the archive's size, about 730 items, "probably not far off the mark".^[4]

There was, strictly, no "firm". No word in the tablets names the business, and its unity is an inference from the way the texts behave. Stolper stated the problem and its solution plainly:

The Murašû texts are records of a single business house. It had no ancient name… each was drafted in the interests of one or more members of the family descended from Murašû, or in the interests of the family members' agents.

Stolper, Entrepreneurs and Empire, p. 18^[5]

No term in the tablets can be accurately rendered as "the firm"; what holds the texts together is that they were kept together, in one room, drafted for members of the family or their agents, the members acting as co-proprietors and their agents serving several of them in turn. On that basis, Stolper argued, the fact that the archive "supports interpretation as a coherent record of complementary activities justifies use of 'the Murašû firm' as a historical postulate".^[5] Even the actor whose doings the archive records is a modern reconstruction.

The family ran four generations. Behind it stood the eponymous Murašû, active around 500 to 445 BCE and roughly coeval with Persian rule in Babylonia; property registered to him in 500/499 was still being litigated over in 445. The business proper was carried by two principals in succession. Enlil-šum-iddin, a son of Murašû, dominates the texts from 445 to 421 BCE and did the great majority of the firm's recorded business. His nephew Rīmūt-Ninurta succeeded him and is the central figure from 429 to about 414, after which his rents and taxes were collected not by him in person but by his agents, a sign of the firm's management going indirect as it neared its end.^[6] The one woman in the family tree, Naqqītu, survives on a single fragment.

Round the principals worked a graded staff whose titles map the society they moved in. Servants (ardu) ranged from chattel slaves to trusted managers; one, Bēl-erība, issued contracts on his own account, "qualifying him better as a principal of the firm than as a mere employee".^[6] The most consequential term is mār bīti, "son of the house". In one use it means a steward of a named person; in another, unqualified or as mār bīti šarri, "son of the (royal) house", it renders the Old Iranian *visa-puθra- and means a prince, a male member of the Achaemenid royal family, and these princes were among the firm's most influential customers.^[6] The firm's reach was geographical as well as social. Its base at Nippur sat, in Stolper's phrase, at "a transit point between the provincial capital and the southern center of Uruk, and between the royal residences at Babylon and Susa", and its domain covered the city's hinterland, a region of maybe a hundred kilometres square, threaded by six major canals, sixty lesser waterways and more than 180 settlements.^[6] Most contracts were drawn at Nippur; a handful were written at Babylon and two at Susa, marking the firm's occasional business at the imperial capitals.

The firm was chiefly a manager of other people's land. It held a few fiefs of its own, but that was marginal: in the great majority of its transactions, Stolper found, the house "figured not as a participant in the system of land grants but as an accessory to it, undertaking the management of property which belonged, on a variety of titles, to others".^[7]

Three lines of business run through the tablets. "The primary enterprise of the firm was agricultural management."^[7] The Murašûs leased land and water from their holders, paid the rents and taxes due on them, and sublet the properties to cultivators along with livestock, equipment and seed, keeping the margin. Second, they lent. Against pledges of real property they advanced silver and supplies, and the resulting certificates of obligation are the single largest class of text in the archive. Third, and invisibly, they turned crops into cash: their income came mostly as produce and their outgoings mostly as silver, so they must have retailed grain and dates to Nippur's townspeople for the metal the system ran on, a trade the surviving legal records do not document because its receipts stayed with the buyers. Stolper's summary sets the firm between the three systems it served: it "took by lease, pledge, or mandate the properties held by various members of the 'feudal' order, sublet them along with other necessities of agricultural production, converted their produce into silver, and supplied silver and credit to the firm's customers", mediating between the system of tenure, the system of production, and the system of taxation.^[7]

What to call such a business has been argued ever since. Stolper, following the foundational study by Guillaume Cardascia, cast the Murašûs as agricultural managers and credit brokers standing beside the land-grant system rather than inside it, and specifically not tax-farmers in the strict sense: they did not buy the right to collect a province's revenue, though they advanced the silver with which feudatories met their own assessments. Later economic historians have softened the framing. Reinhard Pirngruber places the firm in a "very tight nexus" of interlocking markets in land, credit and labour, agricultural entrepreneurs who advanced silver to peasants so those peasants could discharge their taxes and who took pledged land on default, and who could themselves fall into debt, market actors rather than simple predators.^[8] Michael Jursa, in the standard recent survey, calls the archive "probably the single most important source of information on late fifth-century BCE Babylonia, and certainly the best studied", while using the looser label of a family working "within the sphere of state-sponsored agriculture and tax farming".^[9] The change of vocabulary marks a genuine shift of emphasis, from an imperial system extracting from its subjects toward a monetising economy integrating its markets; each reading is built from the same body of contracts, read to opposite conclusions.

The holders the Murašûs dealt with were, in the great majority, occupants of military fiefs. The Persian crown in Babylonia rarely farmed its own land for profit; instead, as Stolper put it, "Most of the crown's revenues… were not drawn from the produce or rents of the king's own property, but from taxes and services incumbent on land held by the king's beneficiaries."^[10] The basic holding was the fief called "bow land" (bīt qašti), whose occupants, commonly a set of kinsmen, held it in return for at least nominal military service and a yearly tax, the ilku, rendered mostly in silver. Two grander variants are named beside it, "horse land" (bīt sisî) and "chariot land" (bīt narkabti), the three types graded by the arm each was bound to supply, an archer, a cavalryman or a charioteer. Whole or fractional fiefs could be leased or pledged but not sold, and the terminology reached back to the first Achaemenid years and forward into the Seleucid age (fuller treatment at tribute and taxation).^[11] Over time the fiefs shrank through divided inheritance while the crown's demand for silver hardened, so that subsistence allotments turned into cash-producing rental properties, and it was into that gap that the firm moved.

Fiefs were grouped into associations called ḫaṭrus, and the ḫaṭru is the institutional heart of the archive. About sixty are named, distinguished by ethnic, territorial, professional or military tags, each under a foreman (šaknu) responsible for allotting fiefs, keeping land in use, collecting taxes and ensuring that service was performed. As the foremen appear in the Murašû texts their main role is fiscal, and Stolper drew the institution's several functions together: the ḫaṭru "was in effect a small-scale fiscal district; the institution was a means of producing and extracting taxes for the Achaemenid state", at once a device for insuring agricultural production and for "supporting a standing military reserve, a local garrison force, and cadres of state-controlled workers".^[12] It was, he concluded, "a characteristic formation of the Achaemenid Empire".^[12] The best account of the system remains Cardascia's, whose 1946 study argued that the word ḫaṭru was in origin territorial, naming "un territoire, un canton concédé à une collectivité" (a territory, a district granted to a collective group), the sense of the landholding group itself being secondary.^[13]

Many of the ḫaṭrus were settlements of deported peoples, the documentary face of the empire's practice of moving populations onto the land. Among the groups named in the Nippur texts are Arabs, Armenians, Carians, Cimmerians, Indians, Lydians, Phrygians and Tyrians, each holding bow land and owing service and tax like any other feudatory.^[9] A single receipt can show the whole apparatus in miniature: a bow land held by men "of the ḫaṭru of the Cimmerians" and managed by Rīmūt-Ninurta, the ilku paid partly in silver and partly in kind, and the foreman named as the officer who receives it. The transplanted nations of the empire appear here not as war captives but as taxpaying smallholders on the Babylonian plain.

Whether the men who held bow land still fought, or merely paid, is answered by the archive's finest single document. On 8 January 421 BCE, in the second year of Darius II, Rīmūt-Ninurta had been called up for the cavalry service owed on a horse land in which he held a share. Rather than serve, he contracted with a certain Gadalyama to go in his place, fully equipped, and the deed records the bargain in Gadalyama's own voice:

Gadalyama, son of Rahim-ili, spoke of his own free will to Rimut-Ninurta, son of Murashu, as follows: 'Because Rahim-ili adopted your brother Enlil-shum-iddina, you now possess the share of Barik-ili in the field planted and in stubble, held in tenure as horse-land by Rahim-ili. … give me, (so that) I may go to Uruk in fulfilment of the royal order and (in return for this) I will fulfil the duty which rests on the horse-land as much as is your share.' Upon (this) Rimut-Ninurta heard him… he gave him; so that, in accordance with the royal command, he will go to Uruk to fulfil the duty on the horse-land mentioned.

Gadalyama's contract, Kuhrt 2007 no. 38 (Nippur, 8 January 421 BCE), trans. Kuhrt^[14]

The equipment listed is a full cavalryman's kit: a horse with reins and harness, an iron corselet and hood, a copper quiver, 120 arrows, a sword and scabbard, two iron lances, and a mina of silver for provisions. The document shows the militia was real and functioning under Darius II: a horse land still carried the duty to send an equipped rider to a royal muster at Uruk, in obedience to a "royal order", and the obligation could be met by a paid substitute. Pierre Briant reads the text as proof that around Nippur estates were subject to active military service, the identity of the man who performed it being a matter of indifference to the administration, and takes the assembly at Uruk not for a one-off expedition but a regular muster.^[15] On the same evidence Sean Manning has argued against the older view that the substitution of cash and hired men for personal service marked military decay; borrowing horses and gear and paying a replacement were sensible answers to the perennial problem of keeping a force mobilisable (see warfare and the army).^[16] The muster left its trace elsewhere in the archive too. A cluster of loans from the second year of Darius, secured on bow land, are for silver advanced "for clothing and military equipment" so that the holders could present themselves at Uruk, the technical term rikis qabli, "battle-array", proving that feudatories were mortgaging their fiefs to meet genuine service demands.^[17]

The firm acquired most of its land by lending rather than by leasing. Direct leases to the Murašûs are few; the bulk of the acquisitions came through the certificate of obligation (u'iltu), which accounts for fully a third of the texts, and the commonest kind was a mortgage. A debt was expressed in dates, the debtor was usually a holder of bow land, often a group of coparceners, and the security was the fief itself, its orchards and grain fields, with the ḫaṭru frequently named. The instrument bent the crown's silver demands to the farming year: it converted obligations in metal into obligations in produce, and deferred them to the date harvest.^[18]

Because bow land was legally inalienable, a defaulting debtor could not simply be foreclosed. Instead the pledge changed character: the use of the land passed to the creditor while the debt stood, so that "the debt and pledge began as a hypothec; if unpaid, it was converted into an antichresis". Stolper set this within a wider Near Eastern pattern he called "rent capitalism", the tendency to turn insecure titles to land into more durable, intangible debt-title.^[18] How costly this was to the smallholder is an old dispute: Cardascia thought a debtor's failure to redeem his land relatively uncommon, Mikhail Dandamayev thought it frequent.^[18] Stolper's tie-breaker was archival logic. A paid-off Babylonian debt-record was returned to the debtor or destroyed, so the mortgages that survive in the archive are largely those that were never redeemed, which makes converted mortgages, and the antichretic titles they yielded, a large part of the firm's business. A feudatory who lost the use of his fief this way could stay on as the firm's rent-paying tenant, the two parties, in Stolper's image, exchanging places on the ladder of rents to the Murašûs' advantage.^[18]

The archive's most arresting internal feature is a sudden concentration of these mortgages. Loans by feudatories spike sharply at the very start of Darius II's reign, peaking in the third and fourth months of his first regnal year. Cardascia had read the shifting ratios of the firm's business as a slow evolution; Stolper argued the change was too abrupt for that and needed a short-term cause outside the business itself, and found it in the contested accession.^[19] He weighed two explanations. On the first, feudatories tied to the losing side were liquidating vulnerable land into fluid assets against the risk of dispossession, a hedge; Stolper rejected it, since the mortgages spread evenly across some twenty ḫaṭrus regardless of their affiliations and bore no relation to the size of the property pledged. On the second, which he preferred, the mortgages record not strategy but compulsion. The mobilisation of Ochos, the future Darius II, was centred on Babylonia, and the fiscal and military levies of the contenders fell on the local feudatories at the worst point of the agricultural cycle:

Unable to meet both the demands of their farms and the added, urgent demands of Ochos's fiscal and military levies, many small landholders were forced into debt and obliged to mortgage their holdings.

Stolper, Entrepreneurs and Empire, p. 122^[19]

The crisis, Stolper concluded, showed the system working as designed, but with the benefits diverted: under a divided succession "the benefits of the system went not to the Empire's center, but to contenders for control of the center", while the costs "were transmitted to the lowest echelon of state dependents".^[19]

The land the firm managed was, to a striking degree, the property of the Persian ruling house. The precondition of Babylonian farming was water, and the canals were the crown's: the firm leased land from many holders but acquired water only from a distinct branch of royal administration, the canal managers and the officials called mašennu, and that dependence underpinned everything else it did.^[10] Above the fiefs stood the great estates of the royal family, and these too were partly parcelled into taxable bow land worked through the firm. The estate of the crown prince, run for Darius II's first years by a single overseer, contained both estate land proper and a ḫaṭru of its own; the taxes from its fiefs "did not pass directly to the crown, but went first to the management of the estate".^[20] The estate of Queen Parysatis (Babylonian Purušatu), wife and half-sister of Darius II, likewise "included bow fiefs" in the Nippur region, administered by her own steward and judged in her own court; one telling document suggests that on Darius's accession Parysatis succeeded to holdings that had belonged to the queen-mother of the previous reign, the estate records mirroring the dynastic transfer (see Parysatis).^[20] Beyond the queen, eleven other men are titled "prince" (mār bīti) in the texts, most of them owning land near Nippur and dealing with the firm through bailiffs. The concrete measure of the crown's penetration of the Babylonian countryside is here: royal and courtier estates, worked by local agents, sublet and financed by the Murašûs.

The heaviest of these princes, for the archive's later history, was Arshama (Aršam), the Achaemenid satrap of Egypt, who also held land near Nippur (see Arshama, satrap of Egypt). Nine texts from the original publications name him, and one, the archive's very last document, styles him mār bīti, prince. Others record Enlil-šum-iddin leasing Arshama's estate together with crown land, and a servant of Arshama suing the same principal. Stolper concluded that Arshama's connections with the firm ran "from 425 B.C. or earlier, through 404 B.C.—approximately the same interval during which he held office as governor of Egypt".^[20] The parallel with his Egyptian holdings is close: there too an absentee lord managed by a bailiff an estate divided into allotments held by subordinates who owed tax and service, a correspondence Stolper read as evidence of common structures across the empire's provinces. The same satrap thus appears both in the Aramaic letters from Egypt and in the cuneiform tablets from Nippur (see the Elephantine papyri).

The archive is also a chronological instrument of unusual precision, and it settles a disputed succession. When Artaxerxes I died, his throne was seized briefly by a son, Xerxes II, who was killed within weeks by a half-brother, Sogdianus, who was in turn overthrown by another half-brother, Ochos, who took the throne as Darius II. The classical narrative of these events, from Ctesias, disagrees with the cuneiform record by a year over when Artaxerxes died (see Ctesias).^[21] The Murašû tablets, dated to the day, resolve it. No Babylonian text is dated by Xerxes II or by Sogdianus at all:

Their date formulas proceed directly from the last regnal year of Artaxerxes I to the accession year of Darius II.

Stolper, Entrepreneurs and Empire, p. 115^[21]

The two ephemeral kings were never recognised in Babylonia. Fifty Murašû texts are dated in the forty-first regnal year of Artaxerxes I, running from April to December 424 BCE, and eleven in the accession year of Darius II, the earliest written at Babylon in February 423; several date-formulae specify that the forty-first year of Artaxerxes and the accession year of Darius were the same Babylonian calendar year, and two texts of late 424 still look forward to a coming forty-second year of Artaxerxes, proving that the Nippur scribes were dating a living king rather than perpetuating a fiction.^[21] From this Stolper drew a firm conclusion:

Babylonian texts show convincingly that Artaxerxes died in his forty-first regnal year; that Darius's official reign (and in Babylonia his actual reign) began immediately thereafter… not Artaxerxes' official reign, but the reign of their brother and rival Darius Ochos, as the political situation of the succession would lead one to suppose.

Stolper, Entrepreneurs and Empire, pp. 119–120^[21]

The crisis fell between the ninth and eleventh months of Artaxerxes' last year, between December 424 and February 423 BCE (see Xerxes II, Sogdianus and Darius II).^[21] The dating does more than fix a date. The very contenders of the succession reappear as landholders in the archive: the courtiers and princes whose Nippur estates the firm managed, Parysatis, Arshama and others, are the same people whose struggle for the throne, mobilised in Babylonia, drove the smallholders into the mortgages the archive records. The instrument that dates the crisis is also the ledger of its cost.

The archive stops, and how it stops is itself informative. The firm's active life closed around 414 BCE; its nine latest documents, running down to 404, are anomalous, leases of large flocks of sheep and goats in which no Murašû appears. The lessor is Enlil-suppē-muḫur, once a subordinate of the firm, now the bailiff (paqdu) of the prince Arshama, who owned the herds. Stolper read the shift as an absorption:

On this construction, the Murašû firm was in effect expropriated by a member of the royal family. The Murašû Archive in its preserved form… contains the records of an extinct business.

Stolper, Entrepreneurs and Empire, p. 23^[22]

On this account the archive ends because the firm's residue was folded into a royal estate administration, its records becoming an appendix to the crown's own management. François Joannès, in the most recent study of the archive's close, has nuanced the picture into something messier than a single expropriation. Following its last principal, Rīmūt-Ninurta, into his final years, Joannès traces a personal withdrawal: at the end of his sixth documented year Rīmūt-Ninurta travelled to Susa, where a lawsuit cost him "plusieurs dizaines de mines d'argent" (several dozen minas of silver), an expense Joannès links to the difficult circumstances of Darius II's accession reaching even the head of the firm; and in his last years, "en raison d'une absence ou d'une maladie" (by reason of an absence or an illness), his taxes were paid for him by a servant, in the last mention of him.^[23] On this reading the firm ends less with a seizure than with its principal's costly retreat, after which the remaining operations pass to Enlil-suppē-muḫur and into Arshama's herds. The archive's terminal document is dated 1 November 404, the accession year of Artaxerxes II and the very year Egypt broke away under Amyrtaeus and Arshama's satrapy was lost; whether that coincidence is causal, or the archive had simply run out with the firm's absorption a decade earlier, the evidence does not decide.

The Murašû tablets are legal records, and their silences are as marked as their contents. "Almost all are legal records; none are simple accounts, journals, or summaries, only three are inventories… and none are administrative memoranda."^[24] They were kept as proof that obligations had been created or discharged, against the day of a lawsuit, and they document two things above all: possession of land and payment of the charges on it. There are no letters, no narrative, no politics beyond what a date-formula or a mortgage's timing lets a historian infer. Women appear only incidentally, as a party to a contract or the holder of an estate, never in their own voice. The great events of the empire enter only where they touch a fief or a payment.

The records also point beyond themselves, to an authority the firm's own paper could not replace. Behind the archive stood the royal registry (kārammaru), in which the crown recorded primary titles, chattels and service obligations with binding force. When the ownership of a house was disputed, the decisive claim was not a bill of sale but an entry in the register: "That house is mine. It was written down for me in the king's registry in the twenty-second year of Darius (I)."^[24] Instructions on canal maintenance likewise refer sections of dike to "the registry".^[24] The archive is the record of a private business operating inside a state framework of tenure and obligation that it drew on constantly and never contained. Read for what it is, a working file weighted by discard and by editorial accident toward its latest years, it is a document of practice rather than a chronicle: precisely dated, drawn to formula, and written with no eye on posterity.