The Persepolis Fortification Archive

The archive is bureaucratic memoranda, and almost nothing else. An agency based at Persepolis managed the intake, movement, storage, and outlay of the region's staples, grain, fruit, and cress; sheep and goats, cattle, horses, mules, and poultry; the things made from them, flour, beer, wine, oil, meat, hides, over a territory reaching roughly from the Behbahān plain in the west to Neyrīz in the east. It supported, in the words the tablets keep, everything from livestock up through workers, craftsmen, administrators, travellers, priests, and courtiers, to the gods themselves, who received their offerings from the same stores by the same kind of receipt. A typical Fortification tablet is a small tongue of clay three to five centimetres across, formed around a knotted string, sealed, and recording a single transaction: so much barley issued to such a person at such a place in such a month of such a regnal year. Larger rectangular tablets, often ruled into tables, are the second-order records, monthly and yearly accounts compiled at Persepolis from the little memoranda, which were then meant to be thrown away while the summaries were kept on file. This is why the find survives at all: it is a snapshot of documents in mid-process, the raw slips and the digests together, caught before the routine housekeeping could clear them.

The archive is enormous and only partly read, and the counts have been revised upward as the recording goes on. Herzfeld's first report guessed more than thirty thousand tablets and fragments in all. Current estimates hold that the Elamite material comes from fifteen thousand or more original tablets, of which at least seven thousand are legible and analytically useful; beside them sit some 840 monolingual Aramaic tablets and 250 or more short Aramaic notes added to Elamite ones, up to five thousand tablets bearing no text at all but the impressions of seals, and more than four thousand distinct seal designs read from those impressions, one of the largest bodies of glyptic imagery from the entire ancient world.^[2] About seventy per cent of the original documents are in Elamite cuneiform, the largest single body of writing in that language anywhere; the Aramaic is one of the largest corpora of Imperial Aramaic known. And then there are the freaks that prove how many tongues met in one office: a single tablet in Greek script recording an amount of wine, the only Greek record known from the whole archive^[3], a single one in Old Persian (the only known use of that script for real accounting rather than for display), one Babylonian legal deed about the sale of a slave, which may be intrusive rather than a record of the bureau itself,^[2] two or three apparently in Demotic Egyptian, and one, uninterpreted to this day, apparently in Phrygian.

Behind the heap lies a system, and Richard Hallock, editing the texts, recovered its shape by sorting them. He divided the published tablets into categories lettered A through W. The content categories A to S fall into two halves: A to J follow commodities as they move, are transported, delivered, deposited, and set aside for provisions; K to S record the apportionment of those commodities to their ultimate consumers, the ration itself, the largest single group being the regular monthly rations for work gangs. The remaining categories are defined by form rather than content: the letters (T), the small labels or file-tags (U), the journals (V) that gather two to forty-odd entries under one heading, and the accounts (W) that tabulate a store's intake and outgo over a period.^[4] The whole apparatus runs on a paper trail of clay: an order authorises a disbursement, a clerk at the point of issue writes the memorandum, the memorandum travels to Persepolis, and there it is posted into a journal and rolled up into an account. The archive preserves every stage of that flow at once.

The agency had a head, and he has a name. Parnakka, whom the Greeks would have called Pharnaces, a kinsman of Darius, directed the bureau; under him worked a deputy, Ziššawiš, and a staff of scribes, storekeepers, and supply officers named again and again across the tablets. Parnakka's own orders survive as short letters in his voice, and the office that issued them was, in the fullest sense, a chancery. Its instrument of control was the seal and the sealed document. Nearly every tablet is sealed, and a great many transactions turn on an authorisation the texts call a halmi, a word meaning at root a seal, and by extension the sealed order or warrant that a seal validates.^[5] A commodity moved because a halmi said it should; a traveller drew rations because he carried one. The seal was the signature, and in an administration that ran on clay rather than on trust, it was the thing that made a scrap of clay an instruction.

Because almost every tablet was sealed, the archive is also, incidentally, a picture gallery. More than four thousand distinct and legible seals have been read from the impressions, by a wide margin the richest single body of figural art left by the Achaemenid world.^[2] They are not easy evidence. A seal survives only as impressions rolled or stamped across the clay, no one of them usually complete, often smudged or half-rolled, so that recovering a full design means collating many partial impressions on many tablets into one drawing, a labour begun in earnest only after 2001. The imagery ranges from the royal hero grappling with beasts, through worship before altars, to the stepped and towered structures at the centre of the long argument over Achaemenid fire-cult (see the sacred fire).

The seals also anchor the archive's people. A particular official used a particular seal, so an image on a tablet identifies the hand behind a transaction as surely as a name. Parnakka the director sealed with a fine cylinder, catalogued PFS 9, until at some point in the reign it was replaced and he began to use a second, PFS 16; his deputy Ziššawiš had his own, PFS 83.^[1] One seal used in the office, PFS 93, was an heirloom already old when it was pressed into Darius's paperwork, an inscribed piece naming Teispes, a forebear of the royal line, carried down the generations and still doing administrative work a century or more after it was cut. Mark Garrison and Margaret Cool Root, who have been publishing the corpus since 2001, treat the seals not as decoration on the documents but as a primary source in their own right, the visual language of the empire's centre caught at the moment it was being codified.

The great value of the archive is that it is not propaganda and not literature. Herodotus and the other Greek writers show Persia from the outside, late, and often hostile, fond of the palace intrigue and the inflated number; the royal inscriptions show Persia as the King wished it seen. The Fortification tablets offer neither the enemy's picture nor the official one, but the ledger, and on several points where the Greek image hardened into cliché, the ledger tells a different story (see the sources and how we know).

The sharpest correction concerns women. The Greek picture of the Persian court is the harem, the secluded, passive royal woman as ornament and property. The tablets know nothing of it. They record women, the pašap, working in the royal economy in numbers comparable to men, classified like the men by skill and seniority rather than by sex, and paid accordingly, so that a senior female overseer, the araššara pašabena of the texts, might draw larger rations than the men rationed beside her: on one tablet she takes thirty quarts of wine where her male co-workers take twenty and ten.^[6] Women who had just given birth received enhanced rations for a period, and a distinction is even drawn in the reward for bearing a son and for bearing a daughter, the mother of a boy consistently drawing twice what the mother of a girl drew.^[6] Above the workforce stand the great women of the realm as economic powers in their own right. The tablets show Irdabama, a woman of the highest rank, holding her own estate and commanding a large workforce, her own sealed orders moving commodities across the heartland.^[6] They show Irtašduna, whom the Greeks called Artystone, Darius's wife, with estates and a household of her own, her table provisioned by royal decree, and her stores and workforce moved on her own sealed order.^[6] These are not concubines; they are landholders and employers, administering property and labour and travelling to see to it. Wouter Henkelman has set the tables of Irdabama and Irtašduna beside the table of Darius himself, provisioned from the same accounts by the same clerks: the royal women are visible in the archive doing precisely what the Greek sources say Persian women could not do.^[5] The tablets are, in short, the documentary ground for portraying the women of this empire as agents rather than decoration.

The second correction is to the picture of a slave empire. The base of the Persepolis labour force was the kurtaš (Elamite kurtaš, from Old Iranian *gṛda-), dependent labourers of both sexes and various ages, organised in gangs and paid in kind, drawn from across the whole span of the empire and labelled by origin: Egyptians, Babylonians, Cappadocians, Carians, Lydians, Ionians, Thracians, Bactrians, Sogdians, and more, working side by side in the Persian homeland. They are not free wage-earners in the modern sense, and the relation is not simple; but neither are they the chattel slaves the Greek word for the King's subjects would imply. Weighing the term across the Elamite, Old Persian, and Babylonian versions of the Behistun text, Gian Pietro Basello finds "household worker" a better rendering than "slave" for the Iranian word and its Elamite loan, and observes that the genuinely servile categories the archive knows, marked by other words, are limited in number and in role, so that true chattel slavery is "not economically relevant" to the Persepolis economy at all.^[7] What the tablets show instead is a managed labour force, rationed, tracked, moved between sites, and fed by a state that treated its upkeep as an accounting problem.

The third correction is religious. It is tempting to imagine Darius's Persia as a land of Ahura Mazdā alone, and the tablets forbid it. From the very storehouses that fed the workers, the administration also paid out grain, wine, and beasts for the cult, and the gods it provisioned were many. Offerings went to Ahura Mazdā, the supreme Iranian god, and in the same breath to other Iranian divinities, and to the great Elamite gods Humban and Napiriša, and to the Babylonian Adad, and to mountains and rivers, and to divine beings whose very origin is now uncertain. In one text a priest receives grain for a ceremony, forty measures of it applied for Ahura Mazdā and forty for the deity Mišduši; in another, offerings are apportioned among Ahura Mazdā, the Elamite Humban, and libations to three rivers. The Iranian gods appear in these accounts less often than the Elamite ones, and the royal administration provisioned all of them without evident preference. The tablets also preserve the vocabulary of ceremonies otherwise lost, above all the lan offering, a regular sacrifice served by priests, the šatin among them, and the šip feast of Elamite ancestry that Henkelman has reconstructed as a procession and banquet, on occasion presided over by the chief administrator Parnakka himself.^[5] This is not the enforced Mazdā-worship of later imagination but a working, subsidised pluralism in the court's own back garden, the practical face of the tolerance the dynasty is famous for (see the Magi, who appear here too, drawing their rations like any other religious personnel).

Because the archive fed travellers as well as workers, it is also the inside evidence for the imperial road and courier system that Herodotus describes from the outside. Officials moving on the King's business carried authorisations, sealed by the governor at their point of departure, and drew rations at way-stations spaced along the main roads at intervals of about a day's journey; the vouchers for what they took were sent back to Persepolis to be reckoned. The sealed warrant was the traveller's licence to eat from the King's stores, and it was near-universal: on the travel texts that record such journeys, about eighty-six per cent name a halmi, the sealed document that entitled the bearer to his ration.^[5] The tablets thus record, from within, the same institution Herodotus praises as the angareion, the tireless mounted relay, which they name in its Elamite spelling pirradaziš, from Old Persian *fratačiš, the express courier, reckoned able to carry a message from Sardis to Susa in about twelve days.^[8][9] The Greek historian outside and the clerk inside describe the same machine, two independent witnesses to it. The destinations named in the travel texts stretch to the far edges of the empire, Sardis in the west, India, Arachosia, and Bactria in the east, so that the reach of the roads can be measured in the ration-slips of the men who walked them. One tablet has some seven hundred shepherds driving the King's sheep from Persis to Susa; a Babylonian named Bēl-eṣir shuttles documents between Persepolis and Susa; the whole apparatus of a governed distance is here in receipts.

The archive's language is itself a problem, and the source of its deepest interpretive debate. Elamite, the language of most of the tablets, has no known relatives and was already ancient in Iran when the Persians adopted it for their accounts. The empire's rulers spoke an Iranian tongue, and the world the tablets describe is largely Iranian: Iranian names, Iranian titles, Iranian technical terms. But almost all of it survives only in Elamite spelling, filtered through Elamite grammar and scribal convention, so that recovering the Iranian reality means working back through a screen. Jan Tavernier's Iranica in the Achaemenid Period (2007) is the great instrument for that work, a lexicon that gathers the Old Iranian names and loanwords buried in the Elamite and other non-Iranian sources and reconstructs their original forms.

How thick the screen is has been fought over. Ilya Gershevitch proposed in 1979 that Achaemenid Elamite was not really a working language at all but "alloglottography": Old Persian written down in Elamite signs, dictated in Persian and read back out in Persian by clerks who need not have known Elamite, the Elamite mere cipher. On that view the Persian versions of the royal inscriptions were duplicates for display, and the true medium of government was always Persian. The thesis has been largely rejected by Elamitologists, Henkelman and others showing that the administrative Elamite is a living, rule-governed language with its own grammar and idiom, not a mechanical transcription, though Gonzalo Rubio has argued in its favour.^[10] The discovery of that single Old Persian administrative tablet cut across the whole argument from the other side: it showed at least one scribe writing Persian in Persian script and expecting a colleague to read it, exactly the thing Gershevitch had said could never have been in general use.^[10] What is not in doubt is that the two registers meet in the archive in illuminating ways. The same word pirrašam, Old Persian fraša-, "wonderful," that Darius uses at Susa for his palace and at Naqš-e Rostam for all creation, turns up in a Fortification text as the name of a kind of poultry, perhaps a peacock; and the solemn Behistun phrase for the death of Cambyses, "he died in his own death," is used in one tablet for the quiet passing of an ordinary administrative functionary.^[11] The vocabulary of the King's proclamations and the vocabulary of the poultry-yard were one vocabulary, put to very different work.

Because the tablets are memoranda, they do not speak in sentences the way an inscription does, but a few of them preserve the King's own order relayed into the account. Darius commanded wine and sheep from the palace stores for his wife's household, and the clerk booked it:

200 marriš (of) wine (is) to be issued to Irtašduna the princess. It was ordered by the king.

Persepolis Fortification tablet PF 1795, trans. Hallock^[12]

More often the order in the account is the director's own. Parnakka's letters open with a fixed formula, the addressee named, then "Parnaka spoke as follows," and close with the scribe's colophon; the tablet is sealed with his own seal, and the whole small document is at once an instruction and its own authorisation:

Tell Datapparna, Parnaka spoke as follows: 6 marriš (of) wine (is) to be issued as rations to Zitrina the ullira in the treasury. In the second month, 18th year. Kamezza wrote (the text). …

Persepolis Fortification tablet PF 1789, a letter-order of Parnakka^[1]

The plural cult, likewise, survives only because a priest's draw on the stores had to be recorded:

40 (for the god) Ahuramazda (as) a gift(?), 40 for (the god) Mišduši, he utilized (it).

Persepolis Fortification tablet PF 337, trans. Hallock^[12]

These are not the sentences of a man addressing history. They are the reason the archive is trusted: the wine for the princess, the wine for a treasury delivery-man, and the grain for the god are booked in exactly the same dry hand as the fodder for the mules, and none of it was written to impress anyone. What the King wished the world to believe is at Behistun; what his government actually did, day by day, is here. The nearest the archive comes to a creed is the same prayer the King set over his palaces, that the god keep the land from the army, the famine, and the Lie, the three banes the tablets themselves are the record of holding at bay:

May Ahuramazda protect this country from a (hostile) army, from famine, from the Lie!

Darius I, DPd §3, trans. Kent^[13]

The Fortification tablets are not the only clay archive from the terrace. In 1936–1938 Erich Schmidt's excavations of the Treasury building, at the south-eastern corner of the same platform, turned up a second and smaller body of documents, some 750 tablets and fragments, all but one in Elamite, of which about 138 have been published.^[11] The Treasury tablets are a generation later than the Fortification texts, most of them dated between about 492 and 457, running from the end of Darius's reign into that of Artaxerxes I, and they record something the earlier archive barely knows: the payment of workers partly in silver rather than wholly in kind. Where the Fortification texts pay a craftsman in barley, wine, and mutton, the Treasury texts increasingly convert a share of the ration into coined or weighed silver, the first clear trace in the heartland's own paperwork of a money economy taking hold beside the older economy of stores (see the daric). The two archives together, edited by George Cameron for the Treasury and Hallock for the Fortification, give a run of the Persepolis administration across roughly half a century.

An archive this candid is still not omniscient: it is a snapshot of one region over about sixteen years, in the narrow idiom of a food-and-travel bureau, nearly silent on high politics, war, and law, and framed throughout by the Elamite screen and by the very bureau whose reach it records.

The tablets' second life has been long and is not over. After the loan of 1935 the slow work of reading them began at Chicago, and its central figure was Richard T. Hallock, whose 1969 Persepolis Fortification Tablets edited and translated 2,087 of the Elamite texts and remains the foundation of the field; he left editions of some 2,586 more in typescript at his death, the unpublished series still cited as PF-NN and now being prepared for print.^[4] From 2006 the Persepolis Fortification Archive Project at Chicago, directed by Matthew Stolper, set about recording the still-unpublished thousands with high-resolution imaging, and its findings pour steadily into print. In parallel the physical archive has been going home in instalments: three batches of tablets and a mass of fragments have been returned to the National Museum of Iran, 153 Elamite tablets and 23 uninscribed in 1948, a great body of fragments in 1951, 300 more Elamite tablets in 2004, and 1,783 in 2019, with the reading of them continuing on both sides of the world.^[2] More than 2,400 of the Elamite texts have now been published, still a minority of the whole.

The archive became, in the 2000s, the object of a bitter lawsuit. Plaintiffs holding a United States judgment against Iran sought to seize the tablets, as Iranian state property held in Chicago, to satisfy it, and for years the collection was under threat of being broken up and sold at auction. The scholars fought the claim on the ground that the tablets are an indivisible cultural whole and were only ever loaned; in Rubin v. Islamic Republic of Iran (2018) the Supreme Court of the United States decided, unanimously, that the Foreign Sovereign Immunities Act created no exception that would allow the seizure, and the archive was spared.^[14] A food-bureau's discarded receipts from the reign of Darius were argued, twenty-five centuries on, before the highest court of a republic on the other side of the earth.

In the study of the empire the Fortification Archive is the corrective conscience that sits behind a great deal else. Wherever a Greek anecdote or a royal boast is in play, the tablets are the place to test it against the empire's own books. They are the ground of confidence, and equally of caution, about how the empire actually ran.