The Mesopotamian temple household

What is attested

A Mesopotamian temple was a house. The Sumerian word is é, the Akkadian bītu, and the usage is not metaphorical: the building was the residence of a deity who lived in it as a great householder lived in his, with a bedroom, a table, servants, wardrobes, livestock, agricultural estates, workshops, boats, dependants and debts owed to him. The god was held to be present in a statue. The ritual texts — most of them first-millennium or Hellenistic, and prescriptive rather than reportorial — set out a routine in which that statue is washed, dressed, perfumed, fed at fixed hours, entertained with music, put to bed at night, and carried out in procession on festival days. What relation the image was thought to bear to the deity is itself disputed, and is taken up below. The temple was not a hall where a congregation gathered. Access to the interior was restricted to a defined class of personnel, and the ordinary population encountered the god mainly when he came out.

The documentation of this institution is overwhelmingly economic, and that is its great strength. From the mid-third millennium — the archives of the goddess Bau's household at Girsu in Early Dynastic Lagash — through the Ur III bureaucracy and down to the first millennium, temples generated ledgers. The two densest bodies are Neo-Babylonian: the Ebabbar of Šamaš at Sippar, with something on the order of thirty thousand tablets, concentrated in the sixth and early fifth centuries BCE, and the Eanna of Ištar at Uruk, with roughly eight to nine thousand, largely of the seventh to fifth centuries. Where an archive is held to begin and end is a modern editorial judgement, and both the tablet counts and the date ranges shift between studies accordingly. These are not liturgies. They are receipts, herding accounts, harvest ledgers, rental contracts, disbursement records, litigation minutes and personnel lists. From them we can reconstruct the temple as a working enterprise with a precision unavailable for almost any other ancient religious institution.

The temple's staff was stratified and its offices were property. At the top sat administrators — the šatammu managing the household and, under the Neo-Babylonian and Persian kings, a qīpu representing the crown's interest. Below them, a class designated ērib bīti, 'temple-enterers', held the right of access to the sanctuary itself. Around them worked the prebendaries. A prebend, Akkadian isqu, literally a 'share' or 'lot', was the legal title to perform a specified cultic service on specified days of the year — baking, brewing, butchering, pressing oil, keeping a door, singing — in return for a defined share of the offerings once they came off the god's table. Prebends were bought, sold, mortgaged, split into fractions and fractions of fractions, inherited, and fought over in court. They are the mechanism by which the cult and the urban economy were welded together, and they are documented in enormous detail. Below the prebendaries were dependants including širkū, oblates dedicated to the temple, whose labour was the temple's to direct. The marking of oblates with a star is documented for the Neo-Babylonian Eanna at Uruk, where oblate status generated a great deal of litigation; it should not be carried over as a standing feature of temple staffing everywhere.

The daily cult itself is described in ritual texts, and here the character of the evidence changes sharply. A tablet from Seleucid Uruk, TU 38 in the Louvre's Textes cunéiformes series, sets out the daily offering regime for the Anu cult: meals at fixed points in the day, main and secondary services, with quantities of bread, beer, wine, oxen, sheep and birds specified. It is precise, it is systematic, and it is a document of the late Hellenistic period whose own colophon claims that its rituals were recovered from tablets a scribe named Kidin-Anu found in Elam, where they had been taken generations earlier. That colophon is a legitimation device, and it should be read as a warning label attached to the whole genre: these are compilers' texts, asserting authority through claimed antiquity.

A newly made image was inducted into cult through a ritual, the mīs pî and pīt pî, the 'washing' and 'opening' of the mouth, after which it was treated as the god's presence. First-millennium tablets from Nineveh and Babylon preserve two ritual recensions, a Nineveh ritual and a Babylonian one, together with a set of Sumerian incantations whose arrangement across six or eight tablets is a modern reconstruction rather than something the tablets state. The rite moves the newly made image in stages over two days: out of the craftsmen's workshop, to the riverbank, into a reed enclosure in an orchard, then to the temple gate and finally into its niche, accompanied by invocations of the great gods, the patron gods of craftsmen, and astral powers. Its declared logic is that the image was born in heaven and not made by human hands; the tools of the craftsmen who made it are symbolically cut away and the craftsmen deny their own work. What survives of these statues themselves is nothing. They were wooden-cored, plated in precious metal, jewelled and clothed, and they were therefore looted, melted, or carried off as war trophies — a deportation of gods that Assyrian and Babylonian royal inscriptions record repeatedly.

How the rite worked

The temple day was structured around meals. The sanctuary was opened, a table set before the image, a curtain drawn around it while the god ate, and the food removed when the meal was over. Late texts describe a regime of two principal and two secondary meals daily. The god did not consume the food; the offerings, having been in his presence, acquired value and were redistributed — a portion to the king, portions to the priesthood, and portions to the prebendaries whose right to that share was precisely the content of their prebend. The Akkadian term sattukku covers the regular fixed offerings, ginû the recurrent income assigned to sustain them. This is the pivot on which the whole household turned: the god's table was simultaneously an act of worship and the payroll of the institution.

Prebendary service ran on a calendar. The temple year was divided among prebend-holders by day, and because prebends were inherited and subdivided, individual holdings could be fractions of a day's service in a given month. Documents record prebends changing hands for silver, being pledged against debt, being partitioned among heirs, and generating litigation when the same days were claimed twice. Prices are attested well enough that scholars have studied prebend markets as markets. A prebend was, in practice, a security whose yield was paid in bread, beer and meat that had passed the god's table.

The estate behind all this was large. Temples held arable land worked by tenants and dependants, date orchards, herds contracted out to shepherds who were held liable for shortfalls at annual reckoning, fishing rights, boats, and weaving and metalworking shops. They lent, they collected, they employed, they prosecuted. In the Neo-Babylonian and Achaemenid periods the crown sat over this apparatus through the qīpu and through royal claims on temple income, which is one reason the archives break off where they do — administrative reorganisation, not piety, ends most of them.

Festivals moved the god out of the house. Processions carried statues along prescribed routes, divine statues were carried by river barge to visit the gods of other cities, and the akītu festival, best known from Babylon, involved the god leaving the temple for a festival house outside the city and returning. The ritual texts for the Babylonian akītu survive in fragmentary late copies, and the celebrated elements — the king's humiliation before the god, the recitation of Enuma Elish — come from those late and incomplete tablets, not from a continuous record of performance. What the ceremony looked like in the second millennium, or in any specific year of the first, is not documented.