The structure of the administration of Ur III and some of its administrative initiatives also played a role in avoiding internal rebellions. Many of these features of Ur III were foreshadowed in the structure and administration of the Akkadian empire, but appear to have been much more thorough and perhaps correspondingly more successful. But these initiatives should not be misunderstood as revolutionary: there was still considerable stability in the historical organization of the city-states. When a state was taken into the Ur III empire, it became subordinate to the centre but was largely unchanged in other respects. The palace and its staff – sometimes even including its women – and the other institutions that existed in the independent state were taken over into the civil administration of the provincial government. The provincial governor might even keep the title of ensi used by the former ruler, and indeed the new governor might even be the former ruler. The stability extended also to traditional differences between northern and southern cities: in the south the imperial administration operated through the temple estates, while in the north it tended to use individuals as agents. As a consequence, when Ur III’s central administration collapsed, the cities were able to adjust to independent operation without much disruption.
There were two fundamental divisions in the administrative structure of the state, which is largely due to the reforming effort of Šulgi. The first division of note is that between the civil administration, which largely recreated or continued the traditional administration of the incorporated territories and whose bureaucracy created most of the documents that have survived, and the parallel and independent military administration, which was much more closely linked to the central power and the ruling dynasty and could serve as a check on particularist and centrifugal tendencies. The second division was that between the core of the Ur III state, consisting of the Land of Sumer and Akkad bounded by the desert, the lower sea, the limit of the land watered by the Tigris, and Šulgi’s northern wall, and the periphery, extending north-west from the core as far as Urbilum and east to the Zagros mountains.
The Inner State
Within the core area the civil administration divided the Land into about 20 provinces, each governed from a capital city by an ensi appointed by the central power. The ensi was responsible for the most part for the administration of temple estates, the maintenance of essential infrastructure such as canals, and the dispensing of justice. In theory, these governors were not secure in their tenure: they could moved from place to place in an attempt to prevent them from developing a base of power in a long-administered territory. Šu-Sin’s uncle Babati had many such postings; amongst them civilian governor of Awal and military governor of Maškan-Šarrum. In practice, however, they tended to be drawn from local power elites as were the occupants of other high offices. Ur-Nammu, for example, made Namahani’s viceroy the provincial governor of Lagaš after he had defeated that king. As a consequence, as the central power waned, and in an effort to retain the support of those local elites, it became more and more difficult for the centre to exercise the prerogative of appointment, and these positions came more and more to be retained in families.
The military administration also divided the Land into military zones, each lead by a šagin (šakkanakkum,) or general. These zones did not quite coincide with the civil provinces: the ensi of Umma, for example, had to deal with several šagins. In general the generals were not native to the area, nor were they from the great families. They seem to have been chosen, perhaps on merit, from groups that had proven their commitment to the royal house. Many of them are even observed to have names that indicate Amorite, Gutian, Akkadian, or Hurrite descent, so they seem also to have been drawn from marginal groups, perhaps because their only power base in the Land would be derived from their military service. In many cases, they were further committed to the regime by marriage ties with the royal family. These military governors received income from royal agricultural estates and properties. Unlike the ensis in practice, the šakinas remained highly mobile and could be posted anywhere in the core or periphery of Ur III.
All the provinces of the core participated in the bala (exchange) institution created by Šulgi in his year 38. Each was required to supply a certain quantity of goods to a bala fund. The kind to be supplied depended upon the province’s specialisations (Lagaš – cereals, Umma – cereals, reeds, timber, etc.,) but it is not recorded that any core provinces were required to supply livestock: that was, seemingly, a duty reserved for the periphery. The quantity depended upon calculations of potential yield made by responsible officials. These dues were in the standard case delivered to centralised collection and distribution points, though they could be delivered directly to some parties who had a standing claim. Once the province had met its obligations to the bala fund it could then withdraw from this fund as it required. The central administration typically distributed a good part of a province’s bala fund within that same province to royal dependents. Actual contributions might differ from those due, in which case the difference would be recorded and a debit or credit carried across. This sytem continued until year 9 of Ibbi-Sin.
Since ED times it had been normal for cities to send offerings and general supplies to the Ekur in Nippur and this custom was regularised in the Ur III system by assigning responsibility for the supply of contributions to various cities of the Land in a regular rotation. The responsibilities were not equally distributed: Lagaš, for example, was required to supply 2 months’ worth, while other cities were assigned one month or a half month only. This responsibility was not assigned to cities in the peripheral regions.
The Outer State
The periphery was also divided into provinces. The overall responsibility for the periphery was vested in the sukkal-mah, or royal chancellor, a court official of Ur, but the actual administration was assigned to officials posted there. In the normal arrangement, each province was governed by a general (šagina,) who was a royal appointee and liable to be temporary in the post, or by a senior captain (nu-banda.) Several of the major core cities, however, were administered by ensis; but this seems to have been merely a terminological difference, since they are seen to be responsible for military affairs in a way that the ensis of the core were not. Each province had a larger settlement that served as the main administrative centre and a number of smaller settlements commanded by a junior captain subordinate to the provincial centre. These military personnel seem, as in the core area, to have been rewarded by the grant of subsistence land holdings on which it is supposed that they raised livestock.
The contributory system of the periphery was also markedly different from the core. Here we are aware of the names of various kinds of obligation, such as gún-mada, šu-gíd, and máš-da-ri-a; but of these we only have reasonable information concerning the first. We are not sure, however, whether this means it was the most significant of the obligations.
The gún-mada (the ‘rental of the land,’) begun like the bala in Šulgi’s 38th year, was levied on all soldiers allocated land holdings outside the core area. The size of the levy depended strictly upon the soldier’s rank and was paid in the form of livestock, presumably collected from the local population. In his 38-9th year Šulgi established the city of Puzriš-Dagan (modern Drehem) to the east of Nippur as a place where those levies – especially the animals – and other levies of livestock could be collected, held, and distributed. The necessity for such a specialised location is evident from the quantities recorded: there could be as many as 200 sheep and goats and 15 head of cattle passing through each day. The officials of Puzris-Dagan possessed assessments of levies due, and deficiencies would need to be made up later. The levy was collected annually in the period September-December, or in some cases twice per year. Once delivered these gún-mada levies and the other levies mentioned formed part of the bala fund described above, and the core was able to draw upon these supplies – especially of livestock – for whatever reason, but presumably especially for temple offerings. In a similar way, the periphery was also able to draw upon the bala fund at need for supplies – especially of cereal.
All administrative texts were in Sumerian, even though there is a suspicion that Sumerian may no longer have been used in common speech, but was a merely learned language, or even a dead one. If that is the case, then the use of Sumerian was probably a further effort to emphasise the legitimacy of the rule of Ur III in the Sumerian Land. In any case, radical reforms were made to the script, in order to facilitate the record keeping of transactions between previously independent states. Something similar was attempted by Akkad, as well as the standardization of the format of records and administrative procedures. This in turn required the establishment of an extensive system of scribal schools.
Again following the lead of Akkad, Ur-Nammu standardised the weights and measures and introduced a new calendar, the reichskalender, and a new uniform system of year-names. We see this reichskalender being used in Puzriš-Dagan, Umma, Girsu, and Ešnunna in royal contexts, but local calendars were not generally displaced.
 vdMieroop, p 77.
 Steinkeller, pp. 20 ff.
 Saggs, p. 234.  Postgate, pp.151f.
 Postgate, p.152
 H&S, p. 79.
 Postgate, p. 122.
 Postgate, p.42.