tionaries and royal wives state that the king gave the seal to his subordinate or wife
(Mayr and Owen 2004 ); the same seems to apply to seals that mention a king of Akkad
(Zettler 1977 ). Thus the king must be regarded as the source of the message of royal gift
seals, if not directly then indirectly through his advisors. While Ur III kings are depicted
on seals of their subordinates, no personal seal of an Akkad or Ur III king is extant,
perhaps because self-deified kings considered themselves beyond the mundane sphere.
At least from the Akkad period onwards, the king must also have been the ultimate
source for images set up in temples. The circle of donors of dedicatory objects,
especially statues, was more restricted under the hegemonic kings of Akkad and
Ur III. While Early Dynastic donors of dedicatory objects include a large number of
non-royals of diverse professions who hardly ever mention an association with a ruler,
from Akkad times onwards practically all donors are either kings, are affiliated to the
crown by kin, or at least include the reigning king in their wish for a long life (Braun-
Holzinger 1991 : 18 – 21 ). Whether the images of Early Dynastic queens and princesses
were commissioned by themselves or by the crown cannot be determined with
certainty, and the few Late Uruk monuments may, in the absence of inscriptions, all be
royal.
Different media target different audiences. All objects dedicated to a deity address,
at least in rhetoric, the gods. Aside from them, foundation figurines were primarily
aimed at future kings who would rebuild a temple; stelae, statues and other dedicatory
objects set up in temples at the people at home; rock reliefs at people on the periphery
of the realm; while seals circulated within the administration and royal gift seals
targeted the royal entourage in particular.
Different image carriers provide different conditions for depiction. While statuary
is restricted to one figure, or exceptionally a pair, objects carved in relief can depict
one or more scenes of several figures. Because they provide space for extensive visual
narratives and long texts, stelae were ideal vehicles for royal self-representation, so
much so that the Akkadian equivalent and loan of the Sumerian term for stela came
to mean “memorial monument set up by a king” (CAD N/ 1 , s.v. narû). The few rock
reliefs that are known represent only one figure or a single scene, probably due to the
effort it took to carve them.
Two principles of composition are crucial for understanding Mesopotamian images:
the distinction between culminating and episodic scenes and the notion of core and
expansion (Suter 2000 : 211 – 225 ). Culminating scenes depict one group of figures at the
climax of a series of events, while episodic scenes juxtapose successive episodes of a
story; the former are allusive and lend themselves to symbolism, the latter are more
explicit and encourage concreteness. In Erwin Panofsky’s ( 1962 ) tripartite scheme of
interpretation, the former require synthesising intuition on the third layer of icono-
logical interpretation, while the latter can usually be understood with a knowledge of
literary sources on the second layer of iconographical analysis. Both can be combined
on the same monument. Depending on the available space, culminating scenes can be
reduced to their core elements or expanded into complex scenes, to which episodic
scenes can be added. The principle of core and expansion is paralleled in verbal
narratives, especially royal inscriptions.
Image and text on the same monument are not one dependent on the other, but
complement one another. In the case of the Eanatum Stela, for example, the inscription
serves the legal case of legitimacy of Lagash’s claim over Umma, while the image’s
–– Kings and queens ––