Urartian Geography and Sargon's Eighth Campaign

by Paul Zimansky
Urartian Geography and Sargon's Eighth Campaign
Paul Zimansky
Journal of Near Eastern Studie
Start Page: 
End Page: 
Select license: 
Select License



PAUL ZIMANSKY, Boston University

THE diversity of geographical reconstructions inspired by the account of Sargon's eighth campaign (714 B.c.)' is a tribute to Assyriological ingenuity and Assyrian obscurity. At issue is more than simply ancient toponymy or the details of a single historical event, but the essence of Assyrian military capacity, the configuration of Urartian settlement and government, and the compatibility of Assyrian histori- ography and archaeology as mutual avenues to interpretation of the past. Seven and a half decades of isolated contributions by more than a dozen scholars have generated a substantial and conflicting literature on the question of where Sargon went on this most elaborately described of ancient military expeditions, and each new discussion of the text, until quite recently, operated with its own assumptions, posited a different route of march, failed to rebut arguments made in previous articles, and reinforced the suspicion that little was known for certain on the problem.

Within the last twenty years, however, several new and largely healthy developments have contributed to the discussion. First of all, the general area in question-eastern Anatolia, the Zagros, and northwest Iran-has become much more familiar to scholars and no longer harbors the large pockets of archaeological terra incognita that it did at the beginning of the century. Between 1968 and 1978 archaeological work in Iran vastly increased the quantity of data bearing on Urartu's character as both a political entity and a cultural assemblage. Surveys, particularly those published annually by Wolfram Kleiss in Archaologische Mitteilungen aus Iran, have modified previously accepted views on the extent of Urartian territorial holdings. When coupled with the work of Charles Burney in Turkey, they provide a fair picture of where the major Urartian sites were and the topography of fortification in the state as a whole. While doubtless many Urartian sites remain to be discovered, it is hardly conceivable that one the size of Toprakkale, Bastam, Cavu~tepe, or Karmir Blur has escaped notice. In addition, controlled excavation at both Urartian and non-Urartian sites has made it possible to be fairly specific about the character of Urartian material assemblages.

1 The primary text is in four columns, consisting la huitiPme campagne de Sargon, Textes cunkiforms of 430 lines, and is a "letter" to the god Assur du Louvre, vol. 3 (Paris, 1912) (hereafter abbrevi- published by F. Thureau-Dangin, Une Relation de ated TCL 3), with supplementary fragments later

appearing in B. Meissner, "Die Eroberung der Stadt

Ulhu auf Sargons 8. Feldzug," ZA 34 (1922): 113- [JNES 49 no. 1 (1990)l 22; and Otto Schroeder, Keilschrifttexte aus Assur 8 1990 by The University of Chicago. historischen Inhalts, II, Wissenschaftliche Veroffent- All rights reserved. lichung der deutschen Orient-Gesellschaft, vol. 37

0022-2968/90/4901-0001$1.00 . (Leipzig, 1922), no. 141.

Secondly, recent Assyriological literature manifests a tendency to go beyond the mere literal readings that were the mainstay of early interpretations-searching for specific correlations between textual microdescription and individual features in the ancient environment2-toward a greater appreciation of the intellectual conceptions and motivations of the framers of the text^.^ This has entailed more sophisticated means of reconciling textual and archaeological information, both with their inherent biases, and looking at larger structures of the narrative in order to evaluate what is and is not to be sought in physical remains. As one of the most poetic and politically involved Assyrian texts, the Eighth Campaign account is in some ways one of the most suspect and in other ways felt to be trust~orth~.~

These developments coincide with, but do not necessarily support, a trend toward a consensus on the general configuration of the route of Sargon's march in the most recently published hypotheses. L. ~evine,' W. ~a~er,~ us car el la,^ and M.

0. W. salvini8 have all opted for an itinerary that would take Sargon northward along the western shore of Lake Urmia and then move inland before making the southward turn to return to Assyria. For reasons that will be presented below, this reconstruction runs at some variance to both the texts and archaeology and has been erected against a straw opponent rather than the most likely alternative. The present article is offered to give pause to those who might presume that the question of the route has been resolved.

At the risk of reiterating material that is covered in virtually every discussion of the Eighth Campaign, let us begin with the ground rules. The basic text,9 formulated as Sargon's first-person letter to the god Assur, stipulates that the campaign began at Calah in the early summer. Sargon crossed both the Upper and Lower Zab and entered Mannean territory. There he reviewed his troops and moved about receiving deputations from Assyria's subordinates before setting off in pursuit of Metatti of

2 Perhaps the most extreme example of a literal approach is Edwin M. Wright, "The Eighth Cam- paign of Sargon I1 of Assyria (714 B.C.)," JNES 2 (1943): 173-86.

3 A. Leo Oppenheim, "The City of Assur in 714 B.C.," JNES 19 (1960): 133-47, offers an early and quite celebrated study of the Eighth Campaign ac- count in this regard.

4 Oppenheim, ibid., p. 135, notes: "In any attempt to utilize our text for a reconstruction of the histori- cal geography of the region it will have to be borne in mind that the sequence of events as given cannot be fully relied upon. On the other hand, Urartu was for Assyria no fabulous faraway country, so that the poet's license must have been rather restricted." Walter Mayer, "Sargons Feldzug gegen Urartu-714

v. Chr.: Eine milit2rhistorische Wiirdigung," MDOG 112 (1980): 14-16, argues that a higher standard of historical accuracy is to be expected in this text than in other royal inscriptions because the letter was addressed to the god Assur.

5 L. D. Levine, "Sargon's Eighth Campaign," in

L. D. Levine and T. C. Young, eds., Mountains and Lowlands: Essays in the Archaeology of Greater Mesopotamia, Bibliotheca Mesopotamica, vol. 7

(Malibu, 1977), pp. 135-51. K. Kessler, "Zu den Beziehungen zwischen Urartu und Mesopotamien," in V. Haas, ed., Das Reich Urartu: Ein altortien- talischer Staat im 1. Jahrtausend v. Chr., Xenia 17 (Konstanz, 1986), p. 68, credits Levine's article with providing the stimulus for revising older notions of how Sargon's itinerary is to interpreted.

6 Walter Mayer, "Sargons Feldzug gegen Urartu," pp. 13-33.

7 Oscar White Muscarella, "The Location of Ulhu and UiSe in Sargon 11's Eighth Campaign, 714 B.C.," Journal of Field Archaeology 13 (1986): 465-75.

8 In Paolo Emilio Pecorella and Mirjo Salvini,

Tra lo Zagros e I'Urmia: Recherche storiche ed archeologiche nell'Azerbaigian iraniano, Incunabula Graeca, vol. 78 (Rome, 1984), pp. 35-51.

9 Besides TCL 3, there is another substantial account of the campaign in Sargon's annals. The two versions do not differ from each other signifi- cantly, and TCL 3, in addition to being more detailed, is presumed to have been composed closer to the events it narrates because it fails to note the death of Rusa I, which occurred shortly after 714


Zikirtu, an ally of the Urartian king Rusa I." This chase culminated in a confronta- tion on the slopes of Mount UauS between Sargon and Rusa, who had come to the aid of Metatti. Sargon was victorious, changed the direction of his march, and launched an invasion of Urartu itself. Subsequently, the Assyrian army moved through six provinces (nag2) under Urartian control, of which the first had recently been added to the kingdom and the fifth lay beside a "sea." The last factor limits the range of possible itineraries for those who would construe the narrative literally, since it is generally agreed there are only two possibilities for this body of water-Lake Van and Lake Urmia. Otherwise, there are no references to sites in Urartu whose location is definitely known, such as Mount Ararat or the Urartian capital at TuSpa. Upon leaving Urartu, Sargon approached HubuSkia and received tribute from its ruler. He then parted company with the main body of his army, which was sent on the direct route to Assyria, and set off with a small picked force over the mountains to sack the temple and palace of Musasir, crossing the Upper Zab on his way. Musasir is generally felt to lie near Sidekan, about halfway between Calah and the southern shore of Lake ~rmia." The whole campaign was apparently completed before the onset of winter, which would have closed the mountain passes, and consisted of a single circuit, since there is no mention of doubling back at any point.

10 TCL 3 regularly records this king's name as "Ursa," but other Assyrian texts and all Urartian texts refer to him as "Rusa." The discrepancy is not so much due to an inexact correspondence between the Assyrian and Urartian liquids, as it is to a divergence between Urartian orthography and pro- nunciation. With one dubious exception, this is the only word in the Urartian corpus that has r in a word-initial position. To the Urartians it was prob- ably both foreign and unpronounceable without modification. Assyrian scribes, in recording the name of Sargon's adversary, vacillated between what they heard and what they read in Urartian texts, which are consistent in writing the name Rusa. I am unaware of any place in which this non- Urartian name has been used as an argument for a peripheral origin of Rusa, but this might be a counter-argument to my contention (see below) that his paternal house must be within Urartu proper. In support of my view, however, is the fact that the paternal home of Sarduri was located nearby, and there is no question of the Urartian character of that name. Translitertion is also a problem for modern students of Urartian geography, since dif- ferent conventions are normally used for each of the countries involved. In this article 1 have adhered to principles outlined in my Ecology and Empire: The Structure of the Urartian State, SAOC 41 (Chicago, 1985), pp. xiii-xiv: place names in Turkey are in Turkish; Persian and Arabic names given as they appear in the listings of the United States Board of Geographical Names; and for Assyrian toponyms I follow Simo Parpola, Neo-Assyrian Toponyms, Alter Orient und Altes Testament, vol. 6 (Neukirchen-Vluyn, 1970).

1' The approximate location of this site has never been in doubt, since it is mentioned in Urartian- Assyrian bilingual inscriptions at Kelishin and Top- zawa which were clearly erected in or near its territory. A more precise identification, based on both archaeological soundings and the apparent survival of the ancient name into the present era as Mudjesir, is offered by R. M. Boehmer, "Zur Lage von Musasir," Baghdader Mitteilungen 6 (1973): 31-40; and R. M. Boehmer and H. Fenner, "For- schungen in und um Mudjesir (Irakisch-Kurdistan)," Archaologischer Anzeiger 88 (1973): 479-521. The evidence for this identification is, however, far from overwhelming, resting largely on: (1) the presence of insecurely dated Iron Age sherds, fortifications, and architecture of more than village scale at the site; (2) the location of the site in what is generally agreed, on the basis of findspots of UrartianlAkkadian bilingual stelae, to be Musasir's territory; (3) the similarity of ancient and modern names; and (4) the topography of the area which accords with Boeh- mer's view of an attack by Sargon from the west; it has some correlations with the celebrated relief of Sargon's sack of the temple of Musasir. The idea of Sargon approaching from the west is quite hypo- thetical, resting on the assumption, nowhere sup- ported in the text, that he was coming from Habhu, the precise location of which is unknown in any event. The overall strength of these arguments would be easier to assess if the size of Musasir's territory were better known. If it was small, they are quite convincing, but if it extended, for example, as far to the north as Yuksekova or Ba~kale, there would doubtless be other candidates for Musasir.

Hypotheses on the route fall into three categories: (1) those that have Sargon go north of both Lake Urmia and Lake Van, marching east of the former and west of the latter; (2) those involving a circuit of Lake Urmia only; and (3) those that see him approach the southern and western shores of Lake Urmia and return to Assyria without circling either lake. In general, the shifting fashion of modern interpretation has moved through the alternatives in the order they are listed here.

F. Thureau-Dangin, the first to publish a map of the campaign, placed the decisive battle of Mt. UauS on the modern Sahend, linked the reference to the "sea" with the north shore of Lake Van, and saw the return to Assyria proceeding via Bitlis, which he identified with the important Urartian frontier fortress of uajais.12 Although his account has served as the point of departure for virtually all subsequent scholarship, there has been a general recognition that it is unrealistically long, particularly in its strategically irrelevant excursion from Bitlis to Musasir.13

The revisions, which began almost immediately after Thureau-Dangin wrote, favored the second category. C. F. Lehmann-Haupt, who knew this part of the world well, having traversed much of it on horseback, accepted the initial part of his itinerary but rejected the identification of Bitlis with Uajais on the grounds that there could have been no conceivable frontier at that location and that no road suitable for an army connected that area with the south. He suggested that the second half of the campaign would be better explained as a passage along the east shore of Lake Van, past Ho~ap and Ba~kale to Urmia, and a descent upon Musasir from the Kelishin pass.I4 He recognized that this last step would be no small feat for an armed force but that it would be consistent with the spirit of the text. He identified Uajais with Ba$kale and, although unaware of the archaeological remains at Mudjesir, located Musasir at Topzawa, in the same vicinity, on the evidence of the famous bilingual stele and some unexcavated ruins found nearby.'' At the time Lehmann-Haupt wrote, the extent of Urartian territorial holdings in Iranian territory was unsuspected, and his suggestions for Sargon's route between Van and Topzawa attracted little notice.

12 TCL 3, pp. ii-xx. Ecology and Empire, in JAOS 108 (1988): 164 where 13 Muscarella, "Location of Ulhu and UiSe," he commends me for not taking sides in the debate;

p. 466, n. 3, states that Burney, van Loon, and Kleiss and his review of Pecorella and Salvini, Tra lo "continue to support Thureau-Dangin's long route Zagros e I'Urmia, in BiOr 44 (1987): 259 where he around the east shore of Lake Urmia to Van." This offers comments moderately critical of the short should not be taken to mean that they all adhere to route but stops short of ruling it out. Only Charles his hypothesis of a trip around Lake Van to Bitlis, Burney and David Marshall Lang, The Peoples of however, from which most of the excessive length in the Hills: Ancient Ararat and Caucasus (London, his route derives. Kleiss, "Alte Wege in West-Iran," 1971), pp. 155-56, explicitly opt for the full Thureau- AM1 10 (1977): 137-41, only makes the case for a Dangin itinerary, albeit with reservations about the trip up the east side of Lake Urmia and says length. nothing whatever about Van. M. van Loon, "The 14 C. F. Lehmann-Haupt, "Musasir und der achte Inscription of ISpuini and Meinua at Qalatgah," Feldzug Sargons 11. (714 v. Chr.)," Mitteilungen der JNES 34 (1975): 205-7, does assert that Sargon Vorderasiatisch-Aegyptischen Gesellschaft 21 (1 9 16): went to Van but came back via Ushnu and Kelishin. 137-40; and idem, Armenien einst und jetzt, vol. 2 In more recent reviews of various publications, (Berlin and Leipzig, 1926), pp. 310, 325. however, van Loon has shown some ambivalence on 15 Idem, Armenien einst undjetzt, vol. 2, pp. 299- the subject of the route. See his review of my 301, 317.

More recent studies have offered different suggestions for a route around Lake Urmia only. Some scholars such as M. van ~oon,'~

consider an approach to Lake Van as Lehmann Haupt did, while others, such as A. A. ~ilin~iro~lu'~

and J. E. ~eade," have argued that the only lake involved is Urmia. The third choice was first suggested by H. A. Rigg, who principally directed his arguments against the Thureau-Dangin itinerary and made no reference to the modifications suggested by ~ehmann-~au~t.'~ His most firmly voiced objection was that the distances involved in other routes were simply too great:

To cover 1500 miles (flat measurement) over some of the most mountainous country in the world; against almost constant fighting, or the menace of it; with days consumed in celebrations, receipt of tribute, and diplomatic gestures; with the necessity of reorganizing conquered territories, at least to secure them from immediate revolt (at which the Urartians were masters); and the meticulously detailed sack of Musasir-plainly is impossible.20

Apparently independently, he rejected the equation of Bitlis with Uajais for the same reasons as Lehmann-Haupt had." He also argued that if Sargon had been the first Assyrian king to go north of both lakes he surely would have said so and that Sargon's inscriptions make no claim that this campaign involved deep penetration of Urartu or indeed that such was even its objective.

Despite his precedence in positing a campaign that was confined to the Zagros, Rigg's work has been largely ignored by those who have adopted the same position in recent years.22 Part of the reason for his obscurity is doubtless that he was no longer active in the field after the outbreak of the Second World War and that his most specific reconstruction is presented only in his unpublished Harvard University doc- toral dissertation.

It is probably Levine's study which has most effectively shaped the discussion in recent years, offering an itinerary that many have accepted or modified only in specific places. His arguments against Thureau-Dangin's proposal and the idea that Lake Van is mentioned, for which his scheme is presented as an alternative, are: (1) if Sargon had encountered both lakes he would have said so, (2) the capital at TuSpa on Lake Van is not mentioned, (3) to move so far north and west would mean that more territory was being covered with fewer place names than in the earlier part of the campaign, and (4) Sargon would not have marched through enemy territory directly without covering his flanks. He also adds that place names beginning with an initial

16 van Loon, "The Inscription of lipuini," pp. 205- represent his hypothesis in my fig. 1. 7, but see also n. 13 above. 20 Rigg, "Sargon's Eighth Military Campaign,"

17 A. A. Cilingiroglu, "The Eighth Campaign of p. 132. He did, however, allow that a circum-Urmia Sargon 11," Jahrbuch fiir kleinasiatische Forschung route would be less objectionable than one that (Anadolu Ara~trrrnalarr) 4-5 (1976-77): 262-65, went around both lakes.

271. 21 Ibid., p. 136. 18 J. E. Reade, "Kassites and Assyrians in Iran," 22 Levine, Mayer, and Muscarella do not mention Iran 16 (1978): 140-41. Rigg at all, and Muscarella even claims he himself

19 H. A. Rigg, "Sargon's 'Eighth Military Cam- first proposed the short route in 1971 ("Location of paign'," JAOS 62 (1942): 130-38; and idem, "The Ulhu and UiSe," p. 466, n. 3). Salvini, in Tra lo Kingdom of Van (Urartu)" (Ph.D. diss., Harvard Zagros e l'lirrnia, does acknowledge that Rigg University, 1936). No specific reconstruction of the proposed a short itinerary, "che nei punti essenziali itinerary is given in the article, but the dissertation preludeva a quello che oggi ci appare come il piu included a fold-out map, which I have used to verosimile" (p. 48, n. 212) but does not elaborate.

FIG1.-Previous reconstructions of Sargon's route

Ui-and Ua-seem to be confined to the southern shore of Lake Urmia, and the province of Ajadi, which has such associations, would be located there in his reconstruction. His thesis that there was one, and only one, Parsua, located in the southernmost part of the Zagros known to the Assyrians, draws his itinerary for the initial part of the campaign southward, so that even with the greatly restricted Urartian segment, it is nearly as long as Thureau-Dangin's. His conclusion that "any proposal that would lengthen the route of march would seem to be ruled out on technical grounds alonewz3 is an implicit counter-argument against the circum-Urmia proposal as well as the Thureau-Dangin route.

Walter Mayer is also much concerned with the length of the itinerary, and he offers various suggestions to shorten it, such as having the army already well into the field

23 Levine, "Sargon's Eighth Campaign," p. 148.

before Sargon starts out and allowing that the text includes excursions of various units that were going on simultaneously with the progress of the body of the army. The map that he presents also shows a somewhat more streamlined alignment of place names along the generally south to north line of march, justified on the grounds that Sargon probably followed the Sanandaj-Sakkez-Mahabad road on his way north.24

Two key site identifications figure prominently in the interpretations of Salvini and Muscarella: UasilUajais with Qal'eh Ismail Aqa, first suggested by Salvini, and Ulhu with Qalatgah, proposed by Muscarella. I shall treat these in detail below, for they have considerable bearing on my own theories. Let it suffice to say here that neither of these scholars offers a direct argument against the route around the lake, although Salvini states firmly that Thureau-Dangin's identification of the Sahend with UauS is un~ustainable,~~

and if Muscarella's identification of Qalatgah is correct, a northerly route would be incompatible with the sequence of events given by Sargon's letter.

Besides the question of these specific site identifications, two other considerations seem problematic in the sub-Urmia routes that have been proposed. First is the question of where Urartu was and was not; much of the emphasis in the new proposals seems at variance with the archaeological evidence. Second, there is a certain struc- tural pattern to the account of the march through Urartu that makes little sense unless the entry point is farther away from the exit. Before the idea of a route limited to the area south and west of Lake Urmia becomes an accepted historical fact, I believe a reconsideration of the more northern approach is in order.

Sargon's account shows sensitivity to a distinction between territory that is truly "Urartian" and territory which is merely under Rusa's political control. For example, the letter states that UiSdiS was a Mannean province which Rusa had expropriated.2h Only after his march through it, upon entering the next province, does Sargon claim to have crossed the border into ~rartu.~~ land so

On what basis is this "Urartian" defined? If mere conquest and annexation were sufficient criteria, then UiSdiS would be Urartian and not Mannean. But the text makes it clear that UiSdiS remained non-Urartian despite its subjugation by Rusa. Some less ephemeral association and deeper cultural unity must be at issue. Archaeological evidence and the findspots of non-portable cultural indicators, which are of limited utility in tracing transient political boundaries of the Iron Age, have an obvious contribution to make in identifying deeper affiliations.

The "Urartian" cultural complex in this sense need not be something that permeated all levels of society, although not nearly enough research has been done on this issue.28 It is certainly unlikely that there was any ethno-linguistic unity in the kingdom of

24 Mayer, "Sargons Feldzug gegen Urartu," pp. 15, project of the Urmia plain, which unfortunately had


little time to do fieldwork before the cessation of

25 Pecorella and Salvini, Tra lo Zagros e I'Urmia, archaeological work brought on by the Iranian

p. 48. Revolution. For an outline of the issues involved, 26 TCL 3, 1. 163. see Pecorella and Salvini, "Researches between the 27 Ibid., 1. 167. Zagros Mountains and Urmia Lake," Persica 10

28 The only serious attempt to attack the problem (1982): 17-20. of fortress hinterland relationships was the Italian

Urartu, which expanded quickly, by military force, over a considerable area of great regional diversity. However, strong uniformities in material assemblages were disseminated by those processes associated with the activities of the apex of the state- fortress-building, maintenance of the cult, and royal administrative activities. The distribution of non-portable and semi-portable inscriptions is but one indicator of this and can be supplemented by much additional evidence. For example, one finds mud bricks of a rather peculiar large, flat type in nearly identical sizes at widely separated sites that are clearly linked to the Urartian state by inscriptions.29 The distinctive, highly burnished red pottery dubbed "Toprakkale Ware" by Charles Burney is a fairly specific, although not particularly abundant, indicator of the presence of the Urartian state, being largely confined to citadels and more common in the central areas than on the periphery of the state.30 The techniques of fortification that are normally associ- ated with sites built by Urartian rulers, including specific patterns of buttresses on fortification walls, and the practice of carving footings in bedrock, have been outlined by ~leiss." There are distinctive building forms, and in most of the areas that the Urartians occupied the earliest datable irrigation works are assigned to them. Even the whole "settlement" pattern of fortifying rock spurs and depopulating older tell sites in valleys has been associated with the conquests of the ~rartians.~~

This "imperial culture" is unmistakable and is almost certainly the criterion that Sargon's scribes had in mind in the conceptual dichotomy between the land of Urartu on the one hand and Mannean provinces that had been added to the Urartian state on the other. From this a delineation of the core territory of Urartu may be offered with some confidence, and it does not favor the sub-Urmia routes.

After the kingdom of Urartu coalesced in the second half of the ninth century B.c.. its expansion may be traced by the location of its monumental inscriptions and fortresses. The various redactions of Sarduri 1's single text are all found in Van. Iipuini, his successor, left inscriptions that may be divided into two groups: those bearing his name alone and those in which his name appears along with that of his successor, Menua. The former all come from Van and nearby Patnos, whereas the latter category includes the Kelishin bilingual and the Qalatgah text, found not far from the southwestern shore of Lake Urmia (see below). Under Menua, who must have died more than half a century before Sargon's campaign,33 Urartu's territory

29 Thomas B. Forbes, Urarlian Archilecture, BAR graphical units that are not coincident with associa- International Series 170 (Oxford, 1983), pp. 137-38. tion or dissociation with the Urartian state. See Forbes notes that Bastam and Arin-'oerd are aaom- idem, "Urartaische Keramik," in PI.-J. Kellner, ed., alous in their use of non-standard bricks, but I Urartu: Ein wiederentdeckter Rivale Assyriens, recollect from my own work on the Hallenbau at Katalog der Ausstellung, Prahistorische Staatssamm- Bastam that bricks of 52 cm x 37 cm (height not lung Miinchen (Munich, 1976), p. 62. preserved), i.e., within the "standard" Urartian 31 W. Kleiss, "Urartaische Architektur," in H.-J. parameters, were in evidence there. Kellner, ed., Urarlu: Ein wiederentdeckler Rivale,

30 Charles Burney, "A First Season of Excavations pp. 28-44. See also Forbes, Urartian Architeclure. at the Urartian Citadel of Kayal~dere," Anatolian 32 Kroll, Keramik urartaischer Festungen in Iran, Studies 16 (1966): 85. Stephan Kroll, "Archaolo- Archaologlische Mitteilungen aus Iran Erganzungs- gische Fundplatze in iranische-ost-Azarbaidjan," band 2 (Berlin, 1976), pp. 173-74. AM1 17 (1984): 16, does point out that this pottery is 33 There are no synchronisms with Assyrian rec- sometimes found outside of Urartu, but its general ords for Menua, but his successor, ArgiSti 1, waged a association with the state is assured. On the other campaign against Assur-nirari V (754-745). To judge hand, the more common types of pottery used in by the length of his annals, ArgiSti had a long reign Urartu vary according to chronological and geo-but was clearly no longer the king of Urartu in 743,

expanded dramatically to the banks of the Araxes, down the Murat to the vicinity of Malatya, and northwest to the area of Erzurum. Thereafter the territorial acquisitions of the state were minor, with the exception of the expansion into what is now Soviet Armenia.

In the area around Lake Urmia in particular, Urartian conquests took place very early in the kingdom's history, and then apparently ceased. The inscriptions of Menua are found farther to the southeast than those of any subsequent Urartian king. Thus by the time of Sargon's eighth campaign, there must have been a relatively sharp distinction between territory which had long had Urartian buildings, Urartian building inscriptions, and Urartian pottery, and territory that was never to manifest these attributes in any abundance.34

Over what area would this prolonged Urartian occupation and integration into the state have extended? As a rule, Urartian settlement and building activity focused on expanses of alluvial land with the primary establishment of state control in the form of fortresses on elevated ground around their margins. Although the pattern of settle- ment and fortification varies somewhat in each of these nodes, major establishments are not found isolated from them. Of the main areas of irrigable alluvium, no one today would question that the plains of Marand, Khoy, Shahpur, Urmia, and Ushnu- Solduz were culturally Urartian in this imperial sense.35 It is also beyond dispute that the whole east shore of Lake Urmia was not.36

The key question for the present discussion is the status of the open area dotted with marshes and crisscrossed by small streams at the southeast end of the lake, of which Miandow-Ab is the primary modern settlement. There is some evidence of an Urartian presence here, most specifically in the inscription carved into a rock face at TaStepe by ~enua,~~ TaStepe is a very minor

which records the conquest of the city of ~eSta.~*

when Tiglath-Pileser 111 defeated his son, Sarduri 11. Thus the reign of Menua is generally assigned to the first quarter of the eighth century.

34Cf. 1. M, Diakonoff and I. N. Medveskaja, "Urartskoe gosudarstvo v novom osveSEenii," Vestnik Drevenej Islorii, 1987, no. 3, p. 205, where it is stated that political control may be effected by preserving local governments and cultures and it is argued that Hasanlu IV was destroyed by Sargon, not by the Urartians. The reference given in that article, I. N. Medvedskaja, "Hasanlu IV i nekotorye voprosy istorii Urartu IX-VIII vv. do n. 6," Doklad nu XXXI Meidunarodnoj konferencii assyriologov,

(Leningrad, 1984) is unavailable to me.

35 Kroll, "Archaologische Fundplatze," p. 128, ex- plains the absence of campaign inscriptions in the northern part of West Azerbaijan as a reflection of their incorporation into the state prior to the devel- opment of a tradition of this kind of writing.

36 W. Kleiss and H. Hauptmann, Topographische Karte von Urarru, AM1 Erganzungsband 3 (Berlin, 1976); sheet 1 shows virtually nothing on the east side of Lake Urmia, and Kroll's subsequent survey ("Archaologische Fundplatze," pp. 13-133) should have detected any Urartian sites there had they existed.

37 For the most recent edition of this text, see
Pecorella and Salvini, Tra lo Zagros e I'Urmia,
pp. 65-69.

38 Salvini has argued that Me5ta is Hasanlu and that the destruction of Level 1V at that site may be equated with this campaign. See Salvini, "Die urartaischen schriftlichen Quellen aus Iranisch-Azerbaidjan . . . ,"in Akten des VII. Inrernarionalen Kongresses fiir iranische Kunsl und Archaologie, Miinchen 7-10 September 1976, AM1 Erganzungs- band 6 (Berlin, 1979), p. 177; and Pecorella and Salvini, "Researches," pp. 10-11. I see no reason to locate MeSta so far to the west, particularly when, as Kroll has noted (Keramik urartaischer Fesrungen in Iran, p. 99.), there is another large fortress that was apparently destroyed by the Urartians at Aslan Qal'eh. The latter is much nearer to TaStepe than Hasanlu is and not separated from it by any geo- graphical barriers, so it would seem a more obvious referent for the inscription. MeSta was in Mannean territory, and in the ninth century the land of Gilzanu is generally felt to lie on the southwestern shore of Lake Urmia, perhaps including Hasanlu. There thus is no inconsistency of terminology if MeSta lay to the

site, however, and there are indeed no large Urartian fortresses or other indications of extensive Urartian governmental involvement with this area, as there are in Ushnu and western Solduz. Certainly there is no major administrative center of the sort that regularly appears in Urartian-controlled areas on the western and northern sides of the lake. For this reason, it seems unlikely that the Miandow-Ab plain would be the Urartian frontier as described in the Eighth Campaign letter, its entrance guarded by a castle on a mountain that shuts it off like a door. It is, rather, the kind of open area in which the Urartian state, with its emphasis on fortification and exploitation of small areas of alluvial land, never seems to have thrived. It was certainly conquered by Menua but not fully integrated into the state.

If the plains on the southeast shore of Lake Urmia are peripheral to Urartu, then the itineraries of Levine, Mayer, and Muscarella are flawed by having two of the five provinces specifically designated as "Urartian" outside of the Urartian cultural area. For Muscarella, there is also the complication that one of those provinces would be ArmarilP, the most central in the way the narrative of the march through Urartu is framed in Sargon's letter and the site of the paternal home of the Urartian kings Rusa and ~arduri.~~


Another objection to the itineraries that have Sargon march northward along the west shore of Lake Urmia and then turn south to return to Assyria is that they are impossible to reconcile with directional concepts that are reasonably transparent in the structure of the text. These may not, in fact, correspond exactly with the way we would describe the course of the campaign were we able to trace it precisely on a modern map, but they should not be utterly disregarded. Sargon's account represents the campaign as initially directed against Andia and Zikirtu. That portion of the campaign, treated in no great detail, culminates in the battle on Mt. UauS. Then there is an explicit statement that the campaign changed direction and turned toward Urartu. The course of the march through Urartu is represented as linear, moving from one side of the kingdom to the other. Sargon's account states that he entered Urartu proper at USqaja, which stood at the "head" of its frontier (r2i misri) (1. 167) and left it at Uajais, at the "foot" (ippit misri) (1. 298). This opposition is incompatible with reconstructions which portray Sargon's route as a loop through a small corner of the kingdom, where the army turns around and recrosses the border not far from where it entered.

How would an Assyrian construe movement from "head" to "foot" geographically? Was there a consistent mental map on which these terms had referents to topography or compass directions? The movement from r2u to S2pu is probably best- conceptualized as from "top" to "bottom" or "up" to "down." "Upper" and "lower" are terms often used to describe opposite, never adjacent, boundaries of fields in

east of Gilzanu-otherwise one would have to assume van Loon in his review of Pecorella and Salvini, Tra that Gilzanu and Mana were different names for the lo Zagros e I'Vrmia in BiOr 44 (1987): 259, who says same place. these should be nearer to Van than the Ushnu valley.

39 TCL 3,1.277. This point has also been made by

southern Mesopotamia, and in those cases where compass directions are associated with them there is a tendency, albeit slight, to favor north as "upper" over other cardinal points.40 Probably more relevant is physical elevation, which the Hittites presumably had in mind in distinguishing upper and lower lands in their empire. "Upper" also has the connotation of "upstream," as reflected, for example, in Upper/Great Zab (zaba eliu) and Lower/Little Zab (zaba Sapliu). The Neo-Babylonian world map is oriented so that its top is northwest, the source of the primary ri~er.~'

If Sargon were using a river for his directions in Urartu, he would have been lost, but a general conception that the rivers flowed toward Assyria would have been viable for the final trek homeward. This too would suggest that moving from the top of Urartu to the bottom would imply movement toward Assyria, not away from or parallel to it as in the reconstructions of Muscarella and Levine.

Whatever the precise nuance of these terms, they represent polar opposites, and the basic idea that Sargon entered Urartu on one side of the kingdom and left on the other comes through clearly.42 Given the general form of Urartu's territory, it would make no sense to call Ushnu/Solduz its "head" and the plain of Urmia its "foot." On the other hand, none of the alternatives is inconsistent with a trek that approaches Urartu from northeast of Lake Urmia and ends southwest of it.

Another aspect of the text's structure may be invoked in support of the circum- Urmia route. As noted above, the action took place in five Urartian nag;, or provinces, and the nature of these provinces is a key to understanding the course of the campaign. There is a formulaic quality to the narration, wherein Sargon enters a province, describes a named site or group of named sites, destroys both it and a given number of unnamed "settlements in the vicinity," and moves on. The number of cycles per province varies between one and three. What sort of units and how substantial were these nag@?Each was subordinate to a single provincial governor, written EN.NAM in both Assyrian and Urartian texts. The Harper letters make it clear that the Urartian state was composed of a good many of them. Eleven provincial governors came to grief in a single invasion by the Kimmerians, and this cannot have been the full number who were in office at the time. There are sixteen different designations for individual governors in the Harper letters, and although some of these may refer in different ways to the same individual, the relatively infrequent recurrence of specific names suggests no single governor was of overwhelming importance. However, there were few enough of them that their personal names were sufficiently significant to merit mention in Assyrian intelligence reports.43

The nagz2, however, seems to have geographic as well as political definition. As Sargon moved from one province to the next, topographic features that marked the

40 For example, for kudurrus, see W. J. Hinke, A pp. 135, 138, who argues that these references to the New Boundary Stone of Nebuchadrezzar I, The frontier show that Sargon was never really very far Babylonian Expedition of the University of Pennsyl- inside Urartu and that the province of Armarili, vania, Series D: Researches and Treatises, vol. 4 which I take to be the midpoint of the march through (Philadelphia, 1907), pp. 39-41. Urartu, "is not a region which one can call Urartu

41 See E. Unger, Babylon (Berlin and Leipzig, proper (that the scribe might have termed it 'Urar- 1931), p. 20; and, more recently, W. Horowitz, "The tian' is something else)." Babylonian Map of the World," Iraq 50 (1988): 43 For a more detailed discussion of this question, 147-65. see my Ecology and Empire, pp. 89-94.

42 Cf. Rigg, "Sargon's Eighth Military Campaign,"

dividing line were often noted. Between Zaranda and UiSdiS, from which the attack on Urartu proper was launched, was a mountain pass. Another mountain separated ArmarilP and Ajadi. The march between Ajadi and Uajais took Sargon across three rivers. Such natural barriers were not encountered by the Assyrian army as it moved within a nagti, although there are allusions to nearby mountains as sites of fortresses, places of refuge for the terrified populace, and the loci of watchtowers. Thus the individual nagti cannot have been larger than a single valley, or one of the expanses of flat, arable land sealed off by mountains.

The number of such plains in Urartu is by no means limitless as figure 2, which was drawn from false-color Landsat images, demonstrates. It correlates well with what we can say about the number of governors, and the distribution of Urartian sites in the plains fits Sargon's account. The kingdom was in fact a kind of terrestrial archipelago, where settlements and fortresses clustered in pockets of fertile land, more or less cut off from one another by mountains and inhospitable terrain, and it is perhaps not irrelevant that a secondary meaning of the term nagti is "island." It would makes sense to administer these "islands" as units, but, just as the text would have it, there is no uniformity of settlement pattern within each pocket: some, such as the plains of Khoy and Urmia, had several sites of equal size, while others, such as Ushnu and eventually Qareh Zia' ed Din, were dominated by a single major center. The unnamed "settle- ments in the vicinity" are unattested archaeologically but, given their number, cannot have amounted to much more than individual farmsteads.

If we accept that the nagti is the basic agricultural and settlement area of Urartu, corresponding to the areas of alluvium in which archaeological remains are concentrated; accept that the five of these that were designated "Urartian" by Sargon's letter should manifest intensive cultural and administrative affiliation with the Urartian ruling dynasty; and accept that the campaign should move from what the Assyrians might have construed as the head of Urartu to the foot-then the choice of routes is quite limited. Neither the sub-Urmia nor the circum-Van routes meets the criteria, but the reconstruction proposed by ~eade,~~

in which Sargon circles Lake Urmia, does.

On what basis has this middle alternative been rejected in recent scholarship? Apart from Levine, whose challenge is based on the length of the route,45 I know of no one who has framed an argument against it. The reason that Levine found the passage north of the Lake unacceptably long had nothing to do with the Urartian section of the campaign but, rather, was tied up in the issue of the location of Parsua, which Sargon visited before confronting his principal opponents. In his work on the historical geography of the Neo-Assyrian Zagros, Levine challenged the traditional view that there were at least two Parsuas, one of which was located on the southern shore of Lake Urmia and another farther south. He regarded the former as an artifact of early scholarship, unduly influenced by the co-occurrence of Parsua and MeSta in an inscription of ISpuini and ~enua,~~

and the unwarranted conclusion that since the TaStepe inscription indicated MeSta lay on the south of Lake Urmia, Parsua must also be found in that vicinity. Levine rejected the idea that simply because two places were

44 Reade, "Kassites and Assyrians in Iran,"p. 141. 46 G. A. MelikiSvili, Urartskie klinoobraznye

45 Levine, "Geographical Studies in the Neo-nadpisi(Moscow, 1960), no. 24. Assyrian Zagros-11," Iran 12 (1974): 99, n. 1.



FIG2.-Major agricultural areas in Urartu and suggested circum-Urmia route

mentioned in the same inscription they necessarily had to be close to each other and pointed out that the Urartians controlled the southern shore of Lake Urmia but did not control Parsua, so these could not be the same place. With the Urartian evidence dismissed, the only other argument for the northern Parsua was the Eighth Campaign itself, and Levine's alternative to the Thureau-Dangin route was offered as a solution to this anomaly:

In summary, it is suggested that there is no evidence for the location of Parsua in the north on the shores of Lake Urmia at any time. All references to Parsua, aside perhaps from that in Ashurbanipal's text can be localized in one area, an area in the mountains of the central western Zagros north-west of the Mahidasht, and including the northern end of the Mahidasht itself. The extent of this territory to the east is at present hard to determine. To the north lay Mannea, to the west, ~amri.~'

47 Levine, "Geographical Studies," p. 112.

It followed that the Assyrians had to cover a great deal of ground in the Eighth Campaign before they even came near Lake Urmia, and Levine admonished that this eliminated the possibility of anything longer than an approach to the southern shore for the route of march.

It is difficult to accept that Sargon would pass over so much ground so quickly in his letter-only a few lines of text are devoted to the Parsua episode, most of them listing people who came from elsewhere bringing tribute, and it is not even mentioned in Sargon's annals. Yet this visit would account for roughly one-third of the distance in the Levine itinerary, much of it moving away from Assyria's principal opponent of the first phase of the campaign, Metatti of Zikirtu. Locating Parsua as far to the south as Rawansar would also imply a much more extensive military reach for the Urartians than is customarily assumed. If so, it is odd that their campaign inscriptions make so few references to the other peoples of the central Zagros whom we know of from Assyrian records. Is this excursion really mandated by the evidence?

Levine's argument that the south shore of Lake Urmia was not Parsua is indeed persuasive, but the rationale for identifying a precise location for Parsua elsewhere is less compelling. At best, the few references in itineraries suggest that they could all be placed in that area, not that they must. Given our ignorance of the eastern Zagros, it is by no means ruled out that Parsua was an extensive territory, like that of the Medes, which might be approached from a number of directions. I would, however, prefer an interpretation such as that offered by Pecorella and Salvini, where Parsua is seen as a scattered and mobile entity.48 While in theory inscriptions could link place names that were not proximal to each other, the Urartian records of conquest generally do have some relevance to the direction of the campaign. Our knowledge of Anatolian geography is inadequate to establish whether or not the practice was universal, but in the cases which are clear, toponyms are given in order of increasing distance from the starting point of the campaign, ending with a limit at which the outward march terminated.49 In the annals of ArgiSti I and Sarduri 11, campaigns are listed indi- vidually, even when there was more than one in a given year. The inscription of ISpuini and Menua that mentions BarSua (Parsua) and MeSta together stresses only those two place names, and it is not easy to dismiss their linkage, particularly since other joint inscriptions of these two individuals do not mention either place. In short, these inscriptions do not appear to be compilations. A single campaign against a Parsua that lay south and east of MeSta would seem the most logical interpretation of that text. A more eastern, rather than more southern, Parsua would fit in this instance, and without compelling evidence that there was one and only one specific southern Parsua, there would be no necessity for adding such length to the Eighth Campaign as Levine deems necessary.

Without this extra length, there is no theoretical objection to a march around the lake. The Sahend would remain the most logical candidate for Mt. UauS, and Zikirtu and UiSdiS presumably lay on the eastern side of Lake Urmia. Can the rest of the scheme proposed by Reade be reconciled with the topography of the northern and western shores of Lake Urmia? With the proviso that one cannot rely too heavily on the Eighth Campaign text for literal truth and exact correlations between descriptions

48 Pecorella and Salvini, "Researches," pp. 11-12. 49 For example, MelikiSvili, Urartskie klinoobraznye nadpisi, no. 39.


and features in the modern environment, 1 would tentatively suggest the following correlations, with increasing conviction as one approaches the end of the Urartian segment of the campaign.

  1. zarandaS0 would be the Plain of Marand. It is the first culturally Urartian area that one would encounter if coming from the southeast," particularly if the equation of Mt. UauS with the Sahend posited by Thureau-Dangin and accepted in all other articles on the route save those of Rigg and Levine, is correct. The association with horses recorded in TCL 352 also supports this identification, since although the plain of Marand is one of the larger ones in northwest Iran, it is relatively infertile. Thus, even today herds of horses, and occasionally camels, can be seen grazing there, where they are relatively inconspicuous in the more intensively cultivated plains farther west in Azerbaijan.


  2. Sangibutu would be the Plain of Khoy. It would not be impossible to go directly from the Plain of Marand to the Plain of Salmas, but to do so would entail a passage along the shore of Lake Urmia, to which there is no reference in this part of Sargon's text." There are approximately a score of Assyrian references to Bit Sangibuti, Sangibutuli, and Bit Sangi, all dating to the reigns of Tiglath-Pileser 111 and Sargon, and it is clear they cannot possibly refer to the same geographical ~ocation.'~ Even in


50 Thureau-Dangin and Levine assume that the nag6 discussed in TCL 3, 11. 170-87 is Subi, not Zaranda, but it is hard to reconcile this with the claim that the people of this province did not catch horses as far away as Subi (1. 172). Furthermore, Subi was known as a Mannean province (I. 172), as was UiSdiS (1. 163), which Sargon had just left, so it is hard to see why a fortress between them would be regarded as guarding the Urartian frontier. The text clearly makes more sense if USkaja is an Urartian frontier fortress commanding the approaches from the two Mannean areas and that Sargon's attack on it opened the door to Urartu which had been closed to his messengers. A site such as Livar in the plain of Marand could easily have performed this role.

5' Kleiss and Hauptmann, Topographische Karte von Urartu, sheets 1 and 2. The only site farther east that would have existed in Sargon's time was near Seqindel, in the Ahar Valley, which was well to the north of any logical Assyrian route of march.

52 TCL 3, 11. 170-73.

53 Cf. van Loon, review of Pecorella and Salvini, Tra lo Zagros e I'Urmia, in BiOr 44 (1987): 259, where he states "in line 233, upon leaving Ulhu, he [Sargon] arrived at the sea (a-na A.AB.[BA]) of the lands (un-nu-te), 'head of the fortified towns of the province of Sangibutu'." The restoration, crossing a break between KAH 2, no. 141 and the main text is certainly possible, but the sign read as AB is also partially broken and could be something else. That leaves only a firm A, which is not much to go on.

54 For a list of references, see Simo Parpola, Neo- Assyrian Toponyms, Alter Orient und Altes Testa- ment, vol. 6 (Neukirchen-Vluyn, 1970), pp. 88, 303.

These are reviewed in Levine, "Sargon's Eighth Campaign," pp. 142-43, who concludes that there must be at least two localities involved: a southern one on the Great Khorasan Road and a northern one "near the Urartian border." I would argue that all the references by Tiglath-Pileser I11 are to the southern one, where the terms Sangi and Sangibutu are regularly preceded by "Bit." "Bit Sangibuti" is used to designate the Urartian province only once by Sargon and that is in the Annals, not TCL 3. The only place where the southern toponym occurs without "Bit" is in TCL 3, 1. 39, and this may be a scribal error prompted by the frequent references to the Urartian province elsewhere in same text. In short, if one assumes there was a southern Bit SangibutiISangi well known to the Assyrians from earlier campaigning, and that the Assyrians applied the "Sangibutu" part of that name to an Urartian province, presumably because its Urartian name was similar, the variation in terms makes a certain amount of sense-allowing for scribal errors in two instances. It is unlikely that "Sangibutu" was any- thing more than an approximation of the Urartian pronpnciation, since the consonantal cluster -ng- is unattested in that language, and it is unclear how an Urartian scribe would pronounce the sounds so written in an Assyrian text. Warren Benedict, "Urartian Phonology and Morphology" (Ph.D. diss., University of Michigan, 1958), p. 18, does find evidence for -nk-. In any event, there is no reason to suppose that this llrartian province had been tra- versed by the Assyrians before Sargon's eighth campaign.

TCL 3 there are two, since the one in 1. 39 associated with Namri, Bit Abdadani, and the Medes cannot be the same as the Urartian province described in 11. 188-268. The settlement pattern of this plain, which is well watered and not dominated by a single Urartian site, is not incompatible with the description of Sargon, which notes both a riparian city of Ulhu that was captured without a fight, and a fortress, Sardurihurda, which is passed over so quickly that it is clear that Sargon made no attempt to capture it.

  1. Armarili would be the plain of Salmas. The letter's passage dealing with this is brief, given color only by the reference, noted above, to two sites containing the paternal houses of Rusa and Sarduri, respectively. The known Urartian archaeological material from Salmas is not particularly impressive either, although the poorly preserved levels at Haftavan Tepe do speak for an administrative presence here. In this reconstruction Armarili is well within the kingdom, not on the periphery as Muscarella would have it.


    1. Ajadi, separated by a pass and rivers from its neighbors and the only plain in the area with "cities" by the "sea," would be Urmia. Salvini and Pecorella have pointed out that the three rivers that Sargon crossed going between it and the next province, Uajais, could be correlated with the Qasemlu Chay, the Baranduz Chay and the Sar or Shahr ~hay~~-the


    2. present interpretation would simply have Sargon encounter them from the opposite direction. The Urmia plain is the largest stretch of alluvium that Sargon is likely to have crossed, a~ld it is probably significant that it is here that the letter mentions the greatest number of significant Urartian sites.56 It is unlikely that this major Urartian province would consist only of the relatively infertile strip of shoreline between the Ushnu-Solduz and Urmia plains, to which it is relegated in the systems of Muscarella and ~alvini.~'
  2. The plain of Ushnu would be Uajais. The province bore the same name as its principal site-also written Uhsi, Uesi, Uazaun, and Use-which was no inconsequen- tial place. Sargon says specifically that it was the largest of Rusa's fortre~ses,~' and the Harper letters make it clear that it was a mustering point for the Urartian troops whose concentration was of concern to ~ss~ria.~'


11 was also closely linked with ~usasir~'and stood, as noted, on the "lower border" of Urartian territory. Since

55 Pecorella and Salvini, "Researches," p. 16. would suspect some sort of proximity between the 56 Significant in the sense that the account gives two sites for the number of times references to them their names and not simply a total number. occur in the same letter. For example, ABL 380 57 For the latter. see Pecorella and Salvini, Tra lo (Deller 3.4) and 409 (Deller 5.I), written by different Zagro.~e I'Urmia, fig. 2. p. 47. correspondents but possibly referring to the same 58 TCI. 3. 1. 299. incidents, mention that certain Urartian governors 59 A new edition of the relevant letters is given by are moving with troops to Musasir and. coinci-

K. Deller. "Ausgewahlte neuassyrische Briefe be-dentally. that the Urartian king is going to Uajais. treffend Urartu rur Zeit Sargons ll.,'' in Percorella ABL 112 (Deller 2.1), albeit in a damaged passage and Salvini, Tra lo Zagros e I'C'rmiu, pp. 97 122. (11. obv. 14-rev. 3), appears to report on a request A BL 444 (Deller 1.2) records that on one occasion from the governor of Uajais for troops from Musasir five Urartian governors concentrated their forces in a time of crisis. In ABL 1079 (Deller 1.4). llrzana there. TCL 3, 11. 300-301 also confirms its impor- of Musasir transmits to the Assyrians the news that tance as a center of both military power and the governor of Uajais has been killed. Musasir and espionage. Uajais also appear together in ABL 1083 (Deller,

6" Even without the testimony of TCL 3. one 3.6) in broken context.


Sargon says that he attacked it "from the rear," one would expect that it would have some appropriate directional orientation in plan.

Most of the published itineraries locate this site in the province of Hakkiri, Turkey, but no large Urartian site has been found there. None is likely to be. Such sites do not exist independently of argricultural areas, and there are no suitable agricultural areas in Hakksri. The equation of Qal'eh Ismail Aqa on the edge of the plain of Urmia, first suggested by Pecorella and salvinib' and deduced independently, for different reasons, by usc car el la,^^ is to be taken far more seriously. Certainly the size of the site is an argument in its favor, and both survey and excavation have established that it was occupied at the appropriate time. But Qal'eh Ismail Aqa does not !ie on what would most logically be regarded as the lower frontier of Urartu, nsr is it well placed to be a staging area for Urartian operations that were of such immediate concern to the Assyrians as to make movements in and out of Uajais one of the preoccupations of the Harper letters. Moreover, it was from Uajais that Urartian campaigns also set out against the Manneans, and in one instance, with specific reference to ~ikirtu.~' It would seem improbable that all of this activity would be focused on a site so centrally located in Urartu's territory as Qal'eh Ismail Aqa. There is another large site, much more closely associated with the Mannean frontier in anybody's reconstruction and enjoying better access to Musasir, that provides a more satisfactory location for U aj ais.

This, of course, is Qalatgah, discovered by MuscarellabJ and even once considered briefly as a candidate for Uajais by Salvini himself.65 It is unquestionably large, and certainly larger than any other site in the vicinity of Urartu's southern frontier, although its relative size compared to Qal'eh lsmail Aqa is not known because no measured plan of it has ever been made.6b Other arguments in favor of identifying Qalatgah with Uajais, none compelling in itself but suggestive in the aggregate, include the possible survival of the ancient name in the modern city of ~shnu,~'

the discovery of a building block with a putative reference to the ancient site in an unclear context at ~alatgah,~~

and the existence of a candidate for "Old Uajais" in what might reason- ably be expected to be administered in the neighboring province-~asanlu.69 Most of

61 Pecorella and Salvini, "Researches." p. 16 and known to have any Urartian remains, let alone to Tra lo Zugros e I'llrmra, pp. 46-5 1. have been the largest Urartian site in the area, it 62 Muscarella, "Location of Ulhu and UiSe." would have to have received its name by transfer- pp. 472 730. ence after Qalatgah had been abandoned, as pre-

See Parpola, The Correspondence of Surgon II, sumably Uajais had taken its name from the "Old Purr I. State Archives of Assyria, vol. 1 (Helsinki, Uajais" mentioned in the letter (I. 285). 1987). no. 29. pp. 28~29. 68 van Loon, "Inscription of ISpuini," pp. 201-7:

64 Muscarella, "Qalatgah: An IJrartian Site in there is, however, nothing in the text linking the Northwestern Iran," Expedition 13 (1971): 44-49. place names with the findspot of the blocks, and 65 Salvini, "Die urartaischen schriftlichen Quel-Salvini. Tra lo Zagros e I'lirmia, p. 24, casts doubt len," p. 174. on whether a variant of the name Uajais is really

66 For a sketch plan of the site, unfortunately not even to be found there. to scale. see Kleiss, "Bericht iiber Erkundungsfahr- 69 This can only be offered as a suggestion, the ten in lran im Jahre 1970," AM14 (1971): 64. argument being as follows: Old Uajais still existed

67 This was suggested by J. V. Kinnier Wilson. and was important enough to be named in passing "The Kurba'il Statue of Shalmaneser 111," Iraq 24 by Sargon's account, but it was in Ajadi, not the (1962): 109-10 and reiterated by van Loon, "In- nag; named after the newer Uajais. Qalatgah was scription of lipuini," pp. 205-7. Since Ushnu is not apparently founded by Menua as the primary

these arguments, rehearsed in a footnote in my study of the structure of the Urartian state,70 have recently been challenged by Muscarella, who rejects the identification of Uajais with Qalatgah because he does not feel it has a clearly defined "rear,"71 because he does not regard the evidence of the inscription or the equation of the place name with Ushnu as having any weight, because he accepts Salvini's equation of Uajais with Qal'eh Ismail Aqa, and because he has what he regards as a much better candidate for Qalatgah in the letter's extensive discussion of ~lhu.~~

The last point is certainly intriguing in view of the rather dramatic springs that pour out from the base of the citadel rock, whose waters fill the pool beside which the Qalatgah inscription was found in secondary usage. Ultimately, however, this is no more convincing than the pieces of contrary evidence that Muscarella correctly regards with skepticism. The passage describing Rusa's irrigation works at Ulhu is fraught with difficulties, not the least of which is that the text is broken at this point.7' My primary objection to this identification just from the passage alone, is that Ulhu is paired with another site, Sardurihurda, whose qualities as a fortress are stressed. Ulhu, although fortified, seems more a royal residence, which its inhabitants made no attempt to defend. At the approach of an enemy the Urartian populace would be expected to run toward, not away from, a powerful fortress like Qalatgah. Muscarella's contention that "nowhere else in the text does one encounter the enthusiasm expended on the water source of a city, and it is worthy of note that no other site in northwest Iran but Qalatgah qualifies for such enth~siasm"~~

strikes me as unwarranted, par- ticularly if one is not swayed by the rather weak philological basis for his understand- ing that the water supply originated within the city. The main point of the passage is

Urartian site in the Ushnu/Solduz Valley shortly after the destruction of Hasanlu IV, which had been the dominant locale in the area previously. It would be quite reasonable for the Urartians to give their new administrative center the same name as the center they had just destroyed, and although Hasanlu was later refortified by the Urartians, it never recovered its former status, no doubt in part because fortresses on mountain spurs, rather than tells, were now the flagships of the royal bureaucracy. It is less easy to understand how Old Uajais wound up in a different province. Sargon's scribe may simply have got it wrong, or else Solduz was administratively separate from Ushnu, which has frequently been the case historically. If so, it might have made sense to administer it from the Urmia district, since the avenues along the shore of the lake foster easy communications with that area, leaving the new Uajais free to concentrate on the problems of Urartu's southern frontier. In any event, the greater antiquity of settlement at Qal'eh Ismail Aqa makes is a less attractive candidate for this kind of name transfer.

'0 See my Ecology and Empire, p. 112, n. 64.

71 On this point I differ. Like many Urartian castles, Qalatgah is built on a rock spur overlooking a plain that slopes away from it gently, and sepa- rated from much higher ground by a shallow saddle. The normal approaches to the site would be from the valley, but a logical tactic of a besieger would be an onslaught from the higher ground. At Qal'eh Ismail Aqa, the situation is different, since the valley of the Nazlu Chay intervenes between the high ground and the eminence on which the site is con- structed. Salvini, Tra lo Zagros e I'L1rmia, p. 50, has suggested that "I1 castello inferiore, congiunto da una cortina di mura con quello superiore, ma da esso distinto, pub essere il kutallu di cui parla Sargon, e la porta (abullu) pub significare la porta principale della fortezza alta." Kutallu, however, generally refers to the backs of things, be they animals or buildings, and is used in idioms suggest- ing obscurity or remoteness. The lower castle at Qal'eh Ismail Aqa, closer to the primary gates of the site and much more accessible to its Urartian inhabitants than the upper castle, hardly seems to fit those connotations.

72 Muscarella, "Location of Ulhu and UiSe," pp. 465-75.

73 For a recent study of the Ulhu episode and a suggestion that Sargon's scribes' inspiration for the associated motifs came from an Urartian inscription, see C. Zaccagnini, "An Urartean Royal Inscription in the Report of Sargon's Eighth Campaign," in

F. M. Fales, ed., Assyrian Royal Inscriptions: New Horizons, Orientis Antiqui Collectio 17 (Rome, 1981), pp. 259-93.

74 Muscarella, "Location of Ulhu and UiSe," p. 469.

that Rusa (Ursa) had created an irrigation system that had brought abundance to his land. Although the modern springs that emanate from the rock at Qalatgah are now dammed up and sufficient to irrigate a small field, the main source of water in the Ushnu Valley is the Gadar Chay, and only this would be sufficient to create irrigation works on a scale described by the Assyrians. C. Zaccagnini's literary analysis of this passage gives one pause in accepting a strict literal reading in any event, since it is structured in parallelisms inspired by artistic concern rather than physical description.

Virtually every alluvial area in northwest Iran held the potential for dramatic agricultural transformation when irrigation was introduced, as it regularly was by the Urartians. The Ushnu plain had long been under Urartian control by the time of Sargon's march, so the idea that its fertility was radically augmented by Rusa's hydraulic management is suspect. A better case may be made for the plains north of the lake, where the major Urartian construction projects are later in date. I have no specific candidate to offer for Ulhu, but a small tell beside the Qotur, in all probability long since washed away, would fit the references in the Eighth Campaign letter well, with Kiz Kalesi near Khoy serving as Sardurihurda.

There remains the question of the return march, which is tied up with the problem of the location of HubuSkia. After departing from Uajais, Sargon received a deputation from the king of Nairi, who had marched four double hours (beru),i.e., more than 40 km," from his royal city, HubuSkia. The location of HubuSkia, which appears elsewhere as the name of a land, is another controversial topic, with some putting it in the valley of the Bohtan Su south of Lake Muscarella arguing for a location farther north, somewhere not far west of Qal'eh Ismail and others suggesting areas farther south and east." While the last hypothesis accords best with the reconstruction of the itinerary presented here, the issue is not critical if one allows that Sargon did not actually go to HubuSkia himself. The text stresses that the deputation came to him from some distance. In the next line it does state that Sargon received tribute in the city of HubuSkia (ina qereb uru~ubu~kia),79

but this is simply a topos and hardly enough to create visions of a substantial detour in the course of the march,80

Finally, Sargon became incensed at the disloyalty of Urzana of Musasir who revolted against him and failed to bring trib~te.~'

Sargon sent the body of his army

75 CAD B, p. 208 80 Cf. Muscarella, "Location of Llhu and UiSe."

76 Salvini. Nairi e Urlu)a!ri, Incunabula Graeca, pp. 473-74. where the absence of any comment on vol. 16 (Rome, 1967). p. 72 and Levine. "Sargon's the march to HubuSkia is taken as an argument for

Eighth Campaign," pp. 143-44. its proximity to Qal'eh Ismail Aqa. 77 Muscarella. "Location of Ulhu and LiSe." 81 Oppenheim. "City of Assur." pp. 135-36. expp. 473-74. amines this incident, coming to the conclusion that

78 For example. Reade, "Hasanlu, Gilzanu and we really do not know what happened but that it Related Considerations," AM1 12 (1979): 178, con- represented a major crisis for the expedition. Since cludes: "if we place Hubushkia in the Ushnu-Urzana's family was carried off by Sargon in the Rowanduz-Pizhder-Mahabad quadrangle, we can attack on Musasir but Urzana himself was appar- hardly be far wrong": and, more recently, Russell, ently elsewhere, I would follow Oppenheim's sugges- "Shalmaneser's Campaign to Urartu in 856 B.C. and tion that he was dealing with the main body of the the Historical Geography of Eastern Anatolia ac-Assyrian army. This would all make a certain cording to the Assyrian Sources." Anatolian Studies amount of sense if Urzana were attempting to 34 (1984): 195-98, places HubuSkia in the vicinit) of obstruct or delay the Assyrian retreat by blocking modern Rowanduz. one of the better passes farther to the south.

79 TCL 3. 1. 308.

back aiong one road while, with a small contingent, he himself crossed the mountains for an attack on Urzana's capital. As Lehmann-Haupt pointed out long ago, the logical route for an army to return to Assyria from the western shore of Lake Urmia would certainly not be through the rugged mountains of Hakkiiri. Even if one rejects the identification of Musasir with Mudjesir, the general location of Musasir's territory is known. From the Ushnu valley it could be approached via the Kelishin pass, where a bilingual stele stood, apparently marking the Urartian frontier. But this was no road for an army, nor was it a particularly good way to get from Urartu to Assyria-the best road for those purposes was via Khane and ~owanduz.~~

Thus it made sense for Sargon to separate himself from the main body of his troops and take a small contingent across Kelishin. A surprise strike at Musasir's capital while Urzana's attention was focused on the main army attempting to cross the easier pass to the south, was a viable stratagem for extricating Sargon's forces from hostile territory- not an irrational excursion for revenge.

The letter's statement that Sargon crossed the Upper zabg3 before reaching Musasir has probably been the main factor discouraging acceptance of this route. Boehmer, for example, assumes that Sargon must have approached Mudjesir from the west84 and Levine's preference of a northern location for HubuSkia is tied to its association with this river.85 This problem has recently been treated in detail by H. F. Russell, who suggests that what Sargon calied the Upper Zab was actually the Rowanduz ~iver.~~ Certainly skepticism about equating ancient and modern views of where the "source" of a given river lies is justified, and the Greater Zab is fed by so maily tributaries in this mountainous area that no specific location can be extracted from this reference.

The arguments offered above obviously do not constitute any kind of proof that Sargon marched around Urmia, but they should at least suggest that there is some merit in that conception. There is no other way to move through Urartu and hit the correct number of provinces and geographical features. The routes that limit Sargon to a path south of lake Urmia do not encounter enough Urartian territory for five provinces, even if one assumes that Ushnu and Solduz were two separate n~gP.Any route north of both lakes would cross many more Urartian provinces, in addition to drawing the Assyrians far away from Musasir. The Ushnu Valley at Qalatgah, from which the cleft in the mountains leading up to the pass at Kelishin is visible, thus seeins a likely terminus for the Urartian part of the campaign. Here, where the southernmost Urartian inscriptions have been found, lies a major citadel proximal to Musasir as well as some of the primary passes between Assyria, Urartu, and Mannean Iran. The various suggested itineraries that return Sargon and his troops through

82 Major-General Sir Charles Wilson, ed., Mur-86 Russell, "Shalmaneser's Campaign," p. 198. ray? Handbook for Travellers in Asia Minor, Trans-Russell's interpretation of the final phase of the caucasia, Persia, eic. (London, 1895), pp. 321 -22. Eighth Campaign itinerary is quite similar to my

83 TCL 3. 1. 323. own but takes Sargon himself across a more south- E4 Boehmer and Fenner, "Forschungen in und um erly pass before breaking with his army to march Mudjesir," pp. 513-14. north to Musasir. 85 Levine, "Sargon's Eighth Campaign," p. 144.



Hakklri are incompatible with the current state of knowledge of Urartian settlement and fortification practices, in which the really mountainous areas were not fortified, but left to defend themselves.

The length of a circum-Urmia march is considerable but not excessive. It is certainly much shorter than Shalmaneser's campaign of 856 B.c.," and, for what it is worth, Ottoman armies in no particular hurry could move from one side of Urartu to the other in roughly one month.8H If Oppenheim is correct that a lunar eclipse is suggested by Sargon's remarks preceding the attack on Musasir, the date of that attack would be 24 October 714 ~.c.~~-allowin~ one more

about month for campaigning than is customarily assumed given the prevailing climatic conditions. Generally the Assyrian army seems to have been able to cover considerable distances in Urartu rather easily, but this did not enable them to effect lasting conquests there. For that they would have needed siege equipment and time to overwhelm fortified locations. The letter focuses on only two such incidents of conquest, Ulhu and Musasir, and these may in fact have been the only ones attempted. Ignoring fortresses, the army could probably have covered more ground on the plains of Marand, Khoy, Salmas, Urmia, and Ushnu than in the mountains between Lakes Van and Urmia.

Any reconstruction of the itinerary on the basis of the available facts involves choices between conflicting pieces of evidence and a selective application of skepticism and credence. Should, for example, Sargon's failure to claim a circuit of the lake outweigh his explicit stipulation of a march from "head" to "foot" of Urartu, which may be argued to be a mere figure of speech? The rather casual nature of the one clear reference to the lake suggests that the issue of where the army went in relation to it was not of overwhelming importance to the scribe, at least in comparison to the places like Mt. UauS which were remote enough to merit elaborate description. The thesis of this article, that Urartian geography is indeed a major concern of the framers of the letter and that the descriptions given cannot be reconciled with an itinerary that largely avoids Urartian territory, may well be as subjective as propositions that revolve around specific sites. The deck of historical facts has been so thoroughly shuffled by Assyrian scribes that the controversy is not likely to be resolved with the information presently available to us. In the future, excavations at Qalatgah and Qal'eh Isrnail Aqa, one of which is almost certainly Uajais if the statement that it was the largest of Rusa's fortresses has any validity, have the potential to resolve the issue of the end of the campaign, but this alone will not settle the question of whether Sargon went north of Lake Urmia. Scholarship on the Eighth Campaign will continue to wander as long as we have to ask poets for directions.

87 Ibid., pp. 171-201. Kanuni 1534 35 nach Matrakci Nasuh," ZDMG

88 For example, Franz Taeschner. "Das ltinerar 112 (1962): 50-93. des ersten Persienfeldzuges des Sultans Suleyman 89 Oppenheim. "City of Assur," p. 137. 

  • Recommend Us