Imperialist Networks: Ancient Assyria and the United States

“Not since Rome has one nation loomed so large above the others”. With this sentence, Joseph Nye opens his book The Paradox of American Power (2002: 1). After the end of the cold war, and especially since 9/11, the notion of an “American empire” and the comparison with Rome as the sole precursor of the U.S. empire have become a mainstay of historians and political commentators alike (e.g. Kagan, 2002; Golub, 2002). Rome also underwent a transformation from republic to an imperial golden age, and reference is made to opinion leaders such as Cato, who said of Rome’s enemies: “Let them hate us as long as they fear us.” This sentiment is echoed by Charles Krauthammer (2001) who asserts that America “is the dominant power in the world, more than any since Rome. Accordingly, America is in a position to reshape norms, alter expectations and create new realities. How? By unapologetic and implacable demonstrations of will.” For other commentators, Rome is a negative prototype. Chalmers Johnson (2004: 15-16) equates America today with the end of Rome’s Republic and its transition to the undemocratic Imperium Romanum, while Michael Hardt and Antonio Negri (2000: 164-167) draw a parallel between the rhetoric of a Pax Romana and a Pax Americana, both of which are based on permanent war. A book-length treatment by journalist Cullen Murphy (2007) pulls together a set of parallels and focuses on the function of the imperial capital.1


Introduction
"Not since Rome has one nation loomed so large above the others". With this sentence, Joseph Nye opens his book The Paradox of American Power (2002: 1). After the end of the cold war, and especially since 9/11, the notion of an "American empire" and the comparison with Rome as the sole precursor of the U.S. empire have become a mainstay of historians and political commentators alike (e.g. Kagan, 2002;Golub, 2002). Rome also underwent a transformation from republic to an imperial golden age, and reference is made to opinion leaders such as Cato, who said of Rome's enemies: "Let them hate us as long as they fear us." This sentiment is echoed by Charles Krauthammer (2001) who asserts that America "is the dominant power in the world, more than any since Rome. Accordingly, America is in a position to reshape norms, alter expectations and create new realities. How? By unapologetic and implacable demonstrations of will." For other commentators, Rome is a negative prototype. Chalmers Johnson (2004: 15-16) equates America today with the end of Rome's Republic and its transition to the undemocratic Imperium Romanum, while Michael Hardt and Antonio Negri (2000: 164-167) draw a parallel between the rhetoric of a Pax Romana and a Pax Americana, both of which are based on permanent war. A book-length treatment by journalist Cullen Murphy (2007) pulls together a set of parallels and focuses on the function of the imperial capital. 1 In this paper, I present an alternative to these equations by comparing the American and the Assyrian empire -the latter dating to the early 1 st millennium BCE, based on my contention that they have structural similarities not shared by Rome. I will first give a brief outline of the mechanisms and structures of the United States' empire. In a second section, I discuss the historiography of the Assyrian empire and elaborate on practices of power. I integrate my own fieldwork in Eastern Syria as an illustration. In a concluding section, I return to the American empire.

The United States as a Network Empire
Comparative historical analyses are important for any historically oriented archaeology and anthropology since a tertium comparationis, the complex of commonalities and differences between two or more entities, delivers historical insights that cannot be derived from singlecase accounts (Kaelble, 1999). Eisenstadt's (1982) ideas on an axial age, but also Canfora's (2006) history of democracy prove the point. Cultural history and approaches in the vein of a postprocessual archaeology shy away from such avenues to generalization, to their own loss. Generalization, however, should not be the goal of research, as it was in processual archaeology and is drive for the discovery of laws and rules. Rather, it is a means to other ends. My approach attempts to track similarities in two historical cases in search of new insights into both of them. 2 We can mobilize knowledge about American imperialism to understand Assyria better. But ancient Assyria equally enables us to sharpen our understanding of the structures and dynamics of the American empire.
My initial inspiration for this comparative project came from reflections on a number of disparate sources. Among them are Liverani's model of Assyria as a "network empire " (1988; 1992), Johnson's The Sorrows of Empire (2004) and Hardt and Negri's (2004) recent volume Multitude. Apart from such general readings, this paper is also driven by a reanalysis of fieldwork I conducted long ago in eastern Syria, and reflections on reports by the U.S. Department of Defense of plans for the American military's future. As will become apparent, Liverani's geographic model of a network empire, in connection with the militaristic character of that polity, is applicable to the contemporary U.S. It sets both America and Assyria apart from most other historically known imperial powers, and especially from ancient Rome, so often invoked by U.S. scholars as a precursor to the present global configuration.
I begin by briefly discussing structural elements of network empires. Traditionally, scholars have divided empires into a core zone and a periphery, or an imperial core and colonies. 3 Since theories of imperialism date to colonial times, when whole swaths of foreign continents were occupied by imperial powers, it is understandable that both zones have been conceptualized as territorial entities. However, in network empires, only the core is a territorially dominated region. In the case of the United States, the limits of the nation-state coincide with its core. 4 Within it, we can differentiate an imperial center and its hinterland. It would be overly simplistic to see the capital Washington alone as the imperial center. While most political decisions that directly concern the imperial periphery are taken there, the empire's center also comprises economic nodes such as Wall Street and major military units in places such as Tampa, Florida. Thus, the center of the imperial core, if once spatially centralized, has become dissipated with advanced communication technologies. The core's hinterland is the vast expanse that provides the imperial center with resources, including materials, scientific industries, financial capital and especially consumers. In imperialist eyes, both the center and the hinterland profit from the imperialist project.
The empire's periphery consists of a network of interconnected nodes in a sea of interstices. The U.S. imperial network is fundamentally a network of military bases (Johnson, 2003;. This global net has been divided into six regions with command centers, ruled by Commanders in Chief or CINCs from bases in the U.S. and Europe. The command center with the longest tradition, CENTCOM (Central Command), has a defined operational field in the Middle East, including countries reaching from Pakistan to Egypt and from Kazakhstan to Yemen (figure 1). On the next lower level of the network are 13 large installations, mostly huge air bases such as Okinawa in Japan and Ramstein in Germany, and some sea-ports that serve medium-sized installations. The lowest level are specialized bases, such as those for training, repair, and, in war zones, forward operational bases. The long-term configuration of the major air bases and ports stabilizes an otherwise highly dynamic, shifting network of small units. New bases are constantly established and older ones abandoned according to the immediate needs of war. Where the U.S. becomes involved in wars, the meshes of the network thicken abruptly, and the nodes rapidly increase not just in density but in size as well. 5 Fig. 1 Node articulation is a major concern of all networked systems. The U.S. military has two fundamental means of network links. For material transport of people, arms, and provisions, the military uses air routes. Thanks to air fuelling technology, connections are both fast and global. The other means of articulating nodes is through the internet, which origi-nated in ARPANET in the 1960s, a section of the Department of Defense (DARPA, n.d.; Salus, 1995). Both types of articulations remain materially invisible. The spaces needed for air traffic and internet connections are part of the so-called "Global Commons," that is the earth's atmosphere, outer space and digital space. The U.S. government's National Defense Strategy report from 2002 expresses a doctrine of unhindered use of these spaces, including outer space (Grondin, 2009), as well as the exclusion of anyone considered hostile (see also Lemann, 2001). This network of bases and connections allows the American empire to intervene rapidly almost anywhere in the world. However, U.S. strategies of imperial expansion typically do not include territorial annexation of whole countries and the installation of core institutions, such as a state administration and tax collection, or universities, postal services and the like. Even the invasion of Iraq was not originally meant to be an occupation, but to establish a monitoring presence. In Donald Rumsfeld's imagination, this was supposed to be a small "shock-and-awe" intervention with a rapid return to low-level interference (Moustakis and Chaudhuri, 2007), followed by the establishment of a few controlling bases and the intrusion of private businesses. The "establishment of democracy" was at best the legitimizing ice on the cake of increased political-economic power by way of military control.
Military bases in foreign countries are off-limits for local populations and have their own police, jurisdiction, monetary system, cultural affairs and sports organizations. The most famous of these entities is certainly the Baghdad "Green Zone" with the largest U.S. embassy in the world, which forms a not so small city for the occupiers within the capital of Iraq (Langewiesche, 2004;Chandrasekaran, 2006). In general, personnel on U.S. bases in sovereign countries interfere minimally in political affairs in their host nations unless there is a perceived threat to American interests. However, in most cases, bases are still seen by locals as a foreign, hostile intrusion, and tensions between soldiers in the nodes and surrounding populations run high especially in zones of active conflict. 7 This situation leads to concerns about attacks on bases, so that contact is intentionally further restricted, and bases have an Americanized interior that contrasts starkly with their surroundings. 8 This structural-cultural divide emphasizes the cultural distance between the imperial network and its interstices. The specific cultural, but also political sensitivities of the interstices are to a significant extent constituted by these relations.
The establishment of new bases is a springboard into as yet uncontrolled areas. The defeat of a hostile country is the golden opportunity to expropriate large tracts of land that are needed for various types of new nodes (Foster, 2004: 55-64). It is no coincidence that Germany, Japan and South Korea have the greatest number and largest bases of all those listed in the Base Structure Report. Nowadays, Kandahar airport, Bagram and Balad Air Base in Iraq (Ricks, 2006) can be added to this list.
This brief structural analysis reveals the six key elements of network empires, valid for the U.S. as well as for ancient Assyria: • an imperial center, located in a territorially controlled core zone; • the hinterland of the imperial core • a periphery with more or less densely spaced nodes of a network • a hierarchy of these nodes that is based on specific functions • a system of network articulation that is highly efficient • interstices in the network, the areas that the empire seeks to monitor The structure of the periphery of a network empire contrasts starkly with that of territorial empires, particularly the Roman one with its symbolically marked boundary, the limes, as well as clearly demarcated provinces (e.g. Kennedy and Riley, 1990;Klee, 2005). 9 The Imperium Romanum had a clear interior, but also an exterior that was beyond the empire's reach. In a network empire, the claim to domination and control is not limited by any symbolically and militarily marked lines on the ground. Imperial power strives towards global reach, even though interstices are never completely under its purview. The ideological assertion of world-wide scope leads to the curious effect that enemies of empire can only exist within its own realm, in the interstices of a periphery thought of as global. Ideologically, such empires do not recognize the possibility of land beyond their periphery (Rhodes, 2005). Wars against rebellious coalitions in the interstices necessitate constant adjustment of the imperial network by thickening the meshes in one place and thinning them in others, as is directly visible today with the closing of bases in Germany 10 and the opening of new ones in Iraq, Afghanistan, and elsewhere in CENTCOM's purview. Paradoxically, network empires claim global control in the face of unending war, because there is no outside of empire, but there still are powerful enemies. In the following section, I turn to the Assyrian side of our comparison. I give a brief overview of Assyria's history and then discuss two conceptualizations of practices of imperial power, one traditional, the other more in line with a view of Assyria as a network empire.

Constructions of Assyrian Imperialism
The core of Assyria with its important cities Nineveh and Nimrud was a region located on both sides of the river Tigris in northern Iraq, close to the modern Iraqi city of Mosul (figure 3). Its history reaches back into the early 2 nd millennium BCE, and the first, brief phases of expansion date to the Middle Assyrian epoch in the 13 th and early 11 th centuries. I limit myself here to "Neo-Assyrian" times, the major period of imperial expansion, lasting from 911 to 612 BCE. These 300 years can be divided into three major phases. In the 9 th century, two kings, Ashurnasirpal II and his successor Shalmaneser, led almost yearly military campaigns to the West, expanding the empire's reach beyond its former Middle Assyrian extent into the regions of Sam'al, Hamath and others (figure 3). Ashurnasirpal II also founded a new capital at Nimrud, starting a tradition of building vast and lavishly decorated palaces (figures 4a and 4b). The second imperial phase lasted from the late 9 th to the mid 8 th century BCE and is traditionally seen as a period of crisis (Kuhrt, 1995: 490-493;van de Mieroop, 2004: 230-232). In this epoch, the Assyrian empire did not grow appreciably. The third and last phase, from 745 to 612, is marked by a rapid and widespread expansion. In less than 150 years, Assyria ventured into the mountainous Taurus regions to the north, its relations to once revered Babylonia changed from hegemonic to an occupation (Cancik-Kirschbaum, 2003: 71-81), and Egypt was conquered for a brief time (Eph'al, 2005). Concurrent with this expansion, a reorganization of some peripheral regions into provinces was attempted. Assyria's demise was sudden, and historians disagree on the extent to which imperial overstretch or a two-pronged attack by the Medes and Babylonians in the east and south contributed to its collapse (Liverani, 2001;Cancik-Kirschbaum, 2003: 95-100).   Historians have based their accounts of Assyria mainly on royal annals (Luckenbill, 1926;Grayson, 1976). These documents recount military campaigns against foreign enemies, crushing of revolts, victory, tribute collection and deportation as a standardized narrative. Their propagandistic nature is apparent in their hyperbole, and the arrangement of a series of conquests into a pre-conceived, artificial sequence in an ideological chronospace (Renger, 1986;Yamada, 2000;Heinz, 2008: 121-140). The imagery that underscores these exploits is equally selective, never showing the deaths of Assyrians but displaying graphically those of enemies. Enemy rulers are skinned alive, put on stakes and beheaded (figure 5; Bahrani, 2008;Crouch, 2009;Fuchs, 2009), while soldiers retrieve the heads of the adversarial army in order to collect remuneration (Dolce, 2003). In contrast to royal annals, letters from administrative offices in the network nodes address shorter-term governmental problems. These should be of greater historical value, as the senders often report difficulties in meeting the king's orders, loss of power over whole regions, and defections. Even the letters, however, are unreliable in that they show a palpable fear of addressing major problems squarely to the king (e.g., Parpola, 1980, passim).
Historians of Assyria have paid undue attention to the propagandistic monumental inscriptions of the kings. The resulting deeply engrained modern perceptions of Assyria would be worth a study in itself. However, for the purpose of this paper, I simplify a majority opinion held since the 19 th century and set in stone by Emil Forrer's (1920) work on provinces of the Assyrian empire. Standard historiography conceptualizes Assyria as a territorial empire with a periphery organized into small, more or less dependent, contiguous provinces ruled by governors (Postgate, 1992). This view is deeply influenced by a set of implicit ideas. First, the functioning of an empire is likened to the mechanisms of a state's government. Second, politics is imagined as a set of institutional structures that can be described adequately by referring mainly to the actions of a few "actors" at the top of an imperial hierarchy. The reproduction of an imperial system through the complex and contradictory practices of a whole apparatus is given short shrift.
19 th and early 20 th century historians of Assyria fell prey to the political conditions of their own lives. They de-historicized practices of emerging nationalist states that Foucault (1991) has aptly characterized as a "modern governmentality." These can be briefly summarized in terms of three main elements: • the fusion of a disposition and a praxis of power that has as its subject a territorially circumscribed population • technologies of subjection that are administrative and disciplinary • a governmental knowledge that is based on political economy and statistics (Asad, 2002) Specialists on the ancient Near East, whether historians or archaeologists, have imposed on Assyria these inappropriate elements of modern governmentality. The anachronistic paradigm of the Assyrian empire presupposes the presence of four "dispositives." 11 First, Assyrian power is exercised through military campaigns, and once we have documentation that a territory has been reached, it is reasonable to assume that this region became dependent, a client state or an imperial province (e.g. Parker, 2006: 85). Expansive practices of power are military in nature and lead to occupations. Second, the goal of expansion is the establishment of control over a foreign country through the installation of an Assyrian governor or a treaty that ensures imperial relations, leading to reliable taxes and tribute for the imperial core. Third, philologists and historians have inferred that economic documents were written in Aramaean on parchments which rarely survive (e.g. Parpola, 2003; but see Fales, 2007: 97-98). The administrative grasp of Assyrian power, or, if one wants, its bio-politics, is constructed on the presumption that there is a gaping hole in the historical archive (Postgate, 1979: 195). Fourth, the territorial basis of imperial control is asserted through the installation of monuments in the periphery: if the Assyrian king reached a specific geographic place and set up a monument, scholars imagine not only that the surrounding region became part of the imperial periphery, but that regions between this point and the core were included in the empire (e.g. Roaf, 1990: map p. 164;map p. 179;Liverani, 2005: 232). Similar to the various fences and watchtowers lining their own present national borders, historians have turned these monuments, but also rivers or lines of Assyrianized cities into markers of imperial boundaries of a kind comparable to the Roman limes and the Chinese Wall (e.g. Kühne, 2009: 46). Following the traditional model of Assyria as a territorial empire, scholars have determined its boundaries as those of imperial sovereignty (Tadmor 1999).
But even Assyrian kings considered their realm smaller than do present historians. I infer this from an analysis of royal texts that describe the erection of 57 monumental stone stelae (Börker-Klähn, 1982: 177-224). The mentions in these texts are only very rarely matched by real monuments. Almost 80% of those mentioned in Assyrian texts were set up mainly at the outer fringes of the periphery, some in the interstices, but only a few in cities that had been Assyrianized. Seven percent are from the core of the empire and only 14% mark Assyriandominated spaces in the periphery. Using known emplacements of royal stelae for the assessment of the geographical reach of Assyrian power surely overestimates it. This was rather an ideological statement that Assyria would enforce its interests, should they not be followed voluntarily, and not a sign of bureaucratic inclusion into the empire.
Notions of territoriality also play a role in the modern-day reading of other texts. When letters or royal annals talk of "Ruler X of Land Y", as in "Ahuni of Bit Adini", this is assumed to denote a fixed territory with contiguous boundaries to neighboring political entities. However, the real sphere of peripheral rulers' exercise of power was likely restricted to loyal villages. 12 Uncritical projections of modern notions of governmentality back to ancient Assyria also distort ideas about the institutional frame of administrative and disciplinary practices. One of the core notions that determine views of the Assyrian periphery is the term bēl pāḫete, almost always translated as "governor." Šaknu is a similarly often used official position which means a priori nothing other than "appointed" (Postgate, 1980: 67). While I cannot claim major philological insights, I am interested mainly in governmental practices on the part of such people. Based on a review of a number of Neo-Assyrian texts that describe what these officials were doing, such a translation seems to me to produce problems for three reasons.
• First, it is highly unlikely that a term denoting an office remained stable over a period of 300 years. "Histories of terminology" or Begriffsgeschichten have firmly established that even basic terminologies are subject to rapid historical change. One need only consider the history of words such as "governor" or "citizen" in the United States since the 18 th century.
• Second, the assumption that bēl pāḫete or šaknu denote only one administrative position is mistaken. This is clear from Assyrian documents that mention bēl pāḫete in the northern kingdom of Urartu, a hostile polity that demonstrably had no provincial system (Zimansky 1985: 89-93). 13 • Third, Assyrians used the term bēl pāḫete for a variety of purposes. In official, monumental inscriptions, the notion may be similar to our meaning of "governor." However, in royal letters, indicating real practices of power, bēl pāḫete often appear in the plural, acting militarily in foreign lands. The basic meaning of "pīḫatu" is close to our "responsibility", and "bēl" denotes the one in charge. In military contexts the whole notion may come closer to a flexible meaning of a "commander" whose responsibilities were not as clearly set as those of a Roman official.
Assyrian positions of power were not as predefined as we might think when encountering the translation "governor". In fact, we likely have to conceptualize this title as denoting a person who maintained power over a section of a network with a specific number of defined nodes, but with fuzzy edges. What mattered primarily in the network periphery were loyalty and the capability of the military leaders to impose the royal will, not a uniform administrative control over land and people.
To summarize these arguments, I contend that one of the main flaws in the prevailing historiography of the Assyrian empire is an unrecognized universalisation of "governmentality," a set of dispositives firmly anchored in modernity and its state structures. This concept has led to a notion of the Assyrian empire with a colonial structure, one that mirrors the conditions of its scholarly invention under the dominant ideas of British imperialism in the late 19 th century. My own approach differs from this insofar as my comparison between two empires is explicit, and is meant to support our status quo but rather to provoke critical reflection about it.

An Alternative View of Assyrian Imperialism and Practices of Ruling
A major change in the understanding of the Assyrian empire comes from an article by Mario Liverani (1988; see also Lamprichs, 1995). Liverani points out that at least in its early phases, Assyrian power over the periphery was rather exercised by means of a network. Ensuing debates within the discipline of Assyriology about Liverani's thesis have certainly led to new insights about the past (Postgate, 1992). But in my view, the dispute over the characteristics of Assyrian ways of ruling have not gone far enough, as discussion has been mainly restricted to issues of political geography. Two other aspects are missing: practices of power, on which I focus here, and ideologies, which have been discussed many times by others but which warrant more study (Liverani, 1979;Fales, 1981;Morandi Bonacossi, 1996).
With the structures of the U.S. network empire in mind, we can revisit the large-scale processes and structures of Assyrian imperialism. The main geographic elements of a network empire are easily identified. Nodes in a network consist of "strongholds", called dūru in Akkadian, with smaller forts (birtu) in between. Parker (1997: 77) suggests that the smaller forts were established during incursions into enemy territory and served to build up a network that slowly thickened until complete inclusion in the empire. Thus, he maintains that network zones only temporally preceded territorial expansion. On the other hand, Liverani (1988: 91) states: "I do not believe that it [dialectics between colonization and tribute] is only an inheritance from the formative period, and then the beginning of the break-up; I think it is a structural feature." I suggest that the "provinces" were likely ruled in a way that is structurally closer to a network type of control than territorial power (see below). An Assyrian network zone may be compared to the spatial arrangement of U.S. bases outside the mainland, the setting up of "Forward Operating Bases" (FOB's) and the connecting of them into an increasingly dense mesh of control stations that do not lead to territorial annexation.
Major Assyrian strongholds have an urban character. A relatively large number of them have been excavated, and their Assyrianized culture provides another interesting correspondence to the U.S. network empire's Americanized Green Zones and air bases. A prime Assyrian example for an imperial stronghold in the network zone is the city of Tell Ahmar (Til Barsip) on the left bank of the river Euphrates. This city occupied a strategic position, as this was the point where the army would cross the Euphrates when venturing into regions to the west. Its change from a local multicultural entity (Bunnens, 1995;Bunnens et al., 2006) to a completely Assyrianized fortress could not be more pronounced (Thureau-Dangin and Dunand, 1936). The main reason for the establishment of a stronghold in this place was the river, both a formidable obstacle for the military forces (Lamprichs, 1995: 192-193) and a natural traffic route. Other Assyrianized cities close to the core included Tall Šēḫ Hamad (Dūr Katlimmu) and Tell Ajaja (Šadikanni) on the river Khabur, as well as Tell Rimah (Zamaḫu) in the steppes west of the Tigris and others (cf. Kühne, 1994: 59-60).
A whole system of royal roads or harrān šarri crossed the northern Mesopotamian steppes (Kühne, 1980;Kessler, 1997). These royal roads permitted swift movement of troops and served the expedient connection between network nodes. Special authorization was needed for the use of royal roads. Stations were categorized into posting stations for express communication (kallû) and road stations (bīt mardiāte; Eph'al, 1983).
At the time of the establishment of the Assyrian imperial network, some cities were located in the interstices of the network. For example, the Neo-Hittite city-state of Tell Halaf (Guzana) was Assyrianized only after the ninth century BCE (Kühne, 1998: 285). The Cizre plain, close to the Assyrian heartland, was occupied much later than the area of Tushhan further away in the northwest (Liverani, 1992: Parker, 2003. Meshes in the network zone were large enough to leave some urban zones "inside out:" they were both part of the network zone and still formally independent in their internal affairs. Most of the interstices in the Assyrian network were outside of imperial control. Scholars have explained such interstices as "buffer zones" (e.g. Parker, 2003: 551-553); however, such an interpretation may fall prey to Assyrian propagandistic exaggerations. The empire was not able to control all regions within its network reach. Even the Assyrians themselves left some traces of this situation. A few exceptional texts describe attacks by "nomads" during the drilling of wells in the steppe (Cavigneaux and Ismail, 1990: 335). Territories of control may often have barely reached beyond the immediate surroundings of an urban Assyrianized stronghold, and political voids between such spheres, the lands of guerilla resistance against established powers, must not be underestimated because Assyrian texts silence them. Insurgents in the Pashtun and Baluch territories in Afghanistan, who fought the Soviet invasion in the 1980s just as successfully in the long term as they seem to repel the present Western military, are a distant echo of fierce resistance in peripheral interstices in the past. Similar to the ancient Assyrians, governmental American discourse almost never admits defeat, building instead on a firm rhetorical conviction of final victory.
Overall, Assyrian imperial geographic structures can be described as a network empire with strong parallels to the U.S. empire of today. Of course, similarities exist only at a very general level, in terms of the structural logic of expansion of power, the rhythms of military repression in the imperial periphery, the thickening and loosening of network meshes, the focus on links between network nodes, and the lack of power over network interstices.
It is, however, insufficient to point out structural similarities in order to link two empires that are otherwise deeply divided by space, time, language, technology and identity. The practices of ruling, whether military, administrative, or diplomatic, give more credence to my proposition of a close similarity. Expansion of Assyrian imperial power was carried out by means of military campaigns into uncontrolled lands. The Assyrian term girru, generally translated as (military) campaign, has been understood as suggestive of the beginning of a permanent submission. However, its original meaning, as given in the Chicago Assyrian Dictionary, is related to "way" or "route" (Zehnder, 1999: 124-126) and may in many cases be captured better by English words such as "intervention," or "incursion." Girru is an episodic policing practice by the Assyrian army, carried out on a geopolitical level.
The assumption of an intention to directly administer peripheral regions, even after the administrative reforms of king Tiglat-Pileser III in the mid-8 th cnetury BCE, is likely mistaken. There is just one single document from the whole Assyrian period, the so-called Harran census, which focuses on biopolitics, that is on intricate knowledge about the qualities and quantities of a subject population. It lists villages in a region, but data on demography and labor force are vague. Heads of farmsteads are listed, followed only by the term adi nišēšu, that is, "with his people or family" (Fales, 1973). The administrator is not interested in the labor power behind a productive unit, but only in counting such surplus producing units. Equally interesting is the observation that the Assyrians were sticklers when registering deportees and prisoners of war, but left us almost no information about their fate after deportation (Keskin, 2003). We find a similar lack of biopolitical concern in U.S. relations to its peripheries, most crassly expressed in the refusal to count the victims of its wars in Iraq, estimated by a research group at Johns Hopkins University to amount to more than half a million civilian deaths. In the economic realm, a lack of biopolitical "care" appears in Paul Bremer's infamous "Order 39" from September 2003, privatizing all major state-run compa-nies in Iraq and allowing foreign firms to own them entirely, and to transfer 100% of their profits out of the country, without governmental interference. 14 Assyrian rule does not fit under Foucauldian "governmentality." In my view, Assyrian officials did not in the slightest way develop an attitude of care for their colonial subjects. The notion of governmentality, however, is based on at least a minimal recognition of the subjected as subjects. We find a modicum of "imperial governmentality" in the British as well as the Roman empires (Brandt 2009). The idea of colonial rule was that the colonized should change their subjectivity -they were supposed to take on a metaphorical "white skin" and despite that be ruled under conditions that amount to a harsher version of political conditions in the core. Major efforts had to be made to culturally remold such subjects. British administration in India, the enforcement of English as official language, forced Christianization, a willful re-writing of the history of the colonized through establishment of museums, archives and libraries (Stoler, 2002) are all part of a vast project of destruction of native subjectivity in order to create a new colonial subjectivity. The point here is that a politics of desubjectivation necessarily starts with an assumption of subjectivity among the conquered.
Assyrian, and, I claim, U.S. imperialist policies were and are different in that they never accept the subjected peoples have a subjectivity to start with. Liverani writes that the Assyrians thought of their enemies as "belonging rather to the animal world, and as a matter of fact frequently compared to animals" (Liverani, 1979: 310). The brutality of the imagery from palaces, the vivid scenes of torture and killing confirm such an interpretation (Bersani and Dutoit, 1986;Bahrani, 2008: 154-160). However, too much has been made of the attempt to Assyrianize these enemies after submission (Liverani, 1979: 311-312;Postgate, 1992: 261-262). The United States does not even need an official apparatus to produce a world of apparent species difference between friend and foe. The excesses at Abu Ghraib constitute that sharp, categorical divide, the visual demonstration of a fundamental impossibility to recognize inhabitants of the network's interstices as human beings -even when they are not soldiers.
Only those who were conceptually deemed to be part of the imperial core could in the eyes of the Assyrians, receive a subjectivity, whereas interstitial and deported peoples remained in an object status. There was no real interest in a potential conversion of the conquered into Assyrianized subjectivities. Rather, submission of a region resulted in a material asset, a node in a network. The population in such places was to be kept at bay but not made part of the imperialist dynamic. Conquest was not conceptualized as a conquest of people and their spaces, but as an enlargement of a network through the "knitting" of new nodes and meshes of a network. Conquered peoples are free to adopt (or not) imperial culture, administrative structures and habits, as long as they respond to those demands that come from the imperial nodes. They have the choice between a status similar to a vassal -today, anti-imperialist terminology characterizes them as "puppets", to play a double game to become evaders tout court. Structurally, this leaves the empire's forces more flexibility than in territorial empires. At the same time, those who find themselves in the interstices of such a network had much more leeway to form ad hoc coalitions against imperial forces.
The corresponding practices of power in network empires do not concern the administering of and care for subjected peoples. Rather, all that matters is the functioning of network dynamics, a politics that is best characterized as a regime of "requirementality." By this I mean politics that are based on the needs that are transmitted in a hierarchized network. Requirementality is firmly organized in a top-down fashion and contains five main components: • First, transmission of demands in an imperial network needs to be effective and fast. One of the most important innovations of the Assyrian empire was a messenger system, consisting of royal roads and relays at which mules, a chariot and a driver were always available, should a messenger come by. This harrān šarri system, only accessible to a privileged few, was the backbone of the Assyrian network periphery, and is reminiscent of the global reach of American transport and communication structures such as ARPAnet.
• A second element is what I call, following Münkler (2005: 63-67), "time sovereignty." The decision about what happens when rests always with the superior in a power relation, independently of whether the means to carry out a task is available on the spot. Ultimately in the Assyrian case, the determination of the speed of transmission in a network rests with the king. Communication is part of this. The king's praxis of power is to demand materials, labor, information and services. They are to be realized immediately, but the meaning of "immediately" might change over time. For instance, in king Sargon's mail, the word "right now," annūrig in Akkadian is used very often, implying that the receiver of the royal message had no time to waste. 15 In those instances, the king was careful to draw a fine line between prompt execution of orders and expressions of urgency that might suggest his losing of time sovereignty. The dominance structure of time becomes clear in rare exceptions, such as a letter that orders the impaling of cavalrymen if they arrive late (Parpola, 1987). 16 Time sovereignty is also a foundational element of U.S. executive discourse. It is at the core of G.W. Bush's infamous West Point speech that announced the doctrine of preemptive strikes, shattering the most basic principles of American foreign policy and U.N. laws of war. 17 According to politicians, the reduction of forces in Iraq, the withdrawal from Afghanistan and any other such decisions will not be dictated by American public opinion or legislative bodies, not to speak of Iraqi or Afghan interests. It will only happen, as declared so often especially by the Bush administration, when "the job is done." • The third element of imperial requirementality is its unidirectional cascading demand structure. Demands are sent out to a receiver who is largely free how to delegate them to lower levels or to enact them himself. The only indispensable obligation is to respond to the demand from above, while the means do not matter. This has interesting consequences: letters from the imperial peripheries to the Assyrian king reveal the development of massive avoidance of duty, denunciation and passing on of orders sideways or to lower levels (Parpola, 1987). The result is also an executive structure where a sense of the human dimension of politics is entirely lost. Execution takes on its grim double meaning here, an effect of imperialist practices that have never even conceptualized the colonized and subjected as subjects. The American government's request for information from the periphery displays similar procedures, though the top of the power pyramid is a small network of actors, rather than the single figure of a king. Practices at Abu Ghraib are the best known outcome of a cascading system of demands where responsibility cannot be captured because the superior passes on a demand without specifying the persons or means to carry it out (e.g. Hersh, 2004: 25-45). Private contracting of military provisioning, reconstruction and even battlefield tasks are equally part of such a system (Scahill, 2007). The upper echelons in a regime of requirementality are not interested in which unit or corporation carries out a task, not even necessarily in a full success, but only that it is accounted for as done. In such a system with no shred of mutuality of relations remaining, it may not matter so much whether a task was carried out successfully -even bad military news can be ignored at the top (Hersh, 2004: 285). The terror that results from ruthless political and military actions is sufficient for the success of the upkeep of an imperialist network.
• The fourth effect of "requiremental" practices of power is their unforeseeability. Imperial demands lack the routine and orderliness of a tax system and of a repressive system of colonization that is based on brutal re-subjectivation strategies. In network empires, exigencies can be imposed suddenly. Assyrian mid-level operators developed a keen practical and cynical sense of diverting pressures. Those in the lowest echelons had the most to suffer, being coerced into labor and drafted as soldiers. On this level of "imperial service", strategies of evasion and subterfuge are known and consist of running away, often joining non-sedentary groups. 18 Similarly, recent assessments of U.S. reconstruction in war zones have revealed massive instances of false reporting and corruption. Chatterjee (2004) provides an excellent monographic account of these conditions, whereas Harvey (2005) and Foster (2006) analyze the systemic background.
• If a territorial empire mainly works through an administrative dispositive of power, in the case of network empires, the military prevails. Letters of Assyrian kings, but also the structure of the U.S. occupation of Afghanistan and Iraq follow the logic of a machine of armed repression. Correspondingly, power is exerted by way of sharply categorized, disciplined collectives of people, mainly men, and in the Assyrian case also eunuchs. The expenses of empire do not reside in the problem of infrastructure rebuilding, but in the upkeep of the military apparatus. One major impetus for expansion is thus the reproduction and growth of the repressive apparatus itself, which tends to become both the means and ends of empire (Mann, 2003;Harvey, 2005: 78-81).

Assyria: the Case of the Wadi Ajij
In a comparison of the United States and Assyria, the latter has the benefit of hindsightwe know long-term developments from start to end. In the Assyrian case, we have good evidence that imperial practices of power did not remain constant over the 300 years this empire lasted. It is entirely reasonable to detect historical change in U.S. network imperialism as well. I turn here to some archaeological evidence of Assyria that indicates historical developments within a policy of what I call requirementality. I rely on a survey that I carried out in the Ajij region in eastern Syria close to the Iraqi border in the early 1980s. Geographically close to the Assyrian core, this area is extremely inhospitable due to soil salinization and low precipitation (figure 6). Before and after Neo-Assyrian times, there were never more than four contemporary settlements in this region. However for the Neo-Assyrian period, we identified a total of 33 sites. Wilkinson et al. (2005) and Morandi Bonacossi (2008) report contemporaneous increases in habitation density for other areas in the steppe between the Tigris and Euphrates. For the elucidation of policy changes in the imperial periphery, I focus on the 9 th to early 8 th centuries. To reiterate, historians have viewed this time as one of change from major expansion to an intermediary phase of political weakness, setting in at around 810 BCE (van de Mieroop, 2004). In the 9 th century BCE, the dominant site in the Ajij region was a relay or bīt mardiāte on a route leading from the Assyrian capital on the Tigris through the steppe to the network node of Dūr Katlimmu (modern Tell Šēḫ Hamad) on the river Khabur. At this time, the Ajij region was part of the network periphery, and 95% of the regional population was living in a single site (Bernbeck, 1993: 132).
This picture changes radically in the second phase of the Neo-Assyrian empire. The number of settlements increased from three to twelve and the total settled area doubled. The old central site was abandoned and Umm Aqrubba, a new, equally large one, founded. Judging by its topographic features, it had a fortification wall (Bernbeck, 1993: 181;Abb. 37). This small regional center was home to 68% of the regional population. The remaining third of the settlers lived in scattered small farmsteads. Statistical evaluation of 8 th century settlement size reveals a pattern that differs markedly from a normally developed settlement system (Bernbeck, 1993: 132-135). The second largest site was much smaller than its rank would predict, and there was little difference in size to all remaining sites. According to standard settle-ment pattern interpretations (Smith, 1976: 30), this suggests strong administrative regulation. Nearness to wadi bottoms is a major factor in the location of all sites, another difference to places inhabited in the preceding phase. Wadi bottoms are closest to groundwater, and provide the best agricultural land. An emphasis on farming is also indicated by a fivefold increase in basalt grinding stones (Bernbeck, 1993: 126-127). It is astonishing that this hyperarid, salty depression, whose wells deliver mainly bitter, undrinkable water, was turned into a pocket of cultivation. An inscribed monument from the region, the so-called Saba'a stela, helps to date these events more precisely to King Adad Nirari III (Unger, 1916;Börker-Klähn, 1982: 196).
Adad-Nirari showed a heightened interest in the region between the rivers Tigris and Khabur, as we know from several other stone stelae of his at Tell Rimah (Page, 1968), Dūr Katlimmu and in the Sinjar mountains (Tadmor, 1973). The Rimah stela closely parallels the structure of the Saba'a one, and also mentions military interventions in regions further west (Weippert, 1992). As mentioned previously, many stelae in the Assyrian periphery were set up in imperial interstices. In the case of Adad-Nirari, we can identify the motivation for their establishment. The last part of the text on the Rimah stela describes the foundation of 10 forts and 321 associated farmsteads in the steppe (Page, 1968). This part of the text has been chiseled out for unknown reasons but is still readable. The combination of war and new settlements is linked to the standard Assyrian practice of enemy capture, deportation and forced settlement (Oded, 1979;Wilkinson et al., 2005).
Why would Adad Nirari order the settlement of people in an area at the outermost margins of agricultural productivity? There are two likely reasons. First, he and his forefathers were unable to conquer the region of Suhu on the Middle Euphrates (Liverani, 1988;) -ironically, the same region as the recently rebellious Anbar province in Iraq. Such network interstices close to the imperial core were likely seen as a substantial threat, so that "civilizing" the steppe meant creating a security zone between core and interstices. In fact, that move created a space between core and periphery, a zone that was populated by non-Assyrians but that was economically and militarily no longer part of the network zone. It did not take long for this zone to be included in the core of the empire. Second, provisioning the core with economic surplus was an important function of the periphery. However, in the absence of effective means for bulk transport, it was advantageous to produce agricultural surpluses nearby. Thus, the Ajij region changed into a densely settled enclave that became a western fringe of the Assyrian core, a process built on military violence and deportation.
The survey results in this marginal region indicate the need for a revision of the idea of a crisis of the Assyrian empire in the late 9 th to mid-8 th century. This was rather a phase in which Assyrian kings adjusted an imbalance between a small core and an extensive network periphery through an extension of the imperial core. The structural result was a temporary reluctance to expand the periphery even further. In terms of practices of power, the settlement program of Adad Nirari was a slight shift from an imperial "requirementality" towards a policy somewhat closer to power relations that can be described as effects of a governmentality (figure 7; Bernbeck, 2008).
This reevaluation of the late 9 th to early 8 th centuries BCE as a period of consolidation instead of crisis leads me to a re-assessment of the last imperial phase of the Assyrian empire as well. The more "balanced" imperial structures established by the mid 8th century, with a larger core region, compared to a barely expanded network periphery, allowed another major expansion of the imperial network in subsequent times. This 7 th century expansion was so unrestrained that military overstretch and lack of attention to infrastructural needs in the peripheries led to the collapse and disappearance of Assyria. Attacks by Medes and Babylonians were the historical expression, not the cause of the Assyrian demise. The demise itself may have parallels in other processes of imperial collapse. The difference is that the end of Assyria almost coincides with the forgetting of Assyria (see Yoffee, 2005: 151-155). The reason for the latter is that there were few if any "subjected subjects." The impression of a catastrophe at the end of Assyria is to be sought in a geographical and cultural nexus of expansionary politics that was never even able to see in the conquered populations subjected people.
For the Assyrians, they remained subhuman beings. The lack of recognition on the part of the Assyrians must have been so offensive that the final strike on the haughty imperial center and its lavish palaces went beyond mere destruction. This time around, the enemies took revenge by annihilating one substantiation of the kings and courtiers themselves: they carefully hacked out the eyes and faces of Assyrians rendered in victorious poses on the wall reliefs (Bahrani, 1995).

American Requirementality
How do these observations relate to an understanding of the U.S. empire? The U.S. network empire dates back at least to the aftermath of the Second World War, if not the beginning of the 20 th century (Foster, 2004). But recent changes in U.S. imperial doctrine have radicalized pre-existing dispositions and practices of the empire. The Assyrian parallel helps us to see them in a new light. The end of the Soviet Empire produced a unipolar political constellation, one that structurally furthers tendencies towards what I have called a politico-military requirementality. Under such conditions, military expansion becomes the main practicable policy to increase and/or uphold power in general, at the cost of diplomacy and economic relations (Mann, 2003).
Members of the "Project for the New American Century," such as Donald Rumsfeld, Paul Wolfowitz, Dick Cheney and others have worked since the early 1990s towards a double military doctrine that is very close to the one exhibited during the expansionary phases of Assyrian warfare. They call their model "network-centric warfare" (Fukuyama and Shulsky, 2006;Kagan, 2006: 254-286). This is directly derived from the strategies of business expansion in the information age, and translates issues such as "just-in-time production" (Ohno, 1988) into military strategies. However, their implementation is conditioned on two doctrinal changes in imperialist policies.
First, President Bush's speech at Westpoint in June 2002 and the release of the National Security Strategy document three months later (NSS, 2002) set the stage for new, unilateral imperial ambitions on the part of the United States. The most important innovation was the introduction of preemptive military strikes. The right to attack is no longer based on direct threat by another country but on the development of conditions of such a threat. 19 The consequence is the transition from power through diplomacy to pure military enforcement, from respect of international treaties to realizing imperial interests with a submissive coalition of the willing.
Second, the Bush II-phase of imperial America shares with Assyria the conviction that imperial interests should be achieved by a military that carries out policing operations with a specific goal. In neoconservative strategist Frederick Kagan's words, the wars of the future will be "regime change wars" (Kagan, 2006: 365-374), a terminology that fits perfectly Assyrian strategies. This is the end of the Clausewitzian thesis which defined war as the continuation of politics with other means, implying continued existence of the enemy. In network empires, the point is to attack, install a new government, and establish network nodes to interfere in the larger region whenever deemed necessary. Despite the seeming lack of success of the wars in Afghanistan and Iraq, this strategy resulted in fact in resounding success. The United States military has nowadays a very dense network of interference in a region it considers crucial for its economy. If the "Project of the New American Century" (see Mann 2004) and other conservative military think tank strategists have their way, the future will be one closely similar to Assyria's yearly campaigns (Jarecki, 2008). The Global War on Terror is planned as a series of "netwars," as proposed by military strategists Arquilla and Ronfeldt (2001), or "network centric wars" in the Pentagon's terminology (Alberts et al., 1999). Structurally, such wars are similar to Assyrian interventions in network peripheries. They try to imitate the "hit and run" strategies of their adversaries, albeit at a level of technological sophistication, communicative ability and control of time far beyond that of the empire's enemies (Hardt and Negri, 2004: 53-60). 20 Third, the conquered are not recognized as subjects. The rhetoric of President Bush after 9/11 degraded human enemies in Afghanistan to a bestiary, progressively moving into the realm of lower-order species (Bernbeck, 2002). The Abu Ghraib scandal is only the most drastic sign of a complete denial of the subjectivity of those who survive in interstices of a network (Greenberg, 2006). First-person accounts of the wars in Iraq and Afghanistan tell the same story of an enemy who does not even reach the level of humanity (e.g. Luttrell with Robinson, 2007). The much-touted goal of spreading "democracy" has been revealed over and over as a political smokescreen for the imposition of an imperial will (e.g. Canfora, 2007), for the securing of oil resources and the expanding reach of the military network itself. And while the ruthlessness of an apparatus of power waxes and wanes with larger scale historical processes, the Obama administration's geo-political goals constitute at best a shift, but not a complete reversal of pre-existing mechanisms of imperial control and its extension.
Overall, the United States and ancient Assyria display significant parallels. On a geopolitical plane, they are similar in the network structures of their peripheries, with a set of hierarchized nodes that are islands of "imperial home" in a local sea of foreign culture. Highly developed technologies serve to articulate these nodes. Related to these geographic structures is an imperial power practice and disposition which I have characterized as "requirementality." This comprises effective communication, time sovereignty, a cascading mode of realizing imperial demands, and a high level of contingency in decisions, concomitant with unpredictability on the network's insides, and all the more so in the interstices.
Network empires do not recognize any imperial subjects in the network zones' interstices; at the same time, they develop ideologies of unbounded spatial reach. Compared to the more territorially oriented empires such as the Roman, Chinese or British, these empires produce a paradoxical situation for the conquered. A complete lack of recognition of the subjected as subjects results in often extreme imperial brutality, and opens simultaneously an ideal ground for anti-imperial resistance.
We as archaeologists and historians have so far been preoccupied with discovering "how the system worked." My comparison is no exception to this rule, even if written in a spirit of critique. It remains an urgent task for the future to give the millions of victims of imperialism a voice back they have been violently denied.

Notes
1 Such parallelisms and comparisons in the historical disciplines can be traced back at least to Eduard Meyer who insists on the destructive character of both Rome and America (Demandt, 1997: 174-175). 2 Analogies have been a mainstay in archaeological reconstructions, severely limiting our imagination about the past because of their unidirectional character: the present is a foil for the past (Wylie, 1985). The great advantage of two-way comparisons is their mutuality, a relation that is open for both sides of a comparison to enlighten each other. 3 Wallerstein (1979: 20-24) in his World Systems Theory and scholars who follow this paradigm (e.g. Chase-Dunn et al., 2006) differentiate between core, semi-periphery and periphery. 4 Historically, the United States' expansion in the 19 th century shows all geographical signs of a network empire as well, which through fast expansion of the core and a genocidal war against the native population was turned into the territorial nation-state it is today (Zinn, 2002: 125-148). 5 The National Security Strategy Report of 2002 (NSS, 2002: 29) makes these strategies explicit: "To contend with uncertainty and to meet the many security challenges we face, the United States will require bases and stations within and beyond Western Europe and Northeast Asia, as well as temporary access arrangements for the long-distance deployment of U.S. forces." 6 See also Johnson, 2004: 151-186 for the difficulties of using Pentagon figures to account for the reality of the U.S. military network.