NAXARAR, term given to the para-feudal, social pattern that early Armenia apparently shared with Parthian Iran, although it was preserved into the Sasanian period and beyond. The earliest stages of this system cannot be ascertained with precision given the scantiness of the surviving sources. However, its decentralized character and the linguistic evidence point to an early date near or even preceding the beginning of the Christian era. That evidence was first demonstrated in the extensive works of Emile Benveniste, who linked the majority of Iranian loanwords in classical Armenian to Middle Parthian rather than to Sasanian Pahlavi. The system’s best-known period covers the 4th to 7th centuries CE. Earlier scholars compared this system to western mediaeval feudalism, despite the obvious anachronism of deriving a 4th-century Armenian society from one that flourished in 11th- and 12th-century France, but its parallels are unquestionably found in the aristocratic and essentially non-urban Iranian world.

The locus classicus for the study of the naxarar system is the late 5th century CE anonymous Epic Histories (BP), long mistakenly attributed to an otherwise unknown Pʿawstos Buzand (Faustus of Byzantium), which preserved much of the oral tradition far more accurately than more learned sources (see, e.g., BP 4.3). According to the Epic Histories, Armenian society, as in Iran, was divided into three main estates: the magnates (mecamec naxararkʿ) corresponding to the Iranian wuzurgān, the azatkʿ or lesser nobility, identified by the same word in both languages (Mid. Pers. āzād), and the non-noble (an-azatkʿ) merchants and artisans (iamik) and peasants (šinakan), Iranian vastrōšān “farmers,” who were personally free, though the latter could be bound to the soil. Only the Sasanian fourth estate of the clerks (Arm. dpir) is not attested as such in Armenia, its function possibly assumed by the Christian clergy.

The foundation of both societies was comprised of the great families, of whom some fifty can be identified in the 4th-5th centuries (Garsoïan, 1996, pp. 229-32), and among whom the hereditary king was but primus inter pares. The power of these clans derived from their possession of vast principalities belonging to the “eternal family,” past, present, and future, and of hereditary offices. Neither of these could be alienated, as long as a single male member of the family survived, nor were they derived from the authority or will of the king. Thus, the title and office of commander in chief of the Armenian army or sparapet (< OIr. *spādapati) belonged exclusively to the great Mamikonean house, even when its only representative was a small child unable to assume its functions and temporary substitutes had to be named (BP 3.11; cf. ARMENO-IRANIAN RELATIONS at Not even the king’s manifest will could alter the hereditary nature of these offices, and any royal attempt to tamper with them could result only in tragic failure. The position of the Mamikonean in Greater Armenia, then, was the exact counterpart of that of the Iranian sūren. The very terms used in classical Armenian for these various offices (sparapet, aspet, hazarapet)are unmistakably derived from Iranian terminology, and their hereditary character distinguished them categorically from that of the appointed officials of the Roman bureaucracy. Nevertheless, as demonstrated by Cyril Toumanoff (1963, pp. 33-144), the heads of the great families were simultaneously autonomous sovereign lords or dynasts and royal officials or gorcakalkʿ. In this latter capacity they owed service (caiayutʿiwn),and oaths of fidelity are recorded, although no trace of a feudal homage can be found in Armenia.

So deeply ingrained was the concept of hereditary office, that it manifested itself in the religious sphere as well, even after the christianization of Armenia. The patriarchal dignity passed from father to son in the house of St. Gregory the Illuminator until the death of his last male descendent, in direct opposition to the fourth canon of the Council of Nicaea (325 CE) mandating episcopal elections. Other bishoprics were likewise hereditary (BP 4.12). Here again, this practice seems closer to the Iranian Magian caste than to the imitation of Jewish tradition condemned by the Council in Trullo at Constantinople in 692 (Canon 33).

These noble families were theoretically equal in status, since they belonged to the same estate. In fact, however, they were ranked according to a rigid hierarchy governing the throne (gah; see GĀH i.) or cushion (barz) which wasoccupied by their representatives at court functions and corresponded to the “entrance- and drinking places, sitting and standing places according to the dignity of each man’s rank” set by the king of kings according to the Letter of Tansar (Boyce, p. 44). The precise order of this hierarchy cannot be reconstructed, since the surviving Rank List(Gahnamak)is a late document of dubious authenticity; but it was probably related to the military capacities, recorded in a still later and less reliable Military List (Zōrnamak), of those whom the Epic Histories call “the lords with contingents and banners” (gndicʿ ew drōšucʿ tearkʿ,BP 4.3). The definition given by the Histories for the return to normalcy after a period of crisis is the time when “every magnate was [again] on his throne” (BP 4.2). The Armenian sources underscore the rage of king Aršak and prince Andovk of Siwnikʿ demoted to inferior places unworthy of their rank (BP 4.54; MD 2.50). It is the rage of Rostam, similarly humiliated at the banquet of Esfandiār (Šāh-nāma, ed. Mohl, 4.15). Only the four great hereditary marcher lords or bdeašxs,none of whom seem to have outlived the 5th century, lay outside and above this hierarchy. They may have been descended from the four kings who attended Tigran II (r. 95-55 BCE) according to Plutarch (Lucullus 21.6), but both this derivation and their degree of autonomy vis-à-vis the Armenian king are still uncertain.

The great families, avoiding the cities built on a Hellenistic pattern as a threat to their autonomy, remained entrenched in their distant fortresses, surrounded by their retainers, and were only forcibly brought to the royal court (BP 3.8). Once destroyed by the great Sasanian invasion of 364 CE, these cities were not rebuilt (Garsoïan, 1984-85), thus perpetuating the centrifugal Parthian pattern, rather than adopting the Sasanian system of “royal cities” intended to strengthen the centralizing policy of the king (Gnoli, 1989, p. l57).

Although each great family held all of its property as a unit, it was headed by its senior member (tēr, tanutēr,or nahapet), who led the clan’s military contingent (gund) in wartime and administered, but did not own, its joint property, which he could neither bequeath nor alienate by any other means. All the other members of the clan were equally known as sepuh (BP,pp. 558-59), and the order of succession to the leadership, although usually guided by seniority, does not seem to have been preordained. In addition to these and to the clan’s armed retainers or azat, each major family was also represented by one of its sepuhsserving as clan bishop, and consequently ranked as a naxarar, who signed conciliar acts in the name of his family. As far as can be determined by the sources, the lower clergy did not necessarily belong to one of the clans, and priests were ranked among the azat.

Most of our information concerning the naxarar system derives from studies of the early Christian era, although it continued to exist during the period of Persian domination of Greater or Persarmenia known as the Marzpanateand to some degree in Imperial Armenia despite the social reforms of Justinian (r. 527–65) (Adontz, 1970, pp. 127-64; Garsoïan, in press). Despite the gradual disintegration of the naxarar system, some of its aspects can be observed throughout the periods of Arab and Mongol domination on the Armenian plateau; and, in the ecclesiastical sphere, traces may have survived as late as the 19th century. References to occasional innovations, such as references to a “senior” (mec or awag) sepuh can be found in some sources (BP, pp. 545-46). An attempt to revive the system was made in the late 12th century under the Georgian Zakʿarid viceroys of Armenia, but no systematic study of the development and survival of the post-classical naxarar system after the 7th century exists to date.



Sources. M. Boyce, tr., The Letter of Tansar, Rome, 1968.

BP = The Epic Histories attributed to P ‘awstos Buzand (Buzandaran Patmutʿiwnkʿ), tr. and comm. N. G. Garsoïan, Cambridge, Mass., 1989.

ŁPʿ = The History of Łazar P’arpec’i, tr. and comm. R. W. Thomson, Atlanta, 1991.

G. Nedungatt and M. Featherstone, eds. and tr., The Council in Trullo Revisited, Rome, 1995.

Plutarch, “Lucullus,” in Plutarch’s Lives II, Loeb Classical Library, Cambridge, Mass., and London, 1948; 2nd repr., pp. 469-611.

Literature. N. Adontz, Armenia in the Period of Justinian Political Conditions Based on the naxarar System, ed. and tr. N. G. Garsoïan, Louvain and Lisbon, 1970.

N. G. Garsoïan, “Prolegomena to a Study of the Iranian Element in Arsacid Armenia,” in Handes Amsorya 90, Vienna, 1976, cols. 177-234; repr. in idem, Armenia Between Byzantium and the Sasanians, London, 1985, pp. 1-46.

Idem, “The Locus of the Death of Kings. Armenia the Inverted Image,” in R. Hovanissian, ed., The Armenian Image in History and Literature, Malibu, Calif., 1981, pp. 27-64.

Idem, “The Early-Medieval Armenian City-An Alien Element?” in Ancient Studies in Memory of Elias J. Bickerman (= JNES16-17, 1984-85), 1987, pp. 67-83; repr. in Garsoïan, 1999.

Idem, “Notes préliminaires sur l’anthroponymie armenienne au moyen-age,” in M. Bourin, J.-M. Martin, and F. Menant, eds., L’anthroponymie document de 1’histoire sociale des mondes méditerranéens médiévaux, Rome, 1996, pp. 227-39.

Idem, Church and Culture in Early Medieval Armenia, Aldershot, U.K., 1999.

G. Gnoli, The Idea of Iran, Rome, 1989.

C. Toumanoff, Studies in Christian Caucasian History, Washington, D.C., 1963.

(N. Garsoian)

Originally Published: July 20, 2005

Last Updated: July 20, 2005