iii. Safavid Period
1. Under Shah Esmāʿīl and Ṭahmāsp (1501-76). After leaving his refuge at Lāhīǰān, Shah Esmāʿīl, the spiritual director of the Safavid order and founder of the Safavid state, recruiting followers en route (already 450 at Rašt (Afżal al-tawārīḵ 1,60b) and 1,500 at Ṭārom (Ahṣan al-tawārīḵ p. 41) reached the pastures of Arzenǰān in the summer of 906/1500. It is precisely here that 7,000 adherents, mainly Turkmen tribesmen from the provinces of Anatolia and Syria, gathered around their moršed-e kāmel (king). Although Esmāʿīl’s predecessors had also formed troops of their followers, it is certain that the prototype for the military forces of the Safavid state was formed at that particular moment, for the partisans of the Safavid order had been dispersed after the defeat of Solṭān ʿAlī, Esmāʿīl’s brother, in 899/1494. This army had several typical tribal characteristics:
(1) The army was comprised of tribal units, the majority of which were Turkmen, the remainder Kurds and Čaḡatāy. Gradually, as the conquest progressed, the heads of these units would be given a city or even a province as a kind of fief. One may thus view this army as a tribal confederation, similar to the Turkmen dynasties of the fifteenth century. The weapons of the cavalrymen who formed the main battalions were traditionally sabers, lances, and bows and arrows (šamšīr, neyza, tīr o kamān).
(2) The army comprised, in theory, a center and two wings, the left and the right. For battle, an advance and a reserve force would occasionally be assigned. The shah and his non-tribal retinue would always occupy the center. Each of the tribes making up a confederation would have a set position in one of the two wings not only during military campaigns and on the battlefield but also on public occasions such as official assemblies, royal banquets, etc. The long tradition in the Turco-Mongol nomadic world lived on (Della Valle, pp. 350-51; Haneda, Le système, pp. 36-47).
(3) The shah had a royal bodyguard of qūṛčīs (Mongolian “archer”) whose origin goes back to the Mongol period. The qūṛčīs were, in theory, recruited from among the principal gezelbāš tribes and received their pay directly from the royal treasury. This corps was particularly important during the Safavid period, given the fact that it was the shahs’ only personal corps. At the head of each group of qūṛčīs, originally all from the same tribe, was a yūz-bāšī (centurion). In command of the entire force was the qūṛčī-bāšī. The effective strength of this corps was about 3,000 men under Esmāʿīl and 5,000 under Ṭahmāsp (Haneda, L’évolution, pp. 41-50).
(4) There were several other offices traditionally found in the Turco-Mongol steppe such as tūvāčī, amīr-e šekār, etc.
As far as the actual size of the army under Esmāʿīl, one finds several different figures ranging from 7,000 at the battle against the Šervānšāh in 1500 to 40,000 according to Mīrzā Ḥaydar’s description of the battle between Esmāʿīl and Šeybānī Khan Uzbek in 916/1510 (Tārīḵ-e Rašīdī, English translation by N. Elias and E. D Ross, London, 1895, p. 234). Two years after the battle of Čālderān, the entire army of the shah probably amounted to no more than 18,000 men, of whom only 10,000 were really combat-ready, according to a report drafted 20 Jomādā II 922/21 July 1516 of a Safavid deserter to the Ottomans (Bacqué-Grammont, Ottomans et Safavides, pp. 204-08).
The three titles of honor for the military aristocracy were khan (the highest rank), sultan (the next highest), and beg (the third). The number of those who held the title “khan” was quite small at first but seems to have increased with time.
It is true that in the course of the conquest of Iran, diverse non-tribal elements (Iranian, for example) joined the Safavid army. But its fundamental structure did not change until the beginning of the reign of Shah ʿAbbās I (1587-1629). There is a good example which proves this, a military review at Besṭām in the summer of 936/1530 of virtually the entire army of Shah Ṭahmāsp (Ḵolāṣat al-tawārīḵ, pp. 198-204). On that occasion, the tribes marched before the shah, one after the other. Of a total of 105,800 men, certainly an exaggerated number (it includes non-combatants such as religious scholars and civilian officials as well), 84,900 belonged to tribes. It is certain that the troops coming from the tribes always made up the greater part by far of the Safavid army.
One should point out however a new aspect in the army of Ṭahmāsp emerging during this same review and that is the appearance of many types of artillery (tūp, farangī, bādlīǰ, and żarbzan) as well as three or four thousand arquebusiers (tofanġčīān). Considering the youthful age of the shah (16 years) and the confused political situation of this period, this modern corps armed with firearms is without doubt a legacy from Shah Esmāʿīl who had begun to farm a genuine corps of artillerymen and arquebusiers in 1516, after the debacle at Čālderān (Bacqué-Grammont, Ottomans et Safavides, p. 193). It is possible to consider this corps, which was equipped with firearms, as the embryo of the more important and systematic corps organized later by Shah ʿAbbās I, the tofanġčī corps and the tūpčī corps.
2. Shah ʿAbbās and his military reforms. No one has clearly indicated both aspects of the military measures adopted by Shah ʿAbbās. On the one hand, from the perspective of domestic politics, they aimed at reducing the qezelbāš element which had played a predominant role in all aspects of Safavid politics; on the other, from the standpoint of foreign policy, they aimed at creating a corps equipped with modern weapons in order to confront the formidable Ottoman forces. The reinforcement of the qūṛčīs (the royal bodyguard) and the creation of a corps of ḡolāms pertain to the first, while the organization of the corps of tūpčīs and tofanġčīs was part of the second category.
a) Strengthening the corps of qūṛčīs.
Minorsky’s classic description of the Safavid army after Shah ʿAbbās is based mainly on European sources and does not always reflect reality of the army of the period; most important, Minorsky has confused qūṛčī with qezelbāš. In fact, while the term qūṛčī designates the royal bodyguard, qezelbāš signifies the tribal elements in general who had backed royal power at the foundation of the state.
The number of qūṛčīs (10,000-15,000) at least doubled under Shah ʿAbbās. As a consequence, the importance of the social standing and the office of the qūṛčī-bāšī also increased. The shah rewarded the qūṛčīs, who remained loyal to him in the period before and just after his accession to the throne. Moreover, he placed several of them in other important positions such as governors of the large provinces, īšīkāqāsī-bāšī, etc. That entailed the deterioration of the power of the qezelbāš amirs, who had a monopoly on the high offices of state. Although the qūṛčīs had been recruited from the qezelbāš tribes and in that sense had constituted an indispensable element of the tribal confederation which characterized Safavid society in the period preceding the reign of Shah ʿAbbās, one should not confuse that regular corps with the qezelbāš, the tribal elements as a whole. By favoring one of the elements of the tribal society the shah intended to reduce the influence of the whole. By the end of Shah ʿAbbās’ reign, the qūṛčī-bāšī was the most powerful amir of all the leaders of the state.
b) The creation of the ḡolām corps.
Babaev (p. 23) attributes the date of the creation of the corps of ḡolāms (made up of Armenians, Georgians, Circassians, etc.) to the period just after Shah ʿAbbās’ accession. However, the sources on which this is based only mention the reorganization of the army in general. As far as we know, no source precisely dates the actual creation of the corps of ḡolāms, and it is quite certain that Shah ʿAbbās set the reform in motion immediately after the assassination of Moršed-qolī Khan, his lala (tutor) and the leading qezelbāš personality in 997/1588-89, for in the narrative of the events of 998/1589-90 in Rawżat al-ṣafawīya, there is mention of the nomination of a certain Yol-qolī Beg to the post of chief of the ḡolāms, the qollarāqāsī (fol. 292b). The nomination may be considered a sign of the first attempt at reform by Shah ʿAbbās.
No one doubts the important role which Shah ʿAbbās played in the systematic establishment of the ḡolām corps. Nonetheless, the term qollarāqāsī is already found in the narrative of the events of 991/1583-84 (Rūz-nāma-ye Monaǰǰem Yazdī 46a) and refers to the participation of the ḡolāms in battle as a royal guard (ḡolāmān-e ḵāṣṣa-ye šarīfa) under Sultan Moḥammad Ḵodābanda (1578-87) (e.g. Eskandar Beg, p. 338, Rūz-nāma, 46a). Since it is difficult to imagine that the ḡolām system was evoked during the ten years of the confused reign of Moḥammad Ḵodābanda, one may surmise that it had already reached a certain level under Shah Ṭahmāsp; note especially Ṭahmāsp’s successive military campaigns towards the Caucasus beginning 947/1540-41.
The effective strength of the cavalry corps under ʿAbbās was probably 10,000 or 15,000 men, as described in all the European sources. The weapons of both the ḡolāms and the qūṛčīs were the traditional ones. Numerous ḡolāms were appointed to high positions of state and at the end of Shah ʿAbbās’ reign almost all the important offices (aside from the civil and fiscal offices traditionally given to Iranians) were occupied by former ḡolāms or qūṛčīs. Backing the shah both politically and militarily, the regular corps of ḡolāms and qūṛčīs were the two pillars of the state after the reforms of Shah ʿAbbās.
c) The organization of the tofanġčī and tūpčī corps. As mentioned above, these two corps had been established long ago. But it was Shah ʿAbbās especially who made use of the modern weapons of these corps to reconquer lost territories. The tofanġčīs were recruited from different regions and organized under the name of their region, e.g. “tofanġčīān-e Esfahān, tofanġčīān-e Māzandarān,” etc. At the head of each group a mīn-bāšī (chief of a thousand) was appointed. No tofanġčīāqāsī is mentioned in the chronicles during Shah ʿAbbās’ reign. Neither the corps nor any of its members had any political importance to compare with the ḡolāms.
As for the tūpčīān, there is very little information in the Persian chronicles, despite some references to them by Europeans (Taḏkerat al-molūk, p. 33). Neither the structure of the corps, nor its actual strength under ʿAbbās is known. Despite its effectiveness against the Ottomans, the importance of the corps within the army was not very considerable. Although Minorsky saw a purely Iranian element in these corps equipped with modern weapons (Taḏkerat al-molūk, p. 32), yet one does find certain links between them and the ḡolām element. It was a ḡolām, Qaračqāy Beg who led this corps against the Uzbeks (Eskandar Beg, p. 620) and Ottomans (ibid., p. 697) under Shah ʿAbbās. In the period after Shah ʿAbbās there were also tofanġčīāqāsīs of ḡolām origin. (Ḏayl-e tārīḵ-e ʿālamārā-ye ʿabbāsī, p. 269).
Through the measures adopted by Shah ʿAbbās, the tribal forces who had been led by the qezelbāš amirs, were placed under the command of the provincial governors, for the most part ḡolāms or qūṛčīs. They lost their importance, their identity, and even their particular names. They also ceased to be called qezelbāš (Haneda, L’èvolution, p. 56).
Thus the reform of Shah ʿAbbās contributed in the short term to weakening qezelbāš influence and to strengthening central power. But in the long term, by crushing the base of tribal support which had allowed Shah Esmāʿīl to accede to the throne, the reform was to cause the collapse of the dynasty.
A. Sources. As there is no manual for the Safavid army, we must piece together its description from Persian and European sources.
Persian sources: Ḥassan Rūmlū, Aḥsan al-tawārīḵ, ed. ʿAbd-al-Ḥosayn Navāʾī, Tehran, 1357 Š./1978.
Qāżī Aḥmad Qomī, Ḵolāṣat al-tawārīḵ, ed. E. Ešrāqī, Tehran, 1358 Š./1979.
Jalāl-al-dīn Monaǰǰem Yazdī, Rūz-nāma-ye Monaǰǰem Yazdī, Ms. B.M. Or. 6263. Fażlī al-Rūzānī, Afżal al-tawārīḵ. 1) Ms. Library of Eton College, no. 172. 2) Ms. B.M. Or. 4678. Mīrzā Beg Jonābādī, Rawżat al-ṣafawīya, Ms. B.M. Or. 3388. Eskandar Beg Monšī, Tārīḵ-e ʿālamārā-ye ʿabbāsī, ed. Īraǰ Afšār, Tehran, 1334 Š./1955. Idem, Ḏayl-e tārīḵ-e ʿālamārā-ye ʿabbāsī, ed. S. Ḵᵛānsārī, Tehran, 1317 Š./1938.
European sources: J. Chardin, Voyages du Chevalier Chardin en Perse, 10 vols., Paris, 1811.
E. Kaempfer, Am Hofe des persischen Grosskönigs (1684-85), Leipzig, 1940 (repr. Tübingen, 1977).
R. Du Mans, L’Estat de la Perse en 1660, Paris, 1890.
J. B. Tavernier, Les six vovages . . . qu’il a fait en Turquie et Perse aux Indes, 2 vols., Paris, 1676.
P. Della Valle, I viaggi, F. Gaita et L. Lockhart, Rome, 1972.
B. Studies. K. Babaev, “Voennaya reforma Shakha Abbasa I (1587-1629),” Vestnik moskovskogo universiteta, Vostokovedenie, 1973-71, pp. 21-29 is a good synthesis of Minorsky’s work and Falsafī’s monograph (N. Falsafī, Zendagānī-e Šāh ʿAbbās-e awwal, 5 vols., Tehran, 1956-1963).
J.-L., Bacqué-Grammont, Ottomans et Safavides au temps de Šāh Ismāʿīl (doctoral thesis for the University of Paris I, unpublished) is useful for the army of Shah Esmāʿīl (to be published in 1986 as Les Ottomans, les Safavides et leurs voisins. Contribution à l’étude des relations internationales dans l’Orient islamique de 1514 à 1524).
M. Haneda, Le système militaire safavide. Le chah et les Qizilbāš (doctoral thesis for the University of Paris III, 1983, unpublished). Idem, “L’évolution de la garde royal des Safavides,” Moyen Orient et Océan indien I, 1984, pp. 41-64.
L. Lockhart, “The Persian Army in the Safavid period,” Der Islam 34, 1959, pp. 89-98 (the only article fully devoted to the Safavid army, but now outdated). In spite of some defects, the brief study by V. Minorsky, Tadhkirat al-mulūk, London, 1943, pp. 16-19, 30-33, is still valuable for a general view of the Safavid army.
K. M. Röhrborn, Provinzen und Zentralgewalt Persiens im 16 und 17 Jahrhundert, Berlin, 1966, is an excellent study not only of the military history, but also of the political and institutional history of the period.
Originally Published: December 15, 1986
Last Updated: August 12, 2011
This article is available in print.
Vol. II, Fasc. 5, pp. 503-506