255
CHAPTER 1
TACTICS TECHNOLOGY AND MILITARY ORGANIZATION IN PRE-ISLAMIC ARABIA
Of all those areas that were to become parts of the Muslim world, less is known about military technology and organization in the original Arab homeland than anywhere else, with the possible exception of Berber North Africa. Pictorial representations are virtually unknown from the Arabian peninsula itself, while the dating of much supposedly pre-Islamic verse is highly debatable. Even those verses that almost certainly came from the Jāhīlīyah are probably so late that they should most usefully be regarded, along with the rest of such poetry, as illustrating the cultural situation in which Islam was born, rather than the general background of the Arabian peninsula as a whole.1
Descriptive material from neighbouring, or only partially Arab, cultures may be the most reliable when studying the Arabs' general background from the standpoint of military technology. These sources indicate that some widely accepted views portray the Arabs as more backward than they really were. It has, for example, quite recently been maintained that they knew neither the coat of mail nor the helmet.2 Others overemphasize one external influence. Thus some say that the military traditions of northern Arabia were thoroughly Byzantine like Syria and Egypt,
------------------------
1. U. Montgomery Watt, The Majesty that was Islam, (London 1974), pp. 77-78 and 90-92; C. Pellat, "Jewellers with Words," in The World of Islam, B. Lewis edit., (London 1976), pp. 143-145.
2. A. S. Melikian-Chirvani, Le Roman de Varqe et Golšâh, Arts Asiatiques, XXII (1970), pp, 38-39.
256
with short lamellar and scale cuirasses, mail hauberks and round shields, and that Hellenistic mail armours were common in central Arabia,3 or that the Arabs of western Arabia, Yemen, Ḥijāz and the Syrian frontiers fought in the Sassanian style with helmets, mail hauberks, long straight swords and lances.4 Some tend to oversimplify by apparently relying on one source to emphasize the predominant role of archery among the ancient Arabs,5 or by maintaining without qualification that Arab archery was inferior to that of their foes.6
Surprisingly, perhaps, one must often return to earlier scholars such as A. van Kremer to find a more reasoned analysis of the military capabilities, equipment and organization of pre-Islamic Arabia.7 Yet even here, I believe, there remains a tendency, encouraged by the Muslim's own view of his cultural origins, to portray the pagan Arab as excessively "simple," that is backward.
The nomadic, semi-nomadic and settled Arab tribes of northern Arabia and the Fertile Crescent were almost certainly more sophisticated
------------------------
3. M. V. Gorelik, "Oriental Armour of the Near and Middle East from the eighth to the fifteenth centuries as shown in works of art," in Islamic Arms and Armour, R. Elgood edit., (London 1979), p. 31.
4. M. Lombard, Les métaux dans l'ancien monde du Ve au XIe siècle. (Paris 1974), p. 35.
5. D. J. F. Hill, "Some Notes on Archery in the Roman World," Journal of the Society of Archer Antiquaries, I (1958), p. 2.
6. D. D. Latham, "The Archers of the Middle East: The Turco-Iranian Background," Iran, VIII (1970), p. 97.
7. A. von Kremer, Culturgeschichte des Orients unter den Chalifen, I, (Vienna 1875), p. 78.
257
militarily than their southern cousins, while their material culture was in general a poorer version of that of their other neighbours. What has, however, rarely been emphasized is the degree of contact between north and south within the Arab area.8 Similarly, it would be hard to overstate the vital military role of those northern tribes in warfare between the Hellenistic and Iranian worlds. Most of the detailed references come from the Romano-Byzantine side, where six cohorts of presumably Arab infantry archers, the Ulpiae Petraeorum, are among the earliest to be mentioned.9 By the 4th century light cavalry of the Bakr and Taghlib tribes were playing a far more prominent, though occasionally equivocal, role between the Sassanian frontier and the Byzantine limes in the Jazīra area.10 Arab troops, in particular archers, were also recruited to defend these Byzantine limes, and even to command in such positions.11 By the 7th century Arabs, Christian and pagan, could form the bulk of a Byzantine army such as that at Mu'ta in 629 AD.12 One possibly apocryphal source even has a
------------------------
8. K. S. al Asali, South Arabia in the 5th and 6th centuries CE with reference to relations with Central Arabia, (Unpub. Ph. D. thesis, St. Andrew's Univ. 1968); I. Shahid, "Byzantino-Arabica: The Conference of Ramla, AD 524," Journal of Near Eastern Studies, XXIII (1964), p. 130.
9. Von Hubert van de Weerd and P. Lambrechts, "Note sur les corps d'archers au haut Empire," in Die Araber in der Alten Welt, I, F. Altheim and R. Stiehl edits., (Berlin 1964), p. 663.
10. F. Altheim and R. Stiehl, "Die Araber an der oströmisch-persischen Grenze im 4. Jahrhundert," in Die Araber in der Alten Welt, II, (Berlin 1965), pp. 324-327.
11. F. Altheim and R. Stiehl, "Šāpūr II und die Araber," in Die Araber in der Alten Welt, II, (Berlin 1965), p. 344.
12. A. J. Butler, The Arab Conquest of Egypt, (Oxford 1982), pp. 150-151.
258
Christian Arab in command of a mixed Coptic, "Roman" and Arab force defending Byzantine Tinnīs during the Muslim invasion of Egypt.13
Such evidence might suggest that Arab troops could be more than an ill-equipped rabble of tribal auxiliaries. Nevertheless, this same evidence might also indicate that the Arabs were not a homogeneous whole as far as their military capabilities were concerned. Clearly one must specify just whom one is considering in those culturally and linguistically mixed areas bordering the Fertile Crescent. This is particularly true of earlier centuries when the Arabization of the region had hardly begun. Palmyra and Dura Europos, from the 3rd or 4th centuries, could be seen as Aramaic islands13A in an Arab sea, but although their armed forces were partially, perhaps even predominantly, recruited from surrounding Arab tribes, they should still be regarded as Syrian.
Hatra, by contrast, has been considered both as essentially Arab and as a melting pot of the military traditions of the Semitic east.14 Here, in the 2nd century AD, there were sophisticated fortifications manned by townsfolk and neighbouring semi-nomadic peasantry. In open battle, meanwhile, bedouin horsemen could drive back Roman cavalry while bedouin archers, many of whom were horse-archers, threatened the life of Trajan himself.15 Clearly the Arab peoples of these areas were ideally placed to learn from a variety of sources. The Nabateans, for example, probably adopted
------------------------
13. Ibid., p. 353.
13A. F. Altheim and R. Stiehl, "Odainat und Palmyra," in Die Araber in der Alten Welt, II, (Berlin 1965), pp. 269-270.
14. F. Altheim and R. Stiehl, "Hatra," in Die Araber in der Alten Welt, IV, (Berlin 1967), p. 244.
15. Ibid., p. 249.
259
the war-horse from their Syrian neighbours. As early as the 1st century BC they had evolved their own essentially bedouin tactics of repeated attack and retreat, karr wa farr, using both camels and cavalry, though such a strategy may also reflect Parthian influence.16
The Ghassānids and Lakhmids were the last in a long line of client Arab states to be involved in the age-old frontier struggle between Hellenism and Iran, yet even they appear to have remained dependent on their respective protectors for much war material. It has been suggested, that the Byzantines merely loaned the farmer full military equipment in time of war, otherwise storing it in their arsenal at Buṣrā.17 The Sassanians apparently did much the same for the heavy cataphract cavalry of the Lakhmids, drawing upon their arsenals at cUkbarā' and Anbar.18 These arrangements may account for an apparent contradiction between the Arabs' military capabilities in some spheres and their supposedly limited skills in such areas as siege warfare. Belisarius himself stated that the Arabs under his command were unable to build fortifications,19 whereas the Arab rulers of Ḥīra were already known to be using complicated siege engines in the 3rd century.20 Far to the south, in Yemen, the Sabaean kings were also known to have built fortresses
------------------------
16. Ibid., pp. 296-297; E. Merkel, “Erste Festsetzungen im fruchtbaren Halbmond,” in Die Araber in der Alten Welt, I, F Altheim and R. Stiehl edits., (Berlin 1964), pp. 299-301.
17. Altheim and Stiehl, "Dhū Nuwās," in Die Araber in der Alten Welt, V/1, (Berlin 1968), p. 316.
18. Altheim and Stiehl, "Šāpūr II und die Araber," pp. 351-352.
19. J. R. Partington, A History of Greek Fire and Gunpowder, (Cambridge 1960). p. 186.
20. Ibid., p. 189.
260
a century or so later.21
Certain specifically, though not solely, Arab military characteristics were, however, already appearing. These would be seen again during the first wave of Muslim expansion. Hatra and Ḥīra, for example, developed from nomadic encampments and may be seen as the prototypes of Kūfa, Qayrawān, Baghdād and Sāmarrā.22 There was also an exceptional emphasis on ruses and tactics,23 a desire to exercise maximum force and at the same time refrain from resisting "à l'outrance" in the face of impossible odds,24 all of which indicate an, eminently rational attitude towards warfare. Finally there was the primacy of the bow among the desert Arabs who were described by a Syrian, probably in the 5th century, as "Saracens who pass their life in archery and raiding."25 But were these raiders infantry or cavalry, and if the latter did they shoot at rest, like Byzantine horse-archers, or in motion like the Parthians? Typically, perhaps, such contemporary descriptions pose as many questions as they answer.
Though they are largely seen through the eyes of their more settled neighbours, we can thus build up a picture of these north Arabian warriors. Far less may be said of central and southern Arabia. Nevertheless, all these areas were in sufficiently close
------------------------
21. J. Ryckmans, L'Institution Monarchique en Arabie Meridionale avant L'lslam. (Louvain 1951). p. 231.
22. F. Altheim and R. Stiehl. Die Araber in der Alten Welt, III, (Berlin 1966), p. 7.
23. H. Lammens. "Les Aḥābīs et l'organisation militaire de la Mecque au Siècle de l'Hégire," Journal Asiatique, 11 ser. VIII (1916), p. 432.
24. Montgomery Watt, op. cit., pp. 32-34.
25. A. H. M. Jones. "Asian Trade in Antiquity," in Islam and the Trade of Asia, (Oxford 1970). p. 6.
261
contact for them to have probably shared many military traditions. The Nabateans were, of course, vitally concerned with trans-desert trade routes from the Gulf, while Palmyra's links with central Arabia were of long standing.26 Similarly, Lakhmid l Ḥīra played a vital role in Arabian politics, having a direct link across the peninsula to the Ḥimyarite south. Ḥīra was, in fact, a centre through which outside, that is Iranian, cultural and military influences penetrated south.27 Such cultural contacts may have been reinforced during an apparent Yemeni domination of central Arabia and even, perhaps, the entire Gulf area in the early 4th century.28 Iranian military influences would similarly have been felt during the subsequent Sassanian counter-attack under Shāpūr II.29 One feature does, however, differentiate the armies of northern Arabia from those of the southern and perhaps also central regions. This was the south's lack of cavalry. Although under Sabaean rule Yemeni kings employed regular troops and auxiliary contingents from both south Yemeni and central Arabian tribes,30 the limited available inscriptions refer only to infantry and camel-mounted warriors.31 Indeed one Yemeni leader was, in desperation, obliged to request a force of asāwira heavy cavalry from the Sassanian king.32 Such Sassanian heavy cavalry were, in fact, stationed in
------------------------
26. Al Asali, op. cit., p.87.
27. Shahīd, loc. cit.
28. Al Asali, op. cit., pp. 52-55.
29. Ibid., pp. 55-59.
30. Ryckmans, op. cit., pp. 134 and 231.
31. Al Asali, op. cit., pp. 101-102.
32. Ibid., p. 304.
262
Oman. The arrival of Iranian asāwira and marāziba, or "marcher lords" followed Khusrau Ānūshīrvan's reorganization of the Sassanian province of Mazūn, that is the coastal and settled parts of Oman, in the 6th century. How far this class of Persian landed military aristocrats influenced the local Arab aristocracy is unclear. What is known, however, is that the julandā,or local Arab leadership, survived the collapse of Sassanian authority and continued to dominate Oman until the end of the 6th century.33 The least one may assume is that here was another important and obvious channel via which Sassanian military technology probably filtered into Arabia.
Another interesting, and in a way complementary, north-south link was that recorded between the Arab Ghassānids and the Ethiopian conquerors of Yemen,34 both of which peoples were of course within the Byzantine sphere of influence. Both also employed cavalry with great success against their central and southern Arabian foes, the Ghassānids in particular being famed as cavalry and as mobile camel-mounted infantry.35 Ethiopian and Nubian peoples across the Red Sea may, for various reasons, be seen as a rather distorted mirror of their Arab neighbours in the Yemen and Ḥijāz. Because of its close political links with Byzantine, we often know more about the military traditions of the western coast of the Red Sea than we do of the
------------------------
33. J. C. Wilkinson, "The Julanda of Oman," Journal of Oman Studies, I (1975), pp. 99 and 103-104.
34. I. Shahīd, "Ghassān," Encyclopedia of Islam, second edition, II, pp. 1020-1021.
35. D. R, Hill, The Mobility of the Arab Armies in the Early Conquests, (Unpub. M. Litt. thesis, Univ, of Durham 1963), p. 19; F. Altheim and R. Stiel, "Das christliche Aksūm," in Die Araber in der Alten Welt V/2, (Berlin 1969), p. 232; Shahīd, "Ghassān," loc. cit.
263
eastern. Like Arabs, the Ethiopians were regarded by Pliny the Elder primarily as archers.36 The Bāja and nomadic northern Ethiopians also used large numbers of camels in their warfare,37 while according to Arab tradition the tribes of Yemen and Ḥijāz learned how to use the bow from the Nubians.38 As mentioned above, the Ethiopians apparently made extensive use of cavalry during their operations in the Yemen. It has also been suggested that many of their warriors remained in the area to serve as mercenaries after the Arab and Sassanian counter-attack, though whether as infantry or cavalry is unknown.39 This introduces the difficult question of the aḥābīsh. Were the soldiers so-named in the Ḥijāz Arabs or Ethiopians? In Yemen they were apparently the latter.40 Whatever the answer, it is likely to reinforce further the contention that the military traditions and combat styles east and west of the Red Sea had much in common.
Competing with the Ethiopians for the control of Yemen, and perhaps also having its own influence on southern Arabian military traditions, was Sassanian Iran. While there is come doubt that the above-mentioned asāwira (Pahlavi: asvārān) heavily armoured Persian cavalry actually reached the area, Dayiamī infantry certainly did arrive.41 At that period, as later, these troops had a fine reputation
------------------------
36. D. H. F. Hill, loc. cit.
37. Altheim and Stiehl, "Das christliche Aksūm," p. 226.
38. Al Maʾūdī, op. cit., vol II, pp. 382-303.
39. Lammens, loc. cit.
40. Lammens, op. cit. pp. 433 and 441-442; W. Montgomery Watt, "Aḥābīsh" Encyclopedia of Islam, second edition, III, pp. 7-8. 41. A. Christensen, L'Iran sous les Sassanides, (Copenhagen 1936), p. 362.
264
and fought with the weapons that they were later to make famous throughout much of the Muslim world.42
While documentary evidence concerning pro-Islamic Arab warfare is generally sparse, purely descriptive references to the warriors involved are even more so. One such status that the Saracens were "long-haired".43 This is of some use when identifying possible Arab warriors in the limited available works of art, as most other warriors appear to be closer cropped! In specifically Arab regions such as Hatra and south Arabia, similarities of hair-styles may only be useful as evidence of cultural links between otherwise barely related peoples (Figs. 1, 27 and 28). Where, on the other hand, there is some doubt as to the nationality of the subject, as in 3rd century Dura Europos (Fig. 42), or where a story such as the "selling of joseph" on a 6th century Byzantine ivory throne might already suggest an Ishmaelite Arab subject (Fig. 23), such a written description may be used as further evidence that the long-haired subject is an Arab. Elsewhere it can only suggest that the figure in question could represent an Arab (Fig. 21).
It has also been suggested that the wearing of two swords, or of a sword and a dagger, was characteristic of the pre-Islamic Arabs.44 While it may indeed have been typical of much of the Arab area, appearing at Hatra and elsewhere, the fashion itself is, I believe, of Iranian origin and was widespread throughout the east. Hence it is of little value when attempting to identify a subject.
------------------------
42. C. E. Bosworth, "Military Organization under the Būyids of Persia and Iraq," Orians, XVIII-XIX (1965-1966), p. 147.
43. Butler, op. cit., pp. 15-151.
44. H. Seyrig, "Armes et costumes iraniéns de Palmyre," Syria, XVIII (1937), p. 30 and 30 n.9.
265
There are indeed few pictorial sources that can definitely be said to show Arab warriors, except some in Yemen. Those of Hatra are overwhelmingly Iranian in style. One may, however, assume that this was equally so of Hatra's military equipment, at least that which was used by an élite who are probably reflected in the surviving sculptures. Here swords are long and straight, with long, simple or flared hilts. They are generally slung from a loose sword-belt, often with a characteristic Iranian scabbard-slide (Figs. 24, 29 and 30). Rarely does the Mediterranean-style baldric appear (Fig. 26), though even here it carries an Iranian-type sword. Surviving weapons are limited to some very rusted 2nd century iron spearheads, plus bronze arrowheads of types A, B, G, M and T, two- and three-bladed, both tang and socket types. A horse-bit also excavated at Hatra is, perhaps significantly, of the snaffle-psalion type at that time generally associated with Central Asia and to a lesser extent Iran, rather than the simple snaffle bit or bridoon used in Rome and the west. For the armour of Hatra we have even less information. A fragment of 2nd century iron mail, fused into a mass, and a domed bronze shield boss with a long tube or spike at the front (Fig. 31) have been excavated. A strange, rather Hellenistic, 1st century statuette (Fig. 26), wears either a scale skirt or a short-sleeved scale hauberk under a sleeveless shirt. Various aspects of Hatra costume will, in fact, be echoed in Umayyad works of art, while a few styles such as the head-band of a supposedly 1st century bust in Mosul Museum will persist in various forms throughout much of the early Islamic period (Fig. 28).
In contrast to the Iranized élite of Hatra, we may see ordinary north Arabian tribesmen among those unfortunately disarmed Arab captives in a rock-carved "triumph" of Bahrām II at Bīshāpūr (Fig. 52).
266
From the other side of the desert barrier comes an undated carving from the Rushaydah area (Fig. 13) which the directorship or the Suwaydā Museum considers might portray a Ghassānid warrior. No reason for this attribution appears to have been published, although the carving's very simplified sword does have similarities with early Islamic weapons.
From a later period we have a Coptic textile showing the "Sale of Joseph" (Fig. 136). Here the Ishmaeilite trader appears as a dark-faced horseman armed with a mace, later to be regarded as a typical Arab weapon, and wears a head-cloth beneath a pointed helmet. Surely this is an Arab warrior, crude as the representation might be. But since this textile is probably from the 7th century, it should perhaps be placed among evidence for the first Islamic period.
By association alone we may regard the camel-riding warriors of 2nd and 3rd century Palmyrene and Dura Europos sculpture as Arab, in equipment if not necessarily in origin (Figs. 10, 11 and 32). These are apparently unarmoured men, armed with large-bladed spears, small round shields slung on the animals' flanks, and in one case a possible quiver on a baldric to the right hip. They are generally less well equipped than their infantry or cavalry counterparts, which I take to indicate that here we have armed traders, not warriors as such. Unfortunately, the only camel rider that I have found from south Arabia is on an undated Sabaean carving in Ṣanʾā' that is too mutilated to be of much use.45 Other equally undated and crude carvings from Sabaean or Ḥimyarite Yemen show spears and an apparently single-edged sword (Fig. 136). The finest such south Arabian frieze shows, surprisingly considering the wealth of evidence against the widespread
------------------------
45. Ṣanʾā' Museum (Cat. No. 31.300.1751), in Jawad ʾAlī, Tarīkh al ʾArab qabl al Islām, vol. V (Beirut 1956).
267
use or horses in the Yemen, a cavalryman with a long spear (Fig. 136).
Inadequate as this total of evidence might be, it does, I suggest, show the Arabs of north and south to be not especially poor in weaponry, nor in military technology, nor in military organization. In certain respects they parallel the supposedly backward German tribes along Rome's northern frontier. Here were peoples who lacked neither martial drive nor basic military capabilities, but who did lack both political cohesion and a reason to put all their efforts into overrunning the Romano-Byzantine or Iranian defences. In both cases a selected few such tribes were "trained" by their future victims for their own ends, namely to act as auxiliary troops under an imperial banner and to keep in check those "untamed" tribes further from the frontiers of civilization. Ironically, the final result seems to have been strangely similar north and south, though separated by two centuries.