There is no evidence for the practice among the early Iranians of taking large numbers of wives or concubines and keeping them in secluded quarters. The Iranian national history ascribes to ancient heroes and kings few children and fewer wives (Geiger, pp. 240-44. On the many sons of Gōdarz and Vištāspa, see below). Among the “Avestan people” (q.v.), the wife enjoyed a relatively elevated social status: she participated in rituals, and was charged with the running of the house, so that the “man and wife” were designated, characteristically, as nmānō-paiti “lord of the house” and nmānō-/dəmąnō-paθni “mistress of the house” (AiWb. Col. 1093; Geiger, p. 243 with n. 5). On the other hand, the institution of the harem (on which see Penzer) was firmly established in the ancient Near East. In Assyria, for instance, royal edicts laid down rules governing court and harem etiquette much the same as those prevailing in the Ottoman seraglio. The women of the harem were kept in seclusion, guarded by eunuchs, and prevented from turning their frequent disputes into seditious plots. They were not allowed to give presents to servants lest bribery was intended, and no one could see them unless first carefully examined by senior officials. When the king traveled, his court and harem traveled with him in accordance with strictly observed regulations (Grayson, 1991, p. 198).

With the conquest of the Near Eastern kingdoms, the Iranians took over many of their practices, and the harem became a tradition with Iranian dynasties and aristocracy as well (see ANDARŪN). Thus, the Medes, whose nobility are alleged to have kept no less than five wives (Strabo, Geography, 11. 13. 11), employed eunuchs as harem guards (Clearchus of Soli apud Athenaeus 12. 514d). Greek sources (collected by Rawlinson, 1873, pp. 216-22; see most recently Brosius, pp. 1-12, 105-18 and passim) alleged that excessive lust for luxury and women resulted in extraordinary influence of palace eunuchs and led to decadence among the Achaemenid kings and nobility. This view is partially based on Greek hostile ideological propaganda, but the idea that some influential “eunuchs” may have been non-castrated senior officials with no connection to the harem life (Briant, pp. 285-88, 945) is not borne out by our sources (cf. Lewis, pp. 16, 18, 20-21, 75-76, 82) and also neglects perfect analogies from the Sasanian, Byzantine, Ottoman and Safavid states.

The high social status of some royal and aristocratic women of the Achaemenid period is well documented (Brosius, pp. 83-182). They received an arduous education, which seems incompatible with the seclusion of harem life. Some at least learned such skills as horsemanship and archery (Ctesias, frg. 16 (56) in Jacoby, Fragmente III/C, p. 471), and even participated in hunting (Heracleides of Cyme apud Athenaeus, 514b). They appeared in public (Brosius, pp. 83-93), traveled with their husbands (pp. 87-8), participated at feasts (pp. 94-7), held vast estates and workshops (pp. 125-29), employed large numbers of servants and professional laborers (pp. 129-82), and at times wielded political power (pp. 105-19; see also Lewis, pp. 22, 75-76; 134-36). The following is an outline of the harem life of the Achaemenid epoch, which served as a model for subsequent periods. Herodotus (1.135), who wrote in the time of Artaxerxes, testifies that each (notable) Persian man had several wives, and a still larger number of concubines (Strabo, Geography 15.3.17, adds: “for the sake of having many children”). This was the case with the Persian king as well (see Brosius, pp. 13-20, 204-5). Wives came to the husband on a well-regulated turn-basis (Herodotus 3.69). They exercised total control over the family’s children until these were five years old (Herodotus 1.136), and they customarily accompanied their husband at dinner banquets (ibid., 5.18) but left when “women entertainers” of the harem came in and the men began merrymaking (Plutarch, Moralia, 140B). The chief consort, the wife, who as a rule was the daughter of a Persian prince and the mother of the heir to the throne (Rawlinson, pp. 216-18), controlled the household. The traditional title *māna-pašnī (or *māna-paθnī) “mistress (of the house),” (from nmānō-paθni see above; on the title see also Sundermann, EIr III, p. 678; Back, p. 200; Huyse, II, pp. 107-8), must have been borne by the Queen Consort, if not by the mother of the reigning monarch (on whose position see Rawlinson, pp. 220-21; Brosius, pp. 21-24). These ladies were subject only to the king; each had her own living quarter, her own revenue and estates and a large number of servants (Herodotus 3.134), which included harem eunuchs and concubines (Diodorus Siclulus 17.38, 1; see also ESTHER, BOOK OF).

The royal harem included three more groups of women, living in separate dwellings. First were the “ladies” of the household, legal wives other than the Queen or the mistress of a noble house. They evidently bore the title *bānūka- “Lady” (attested in a Persepolis Elamite tablet as ba-nu-qa (-na-be), see Hinz, p. 63; see also BĀNŪ). The second group consisted of unmarried princesses and the married ones who lived with their own family (cf. Brosius, pp. 70-82). The title for such a royal princess was *duxçī- (attested in the Persepolis Elamite tablets as du-uk-ši-iš; from duxθrī-, MPers. duxšy/duxt, NPers. doḵt: Benveniste, pp. 43-45; Hinz, p. 89; Harmatta, pp. 129-30). The term meant “daughter” also, and this may have confused the Greeks into reporting several cases of the marriage of a prince with a “daughter” instead of with a “royal princess.” The third group of harem women were concubines, beautiful girls (Plutarch, Artoxerxes, 27; Diodorus, 17.77.6; Esther 2.3) bought in slave markets (Herodotus 8.105; Plutarch, Themistocles, 26.4), or received as a gift (Xenophon, Cyropaedia, 4.6, 11; 5.1, 1; 5, 2, 9, 39) and tribute (Herodotus 3. 97), or collected from different parts of the empire (Esther 2.2-3;), and even captured from rebellious subjects (Herodotus 4.19, 32; Cf. Grayson, 1975, p. 114). The Greeks referred to these maidens as pallaki, a term, which denoted Athenian women of low social rank, but the application was not always justified (cf. Herodotus 3.1-2; Brosius, pp. 31-34). An Old Persian designation for “concubine” is not attested, but it was probably *harčī- (Av. hāirišī- [AiWb., col. 1806]) “woman” (cf. Armenian harč “second wife, concubine”), hence *harčīpati- “chief eunuch” attested in NPers. as harzbad (Tafazzoli, pp. 302, 304). While still virgins, they were kept and groomed in the harem’s “first house of women” (Esther 2.9), and trained as musicians, dancers and singers in order that they might entertain their king or the magnate lord at banquets or throughout the night (Aelian, Varia Historia, 12.1). When admitted to the royal harem, they were quartered in “the second house” (Esther, 2.14). Only two instances of large progenies in the Achaemenid house are known: Artaxerxes I had one son from his Queen and at least 18 sons from his concubines (Lewis, p. 75), and Artaxerxes II had three “legitimate” sons (Plutarch, Artoxerxes, 26) and some 150 sons by his concubines (Justin 9.1). Any child borne to such a concubine was regarded as inferior to the “rightful” offspring, and the Greeks came to call them, nothus “illegitimate.”

The Persians made every effort to safeguard the life-style and honor of their women (see especially Grotanelli). This intrigued the Greeks, and the vast number of the concubines in the royal harem of the later Achaemenids led to the rumor that “they were not less in number than the days of the year” (Diodorus, 17.77.6; see also Quintus Curtius 3.3.24; 6.6.8; Plutarch, Artoxerxes 27; Dicaearchus apud Athenaeus, 13.557b). In reality, the number was not so fixed; nor were all the harem women “concubines.” Dinon (q.v.) had reported that “three hundred women watched over” the Persian king (Athenaeus 514b), and Heracleides of Cymê explained that “these sleep throughout the day in order to stay awake at night, but at night they sing and play on harps continually while the lamps burn; and the king takes his pleasure of them as his concubines” (ibid.). A satrap of Babylonia is alleged to have kept no less than 150 such musicians and singers (Ctesias apud Athenaeus, 12.530d). Quintus Curtius (3.3.22-24) has given a vivid picture of the family and harem of Darius III as they accompanied him on his march towards Issus to meet Alexander. Two chariots carried his mother and “wife” (the Queen Consort); then came their female household, all riding on horses, followed by the king’s children and their governesses, carried in fifteen enclosed (mule-borne) litters (on the device see also Plutarch, Themistocles 26) and a herd of eunuchs, “who are not at alldespised” by the Persians. “Next rode the 365 concubines of the king, these [were] also regally dressed and adorned.” Darius’ mother, wife, and several children accompanied him to the battlefield and fell into the hands of the enemy. His treasure and harem, which had been left in Damascus, were captured by Parmenion, whoreported to Alexander (Athenaeus 13.608) that he had “discovered concubines of the king who played musical instruments, to the number of 329” and nearly 500 household servants.

The Persian word for the harem is not attested, but can be reconstructed as *xšapā.stāna “night station, place where one spends the night,” which developed into MPers. šap/bistān, NPers. šabestān (Bach, p. 260; Huyse, II, p. 176). Subsequently, a vṛddhi derivative of šap/bistān produced the honorific šāp/bistān, “eunuch” (see EIr IX, p. 66; Huyse, loc. cit.; de Blois, p. 36). A building at Persepolis located to the west of the “Treasury” has been called the harem of Xerxes (Schmidt, I, pp. 255-64), and despite some skepticism, the location and the nature of the structure make that identification sound. It consisted of a large hall with adjoining rooms, and a number of identical units, each forming an apartment with a four-columned hall and one or two side-rooms and storerooms, an arrangement which would have been admirably compatible with the function of a seraglio within which the queen and other wives of the king would have had their own living quarters. As one would expect, a thick wall surrounds the whole complex and access to it was essentially through a small entrance located in the southwestern corner, a feature that suggests a greatly protected privacy necessary for a royal harem. Quite appropriately, the entrance to the main hall of this palace shows Xerxes accompanied by two attendants, one of whom is a beardless eunuch.

The institution of the harem remained almost the same during the subsequent ages. Alexander married two Persian princesses (Arrian, Anabasis, 7.4.4; Justi, Namenbuch, pp. 64-65) and also took over the harem of Darius III (Quintus Curtius 6.6.8). The Seleucids were reputedly monogamous, but “the monogamy was official only, for the kings kept mistresses at their pleasure, some of whom might be openly invested with power” (Bevan, II, p. 279). During the Hellenistic period, the queens and royal princesses attained to high political and cultic ranks (ibid., I, p. 197; II, p. 132; Vatin; Macurdy; White and Kuhrt, pp. 204-10). The Seleucid king referred to his royal consort as adelphe “the sister and wife,” while his wife called him “brother” (White-Kuhrt, pp. 127, 282, 204). These literary and official designations were borrowed by the Parthians (Sachs-Hunger, p. 507; Minns, pp. 28, 31, 35), and contributed to the notion of sibling marriage among Iranians.

Very little is known of the harems of the Parthians. Our main authority is Justin (41.3): each man had several wives, and kept them fairly secluded from all but relatives and eunuchs. Contemporary documents from Babylonia and Avroman (q.v.) show that the king kept several queens or legitimate royal wives (Sachs-Unger, p. 507; Minns, pp. 28, 31, 35), but the story of Tiridates’ journey to Rome shows that only one was considered the Queen Consort. Tiridates’ queen greatly impressed the Romans by riding beside her husband and wearing a helmet instead of a veil (Dio Cassius 83.1.2). The case of the Roman slave girl, Musa, who so captivated the old king Phraates IV that he promoted her to the rank of Queen (Karras-Klapproth, pp. 95-96), demonstrates that clever concubines could achieve high social rank with some ease. She poisoned her husband and ascended the throne in joint kingship with her son, Phraataces. Their coins show her on the reverse and give her title as “Queen” (Camb. Hist. Iran, III, p. 293, pl. 6, nos. 4-5, not “Queen of Queens,” as is claimed in ibid., p. 712). That the aristocracy, too, maintained a large harem, is indicated by the life-style of Surena, who defeated Crassus: “He used to travel on private business with a baggage train of a thousand camels, and was followed by two hundred wagons for his concubines, while a thousand mail-clad horsemen and a still greater number of cavalry served as his escort” (Plutarch, Crassus 21.6). The attribution in the Iranian tradition of 78 sons to Gōdarz (see GŌDARZIĀN) and 30 sons to Vištāsp (see ESFANDĪĀR) are reflections of Parthian times, and evidence the life-style of the Arsacid aristocracy.

The picture of the Sasanian harem (šap/bistān) is much the same as that of the Achaemenid period. Ammianus Marcellinus (33.6.75-76) reports (in 363 #%) that the Persian Empire contained “many men of different tongues,” most of whom “are hardly contented with a multitude of concubines; they are free from immoral relations with boys. Each man according to his means contracts many or few marriages.” Two centuries later Agathias (II. 30.6) noted that in Persia, “a man could and did have any number of wives.” In this period, too, the kings married a chief wife—the queen, mother of the heir to the throne—and several wives of lower rank. That the household traveled with the king, even on campaign, is shown by the capture of Narses’ family by Galerius (Christensen, L’Iran, p. 233) and of the harem of Pērōz by the Hephthalites (Ṭabari, I. p. 877). The honorifics and ranks of the royal ladies have been the subjects of heated debate. Five titles are attested for royal women. They are: 1. “royal princess” (duxšy, duxt [also compounded with masculine names without indicating descent]; cf. Arm. Lw. dšxoy, “female ruler, queen, lady of ladies”: Bailey, pp. 91-92); 2. “Lady” (bānūg); 3. “Queen” (bānbišn, from earlier *māna-pašnī or *māna-paθnī”); 4. “Queen of the Empire” ([Ērān]šahr bānbišn, cf. Šahrbānū in later times); and 5. “Queen of Queens” (bānbišnān bānbišn). It was a characteristic of the system of ranks among Sasanian royalty and aristocracy that the highest female rank was not necessarily borne by the chief wife, and could well be given to a daughter or a sister (a system also attested in the Safavid and Qajar periods). Thus, when Šāpūr I calls his daughter Ādur Anāhīd “Queen of Queens” (Back, pp. 331-32; Huyse, p. 46), one should not conclude that she was also the wife of her father (see Harmatta, pp. 127-28; Gignoux, EIr. I, p. 472; Huyse, II, p. 107 with further literature). On the other hand, the same king gives his own mother’s title as “Lady” (bānūg), which, of course, does not mean that she was of lower rank. The title “queen” (bānbišn) was generally borne by the wife of a “king,” i.e., a prince governor, while a “royal princess” would be called a duxšy, duxt, a title which assumed the meaning of “queen” when she married a king, say of Armenia or Georgia.

No doubt many Sasanian kings and magnates kept a number of concubines. As in the olden days, the harem was supervised by eunuchs (sing. šāp/bistān, also šap/bistān; cf. harzbed mentioned above), some of whom attained to extraordinary power as military or civic officials (see EIr. IX, p. 66). In general, the harem of the Sasanians was not elaborate. Pĕrōz had a multitude of children and 30 of his sons were killed with him in his Hephthalite campaign (Procopius 1.4.2), and Ḵosrow Parvēz had eighteen sons (Ṭabari, I, p. 1060) and at least two daughters. But these seem to be exceptions rather than the rule. Of all the Persian kings, Ḵosrow Parvēz was the most extravagant in luxury and pleasure seeking. He searched the empire for beautiful girls (a document specifying their qualities has survived, see Ṭabari, I, pp. 1025-26; Christensen, L’Iran, p. 475), and it was rumored that he had kept some 3,000 of them in his harem (Ṭabari, I, p. 1041). This was clearly an aberration, and the people so profoundly abhorred his keeping of those girls in seclusion and denying them the benefit of marriage and progeny that they counted it as the fourth of the eight crimes for which Ḵosrow Parvēz was tried and executed (ibid., p. 1074). Ḵosrow’s answer (omitted by Ṭabari but recorded in Balʿami, ed. Bahār, pp. 1174-75) is revealing: “I kept them in such ease and prosperity and lavished so much wealth on them that they themselves showed no inclination to leave me for any man. In addition, every year I instructed Šērin [Ḵosrow’s favorite wife] to gather them so that if any one of them wished to marry and leave my palace, I would give her a dowry and send her to a husband. None showed any desire.” The women musicians (and we may adduce singers) shown among the companions of Ḵosrow Parvēz in the hunting scenes carved on the walls of the Ṭāq-e Bostān were undoubtedly meant to represent part of his harem.

This survey may be concluded with the observation that the sources on the Arab conquest of Iran do not bear out the popular assumption (see esp. Neẓām-al-Molk, p. 246) that the Iranian aristocracy suffered from decadence brought about by luxury and concubinage.



Classical authors are cited according to the Loeb editions. Agathias, The Histories, tr. Joseph D. Frendo, Berlin and New York, 1975.

Harold W. Bailey, “Armeno-Indoiranica,” TPS, 1956, pp. 88-126.

Edwyn Robert Bevan, The House of Seleucus, 2 vols., London, 1902.

F. de Blois, “Middle-Persian funerary inscriptions from south-western Iran,” in W. Skalmowski and A. van Tongerloo eds., Medioiranica (Orientalia Lovaniensia Analecta 48), Louvain, 1993, pp. 29-44.

Pierre Briant, Histoire de l’empire perse. De Cyrus à Alexander, Paris, 1996.

Maria Brosius, Women in ancient Persia (559-331 BC), Oxford, 1996.

A. K. Grayson, Assyrian and Babylonian Chronicles, Locust Valley, New York, 1975.

Idem, “Assyrian Civilization,” CAH2 III/2, Cambridge, 1991, pp. 194-228.

C. Grotanelli, “Honour, Women and Sanctuary at the Persian court (Plut. Them. 29-31 and Esther 6-8),” Dialoghi di Archeologia, 3rd. ser., 6, 1988, pp. 135-38.

J. Harmatta, “Sino-Iranica,” AAASH 19, 1971, pp. 113-47.

Walther Hinz, Altiranisches Sprachgut der Nebenüberlieferungen, Wiesbaden, 1975.

Philip Huyse, Die dreisprachige Inschrift Šābuhrs I. An der Kaʿba-i Zardušt (ŠKZ), 2 vols., Corpus Inscrip. Iran. Pt. III, Vol. I, Texts 1, London, 1990.

Amélie Kuhrt and Susan Sherwin-White, From Samarkand to Sardis: A new approach to the Seleucid Empire, London, 1985.

Margarete Karras-Klapproth, Prosopographische Studien zur Geschichte des Partherreiches, auf der Grundlage antiker literarischer Überlieferung, Bonn, 1988.

ʿOnṣor al-Maʿāli Kaykāvus b. Eskandar, Qābus-nāma, ed. Ḡolām-Ḥoseyn Yusofi, Tehran 1352 Š./1973.

B. S. Lesko, Women’s Earliest Records in Ancient Egypt and Western Asia. Proceedings of the Conference of Women in the Ancient Near East, Atlanta, 1989.

David M. Lewis, Sparta and Persia, Leiden, 1977.

Grace H. Macurdy, Hellenistic Queen: A Study in Women-power in Macedonia, Seleucid Syria and Ptolemaic Egypt, Baltimore, 1932, repr. 1990.

Neẓām-al-Molk, Siar al-moluk (Siāsat-nāmā), ed. H. Darke, 2nd ed., Tehran 1347 Š./1968.

N. M. Penzer, The Harem, London, 1936.

E. Minns, “Parchments of the Parthian period from Avroman in Kurdistan,” Journal of Hellenic Studies 35, 1915, pp. 22-65.

George Rawlinson, The Five great monarchies of the ancient eastern world, III, London, 1873. A. J. Sachs and H. Hunger, Astronomical Diaries and Related Texts from Babylonia, III, Vienna, 1996.

E. F. Schmidt, Persepolis I: Structures, Reliefs, Inscriptions, Chicago, 1953.

Ahmad Tafazzoli, “An Unrecognized Sasanian title,” Bulletin of the Asia Institute, N.S. 4, 1990, pp. 301-6.

C. Vatin, Recherches sur la marriage et la condition de la femme mariée à l’époque hellénistique, Paris, 1970.

(A. Shapur Shahbazi)

Originally Published: December 15, 2003

Last Updated: March 6, 2012

This article is available in print.
Vol. XI, Fasc. 6, pp. 671-672 and Vol. XII, Fasc. 1, pp. 1-3