BANDA “servant.”

i. The term.

ii. Old Persian bandaka.


i. The Term

Banda (NPers.) and its precursors bandak/bandag (Mid. Pers.) and bandaka (OPers.) meant “henchman, (loyal) servant, vassal,” but not “slave” (for which see barda and bardadārī). Occurrences of the word bandaka in the inscription of Darius I at Bīsotūn (DB) provide the earliest evidence.

The New Babylonian rendering of bandaka is qal-la-a “subordinate,” and the Elamite rendering is mú li-ba-ru-ri, i.e. lipar-ú-ri (ú meaning “I”) “my henchman” or “servant.” It should be noted that in the Akkadian version the word for “subjects” in section 7 (corresponding to DB l. 19) is not qallu (cf. Arabic qalīl “little,” base QLL) but irpl = ardū, which is the Old Babylonian word wardum (root wrd “to be under”). In the Elamite version the concept is expressed by an abstract noun: li-ba-memú-ni-na hu-ud-da-iš “they rendered service to me” (cf. NPers. bandagī kardand).

Extensions of nouns designating persons and professions by means of the suffix ka occur frequently in the Aryan languages. Thus OPers. bandaka is from banda (cf. the Old Indian masculine noun bandha “bond, fetter”) from the Indo-European root bhendh, also the source of English “bind, band, bond.” OPers. bandaka thus has a meaning similar to that of English “bondsman.” OInd. bandhaka, however, came to mean “one who binds with fetters” or “captures,” having undergone a semantic evolution different from that of OPers. bandaka to NPers. banda.

C. Bartholomae went astray in interpreting the word literally as “one who wears fetters.” NPers. banda, plur. bandagān (Pahlavi bandakān) and its long-attested doublet bastagān (a passive participle from the same root) mean “dependent” or “relative” and are synonymous with ḵᵛīšān and ḵᵛīšāvandān. NPers. ḵᵛīšāvand, a compound of ḵᵛēš and ā-vand (from °βant), has much the same meaning of “own folk” as qawm-e ḵᵛīš, ahl-e ḵᵛīš, ḵīšān or familiares among the Romans and oikeîoi/oikétai among the Greeks. In addition to the meaning of relatives by blood or marriage, all these words carry a connotation of duty to serve (more explicitly conveyed in NPers. ḵānazād). In the modern language, the human relationship of “binding” and “being bound” is expressed in many uses of the verb bastan/band and its derivatives and in locations into which they enter.

In Middle Iranian, bandag (bndk/g) is found in Pahlavi and has the same meaning of “(loyal) servant” as in Achaemenid and modern times. The same is true of the Sogdian βantak. In the Turfan fragments bandag has become bannag through a process of assimilation also seen in the main dialect of Persis (Southwestern Mid. Pers. bng as against Northwestern, i.e. Parthian bndg). Although the Frahang ī pahlavīk (beginning of chapter 13) does not clearly vouch for ʿbd (aḇdīn?) as the ideogram for bandak (bavandak?), the Aramaic ʿaḇd(ā) with the phonetic suffix ak, meaning “servant,” is well attested in Sasanian inscriptions, where the abstract noun for “service” also is found in the Mid. Pers. form bandakīh (written OBDkyhy) and the Parth. form bandakīf (written OBDkpy). The Pahlavi Psalter has OBDk for bandak. In line 16 of the Apsʾy inscription at Bīšāpūr there is a mention of “menservants and maidservants,” bandak ut kanīsak (written OBDk W knysky, the latter word having an anomalous s instead of s¡ = č).

The word occurs as a patronymic in the list of dignitaries in Šāpūr I’s inscription on the Kaʿba-ye Zardošt at Naqš-e Rostam: Bandagān (Mid. Pers. Bndkʾn, Parth. Bndkn, Greek Bandigan) father of Zurvāndāt.

Compounds of banda(k) occur frequently as personal names but otherwise are uncommon. The word ḵarbanda, meaning (1) “ass keeper” or “ass driver” (taken into Syriac in the Sasanian period as ḥarəḇandəqā) and (2) “mule,” probably began as a compound ending in the stem of the verb bastan “to bind,” like dīvband “exorciser,” kārband “industrious, docile.”

Among the many examples of men’s names with the component banda(k) are Āturbandak “Servant of the Fire” (Justi, Namenbuch, pp. 51a, 488), Barāzbanda “Servant of the Wild Boar,” i.e., Bahrām (Justi, p. 349b), Jošnaspbanda (= Gošnāspbanda) “Servant of the Fire of the Stallion” (Justi, p. 355a), Hazārbanda “Servant of the Thousand” (scil., yazatas), or simply “He who has a thousand servants” (Justi, p. 128a), Māhbandak “Servant of the moon” (but see Justi, pp. 185b, 490), Mihrevandak (with v from β) “Servant of the god Mithra,” an Iranian general whom the Armenians defeated in a.d. 571 (Justi, p. 214b), Šāhbanda (Safavid period; Justi, p. 274a), Otrārbanda “toward (the place) Otrār” (a.d. 810-11; Justi, pp. 336f.).

Ḵodābanda (= Banda-ye Ḵōda) “Servant of God” has been a name of rulers and dignitaries from medieval to modern times. It accords with an age-old formula which goes back to Akkadian Warad-ilim and can be traced through Old Testament Hebrew ʿAḇdəʾel/ʿAḇdiʾēl and Syriac ʿAbdallāhā to Arabic ʿAbd-Allāh. It reappears in Christian calques such as Greek Theodoulos and German Gottschalk and finally in Turkish names with suffixed qolī.

Also to be placed in this category (notwithstanding the contrary opinion of W. B. Henning and H. S. Nyberg) are the personal names with the suffix vand (from βand/t) peculiar to chiefs of the Zagros region in the Middle Ages, such as ʿAbbāsvand, Aḥmadāvand, Jalāl(ā)vand, Jalīlvand, Ḥamavand (= Moḥammadvand), Ḥasanāvand, etc. Many of these have subsequently become names of tribes or places. In the toponymy of southwestern Iran, some of the names which evolved in this way (man’s name > tribe’s name > place name) have kept the b of band, e.g., Farrāšband and Esmāʿīlband west of Fīrūzābād in Fārs. The word formation is analogous to that of ḵᵛīšāvand “relative, dependent” (discussed above) and has no connection with the OPers. and Aryan suffix vant.

In the polite speech of modern times, banda “(your) servant” is used as a substitute for man “I” and therefore takes the first person of the verb (e.g., banda ānjā būdam “I was there”). Conversely sarkār or ḥażrat/janāb-e ʿālī “(your) excellency” replaces šomā “you” and takes the second person. The locution has ancient oriental precedents in Akkadian waradka (fem. waradkī) and Old Testament Hebrew ʿaḇdəḵā “your servant,” polite for “I,” and fem. ămāṯəḵā “your maid,” but these take the third person. Expressions such as kamtarīn banda “(your) most humble servant” are still used today in letter endings. In progressive circles, however, such styles and likewise the use of banda for “I” are now generally avoided.



AirWb., col. 924.

W. Brandenstein and M. Mayrhofer, Handbuch des Altpersischen, Wiesbaden, 1964, p. 110.

W. Eilers, “Damawend,” Archiv Orientální 22, 1954, 24, 1956, 37, 1969, pp. 185, 271ff., 315f., 326.

Idem, Westiranische Mundarten aus der Sammlung W. Eilers III: Die Mundart von Sīvänd, Wiesbaden, 1988, pp. 15f.

Gignoux, Glossaire, pp. 19, 20, 48, 49.

Idem, Iranisches Personennamenbuch II/2, p. 55 no. 184 and reverse index p. 209b.

W. Hinz, Altpersischer Wortschatz, AKM 27/1, Leipzig, 1942, pp. 70f.

Justi, Namenbuch, pp. 488, 516, and elsewhere.

Kent, Old Persian, pp. 17 par. 39, 30 par. 75 III, 39 par. 111, 42 par. 122, 46 par. 1321, 51 par. 146 II.

Mayrhofer, Dictionary II, p. 407.

(W. Eilers)


ii. Old Persian Bandaka

The OPers. occurrences of bandaka are all in the Bīsotūn inscription, where it is an epithet of men of high rank whom Darius had chosen to be his generals against the “rebels”: Vidarna, Vindafarnā, and Gaubaruva (members of the seven great noble families), Dādṛši the Persian and Vivāna (satraps), Vaumisa and Artavardiya (of whom we know only that they were Persians), Dādṛši the Armenian, and Taxmaspāda the Mede. All Darius’s generals except his father Vistaspa were his bandakas. The word here means a nobleman “bound” to the king in a relationship which, though subordinate, was freely accepted and probably sealed with an oath. The relationship was symbolized by a girdle, worn by all Persians including the king in the Achaemenid period and throughout the history of ancient Iran.

Undoubtedly a similar relationship existed between the Sasanian king Narseh (r. 292-302) and “the greatest, the best, and the noblest subjects in our possessions” (Paikuli inscription, sections 16, 50, 95). These were all princes with distinguished titles who had rallied to Narseh and become his bandags.

The word bandaka/bandag was used not only with reference to dignitaries but also with the meaning of “servant,” i.e., any person whose relationship to another is one of obedience and/or dependence. Thus in DB l. 19 the conquered countries are said to be Darius’s “servants” (bandakā). The Aramaic word ʿlym “lad, servant” (cf. Latin puer), which is used to translate OPers. bandaka, is used in Aramaic documents found in Egypt as the term for “dependent” or “employee”: e.g., a bailiff employed by Aršām (Arsames), the Persian satrap of Egypt, his groom whose son received an estate, some workers, and a sculptor, as well as a freed slave who, given the change of his status, could not be reduced to servitude again. The conventional form of opening address in letters in Imperial Aramaic is “To A. from Y., his servant” (ʿlym; Cowley, p. 135, Driver, docs. nos. 2, 6, 8, 9). This formula reappears in Sogdian letters of the 8th century a.d., where the writer, being socially inferior to the addressee, declares himself to be the latter’s bandag (Freiman, passim). Selōk, the author of an inscription at Persepolis, similarly calls himself the bandag of his king Šāpūr Sakānšāh (Back, p. 496).

Although these facts seem to justify the conclusion that bandaka/bandag never meant “slave,” the problem presents difficulties. There is no evidence that bandaka ever meant “war captive.” Several other facts must also be taken into account. In Akkadian there are two possible equivalents of bandaka, namely ardu, which means either “slave” in the legal sense or “inferior” in any way (even the king being inferior to his god), and qallu, which means “slave, servant” or the like (see also i, above). Then again the choice of doulos for the Greek translation of bandaka (attested in literary texts such as Darius’s letter to Gadatas) can only be explained if bandaka was taken to signify, literally or figuratively, “slave.” The semantics of Mid. Pers. bandag are equally complex. In the Bīšāpūr inscription (ŠVŠ 16) it may perhaps denote slaves offered together with other possessions by Šāpūr to Apsʾy, but in Mid. Pers. legal texts the only term denoting “slave” appears to be anšahrīk. The “slave and master” (bandag-xwatāy) antithesis found in the Manichean texts and in the Dēnkard (VI, E43 a-e, advice to a bandag on ways to attain a better position; Shaked, p. 211) demonstrates the inferiority of the bandag but no more.

In both the Achaemenid and the Sasanian periods, bandaka/bandag definitely referred to the fact that all individuals were subject to the king and that some were also dependent on men of higher rank. At the same time, the status of the humblest was always depicted, in the vivid language of military-political metaphor, as being like that of the captive and the slave.

The subordination of the bandag was not solely political and economic, as the word was also used to denote the relationship between a worshiper and his god: the expression “slave to the gods” is found in the Dēnkard (Shaked, p. 177). Likewise in religious writings the word bandag is widely used with the meaning of “God’s creatures.” There, too, the girdle (kostī) worn by the initiated Zoroastrian is the symbol of his obedience as a faithful follower of the Good Religion (see, e.g., chap. 18 of the Vīdēvdād).

The relationship which the word bandaka implies probably stems from a prehistoric rank differentiation in tribal élites and in the military retinue of the chief or “king,” already changed in Darius I’s time, when also a non-Persian was able to be his bandaka.

Curiously enough, the word has been adduced in support of diametrically opposed historical interpretations. N. Pigulevskaya (pp. 141-58) saw it as proof of a “slavery-based mode of production” in Parthian and Sasanian Iran, while G. Widengren (1969, pp. 21-44) made use of it for his depiction of “feudalism” in pre-Islamic Iran. For opposite ideological reasons, each writer ignores the “Asiatic mode of production” or “Asiatic despotism.” From an objective viewpoint, it can hardly be maintained that the bandaka relationship was the basis of feudalism in ancient Iran, for many other factors must have been present: the breakup of communal landownership, the diversification of war techniques, and the increasing use of coins for money, to mention only a few. One certain fact, however, is that “despotism” in ancient Iran had a peculiar and remarkable capacity to incorporate great noble families, which were not only powerful but also stable. Feudalism did not enter the scene until after the disappearance of these great families and their replacement by local gentry (dehqāns) in the reign of Ḵosrow I Anōšīravān (531-79). Only then and subsequently did the words bandag and banda perhaps refer to feudal relationships.

See also barda and bardadārī; ḡolām.



M. Back, Die Sassanidischen Staatsinschriften, Acta Iranica 18, 1978.

A. Cowley, Aramaic Papyri of the Fifth Century B.C., Oxford, 1923.

J. Greenfield, “Some Notes on the Arsham Letters,” in Irano-Judaica, ed. S. Shaked, Jerusalem, 1982, p. 11.

H. Humbach and P. O. Skjærvø, The Sassanian Inscription of Paikuli III, Wiesbaden, 1983.

O. Klíma, “Zur Problematik der Sklaverei im alten Iran,” Altorientalische Forschungen 5, 1977, pp. 91-96.

M. Macuch, Das Sasanidische Rechtsbuch “Mâtakdân i hazār dātistān,” Wiesbaden, 1981, pp. 79-84.

A. Perikhanian, “Iranian Society and Law,” in Camb. Hist. Iran III/2, pp. 627-80.

N. Pigulevskaya, Les villes de l’état iranien aux époques parthe et sassanide, Paris, 1963.

S. Shaked, Wisdom of the Sassanian Sages, New York, 1979.

G. Widengren, Der Feodalismus im alten Iran, 1969.

Idem, “Le symbolisme de la ceinture,” Iranica Antiqua 8, 1968, pp. 133-55.

Search terms:

 بنده bandeh    


(W. Eilers, C. Herrenschmidt)

Originally Published: December 15, 1988

Last Updated: December 15, 1988

This article is available in print.
Vol. III, Fasc. 7, pp. 682-685