¹ Transliterations of non-Latin scripts are given in bold face, phonetic interpretations in italics. The use of material from a large number of different languages has made it desirable to employ, as far as possible, a unified system of transcription diverging in a few minor points from that usually followed in this journal for the individual languages. I trust that the spellings given will nonetheless be readily understood by scholars.

² Bartholomae, Chr., Altiranisches Wörterbuch ( = AirWb), Strasburg, 1904, 343.Google Scholar

³ AirWb 1856.

⁴ Kraeling, E. G., The Brooklyn Museum Aramaic papyri, New Haven, 1953, no. 5:4Google Scholar. The Iranian origin of ’zt was recognised simultaneously by Benveniste, E., JA, CCXLII, 1954. 298–9Google Scholar, Gershevitch, I., JRAS, 1954, 126Google Scholar, and de Menasce, P., BiOr, 1954, 161Google Scholar. See also Hinz, W., Altiranisches Sprachgul der Nebenüberlieferungen, Wiesbaden, 1975, 52.Google Scholar

⁵ Hesychii Alexandri Lexicon, 1442.

⁶ ibid., 1469. Note that this form exhibits the Ionian sound shift

⁷ Josephus, , bell, jud., I, 255Google Scholar: Parcorus, the son of the Parthian king, “left with Herod some of the horsemen known as freemen ”. Ant. XIV, 342Google Scholar: the Parthians left “200 horsemen and ten of the so-called freemen ”. The correct interpretation of these passages was suggested as long ago as 1879 by Nöldeke, Th. in his Geschichte der Perser und Araber zur Zeit der Sasaniden, 235 n. 2Google Scholar, but has been overlooked by the editors and translators of Josephus' works whom I have consulted; instead these refer habitually to the much discussed passage in Junianus Justinus' epitome of Trogus' (lost) histories which says (41, 2) that the Parthian army consists not of freemen (liberorum), like that of other nations, but mainly of slaves (servitiorum), and that in a certain battle of 50000 Parthian horsemen only 400 were freemen. This is in blatant contradiction to what Justinus tells us a little later (41, 3), namely that their slaves (servos) go on foot and their freemen (liberos) on horse, so that the majority of their cavalry could hardly have been slaves. It seems most likely that Trogus' source mentioned only that of the 50 000 horsemen 400 were i.e. noblemen, from which he drew the erroneous conclusion that the rest, i.e. the majority of the Parthian army, must have been slaves.

⁸ For the occurrences in the MP. and Parth. inscriptions see Gignoux, Ph., Glossaire des inscriptions pehlevies et parthes, London, 1972, 15Google Scholar (s.v. ’c’t), 19 (s.v. ’z’t), 48 (s.v. ’z’t). Note in particular the common sequence šahryārān (Parth. šahrδārān) ud wispuhrān ud wuzurgān ud āzādān, ‘rulers and princes and grandees and nobles’, suggesting that āzādān came to be applied especially to the lowest rank of Sasanian nobles. Of numerous literary references we could mention: (a) in the sense “noble”: Parth. šahrδār, kōfdār ud āzād, “ruler, mountain-lord, and noble”, Henning, W. B., Selected papers (= SP) II, Leiden, 1977, 447Google Scholar; (b) “manumitted”: MP. mard-ē, ka-š anšahrīg, ī pad dah bahr ē bahr xwēš, āzād bē kunēd, frazand, ī az ān anšahrīg zāyēd, har ēk pad dah bahr ē bahr āzād, “if a man sets free a slave who was to one tenth his property, then of the offspring that is born to that slave, each one shall be one tenth free”, Bartholomae, Chr, Zum sasanidischen Recht, I, Heidelberg, 1918, 47 n. 5Google Scholar; (c) “freeborn”: MP. bandagān pad rāh ī āzādān rawēnd, “slaves walk in the path of the freeborn”, Jāmāsp-nāmag, ed. Bailey, H. W., BSOS, VI, 1930, 58:36Google Scholar; (d) metaphorically: Parth. āzād až harw zāwarān, “free from all the (dark) powers”, Boyce, M., The Manichaean hymn-cycles in Parthian, London, 1954, 98Google Scholar; Parth dādum āzādīft ō tō, man hāmhirz, až …, “I gave you, my attendant, freedom from …”, Henning, , SP I, 308: 15–16Google Scholar. For MP. āzādīh see Shaked, S., “Ambiguous words in Pahlavi (3)”, Israel Oriental studies, IV, 1974, 239–42Google Scholar, with numerous references, and for its use in the sense “hereditary estate” Shaki's, M. remarks in Studia Iranica XII, 1983, 191–2 (note 53).Google Scholar

⁹ Hübschmann, H., Armenische Grammatik I, Leipzig, 1897, 91–2.Google Scholar

¹⁰ Tschenkéli, K., Georgisch-deutsches Wörterbuch I, Zürich, 1960, 8, glossed “frei (nicht hörig); in die Freiheit gesetzter Höriger”.Google Scholar

¹¹ see Henning, , SP II, 320 and note 9; and also below note 32.Google Scholar

¹² Benveniste, E., Vessantara Jātaka, Paris, 1964, v. 1099Google Scholar: “your freeborn (”z‘tčh) and divine persons have become slaves and servants”; Miiller, F. W. K., Soghdische Texte I, Berlin, 1912, 83: “there is neither slave nor freeman (’zṭy)”, translating Gal. 3, v. 28.Google Scholar

¹³ c.f. Benveniste, E., “Notes sur les textes sogdiens bouddiques du British Museum”, JRAS, 1933, p. 42Google Scholar, and the same author's Textes sogdiens, Paris, 1940, 6:109, 10:11, 10:25–6.Google Scholar

¹⁴ Bailey, H. W., Dictionary of Khotan Saka, Cambridge, 1979, 20–1.Google Scholar

¹⁵ and into Eastern Neo-Aramaic, which also has āzādūtā, “freedom”, with the Aramaic abstract suffix. See Maclean, A. J., Dictionary of the dialects of vernacular Syriac, Oxford, 1901, 7.Google Scholar

¹⁶ Wahby, T., Edmonds, C. J., A Kurdish-English dictionary, Oxford, 1966, 5.Google Scholar

¹⁷ Miller, W., Osetinsko-russko-nemečkiľ slovar', I, Leningrad, 1927, 16: “azat adj. ‘svobodnyĭ, frei’”. The word does not seem to appear in the more recent Ossetic dictionaries.Google Scholar

¹⁸ AirWb 1657–9. The Iranian root is evidently cognate with Indo-European *ĝen-, Skt.jan-, Gr. Lat. gignō, Arm. cnanim, etc. There is no trace, as far as I can see, of the expected South-West Iranian forms *dan-, *āδāta-, etc., in any stage of Persian for “give birth” and its derivatives, though Avestan zan- ², “to know”, does correspond to Old Persian dan-, MP., NP. dānistan. It seems that in the course of the differentiation of the two homophonous roots the forms with z- very soon became generalised in all Iranian dialects in words connected with “giving birth”. For Old Persian āmāta- see below, note 38.

¹⁹ ā- occurs with an active participle of zan- in Yasna 9₂₂: haomō āzīzanāitibiš daδāiti xšaētō.puθrīm uta ašava.frazaintīm. Bartholomae considered āzīzanā- to have the same meaning as the simple participle without preverb, and this interpretation has been followed by the translators, e.g. by Lommel: “Hauma gibt den Gebärenden herrliche Söhne und frommen Nachwuchs.” But, in the light of āzāta-, it is not impossible that the word has a more pregnant meaning, e.g. “to noble mothers”, or the like. Some such connotation was evidently assumed by the Middle Persian translator of the Yasna, who rendered this passage as hōm ’z’t’nc (var. ’YLYDWNt’nc) bē dahēd ān ī rōšn pus ud ān ī ahlaw frazand. The expected reading for the word which I have left without transcription is a-zādān-iz, “unborn”, but this does not make any discernible sense. The Sanskrit translation attributed to Neryosangh Dhaval has “to the childless (ajātakēbhyō)”, which — as Darmsteter already noted — is a possible rendering neither for the Middle Persian text, nor for its Avestan prototype. It seems most likely that ’z’t is merely an irregular spelling for our friend āzād (normally written pseudo-historically as ’c’t, though ’z’t is the usual spelling in the MP. and Parth. inscriptions); the MP. translation could then be rendered as “Haoma gives a brilliant son and righteous progeny to the noble ones”. References: AirWb 1658; Lommel, H., Die Yāšt's des Awesta, Göttingen/Leipzig, 1927, 190Google Scholar; Dhabar, B. N., Pahlavi Yasna and Visperad, Bombay, 1949, 64–5Google Scholar; Unvala, J. M., Neryosangh's Sanskrit version of the Hōrn Yašt, Vienna, 1924, 35Google Scholar; Darmsteter, J., Le Zend-Avesta I, Paris, 1892, 93 n. 70.Google Scholar

²⁰ Perikhanian, A., “Notes sur le lexique iranien et armenien”, Revue des études arméniennes, nouvelle série V, 1968, 9–30. The passages quoted below are on pp. 11–12 and 14.Google Scholar

²¹ AirWb 1688–9.

²² Nyberg, H. S., A manual of Pahlavi II, Wiesbaden, 1974, 41.Google Scholar

²³ See especially 1253b–1255b.

²⁴ See the references quoted in notes 8, 12, 13.

²⁵ c.f. Ernout, A., Meillet, A., Dictionnaire étymologique de la langue latine, 4th ed., Paris, 1967, 271.Google Scholar

²⁶ For references for freó and its derivatives see Bosworth, J., Teller, T. N., An Anglo-Saxon dictionary, Oxford, 1898, 333–5.Google Scholar

²⁷ Canterbury Tales, the Prologue, 45–6.Google Scholar

²⁸ von Wartburg, W., Französisches etymologisches Worterbuch, 15/II, Basel, 1969, 163–170.Google Scholar

²⁹ discussed by Bailey, H. W. in BSOS VI, 1932, 953–5Google Scholar; BSOAS XXIII, 1960, 16–21Google Scholar; Dictionary of Kholan Saka, Cambridge, 1972, 20–1Google Scholar; Zoroastrian problems in ninth-century books, 2nd ed., Oxford, 1971, xxx, 10.Google Scholar

³⁰ AirWb 341; the meaning “noble” was first recognised by Bailey.

³¹ AirWb 341.

³² Henning, , SP II, 227, 23: In this bilingual text the plural ’zn”n is glossed by Sogdian [’z]’ṭ’qṭ.Google Scholar

³³ Henning, W. B., Īrān Shināsī II, 1971, 62–3.Google Scholar

³⁴ Bailey, , Dictionary, 21.Google Scholar

³⁵ Tschenkéli, , op. cit. (see n. 10), 9: “(niedriger) Adliger”.Google Scholar

³⁶ Bailey, , Zoroastrian problems, xxx.Google Scholar

³⁷ The Assyrian dictionary of the Oriental Institute of the University of Chicago X/l (M), Chicago/Glückstadt, 1977, 256–7Google Scholar; von Soden, W., Akkadisches Handwörterbuch II, Wiesbaden, 1972, 615–6Google Scholar; for the following see also Eilers, W., Zeitschrift für Assyriologie LI (Neue Folge XVII), 1955, 232–5. The Assyrian dictionary, loc. cit., quotes two passages where Assurbanipal speaks of “men of royal lineage, archers (and) mār banĭ of Elam”; whether mār banĭ here means “nobles”, as stated by the editors of the dictionary, or merely “citizens” is doubtful to me. If it does designate “nobles”, this might indicate that the Elamites, like the Persians, used the same word for “freemen” and “nobles”.Google Scholar

³⁸ von Voigtlander, E. N., The Bisitun inscription of Darius the Great, the Babylonian version, London, 1978, §3Google Scholar. The Old Persian version, in Kent, R. G., Old Persian grammar texts lexicon, New Haven, 1950, 117₇, has āmātā, which, on the basis of the Akkadian version, has been assigned the meaning “nobles”. The true meaning and etymology of the word remain controversial.Google Scholar

³⁹ Same inscription §§13, 32, 42, 43, 47, 50.

⁴⁰ Akkadian text in Herzfeld, E., Altpersische Inschriften, Berlin, 1938, 6:5–6Google Scholar. The Old Persian text is there and in Kent, , op. cit., 138:8 sqq. (§8a).Google Scholar

⁴¹ The fact that Old Persian *āzāta- does not actually occur in any of these inscriptions is striking, but cannot invalidate the conclusion drawn here. My point is only that the Akkadian employed in the chancellery of Persepolis used mār banĭ in a sense otherwise unknown in Akkadian, a sense which reflects Old Persian usage.

⁴² Centenario della nascita di Michele Amari I, Palermo, 1910, 84–95.Google Scholar

⁴³ ibid., 92.

⁴⁴ Greenfield, J. C., Porten, B., The Bisitun inscription of Darius the Great, Aramaic version, London, 1982, v. 48, corresponding to §§42–3 of the inscription.Google Scholar

⁴⁵ Cowley, A., Aramaic papyri of the fifth century B.C., Oxford, 1923, no. 30:19, 31:18.Google Scholar

⁴⁶ Lindenberger, J. M., The Aramaic proverbs of Ahiqar, Baltimore, 1983, v. 217.Google Scholar

⁴⁷ Similarly, the much discussed Aramaic bar nāšā, “man” (literally “son of man”), is clearly influenced by Akkadian mār awīlim.

⁴⁸ See the references in Brockelmann, C., Lexicon syriacum, 2nd ed., Halle, 1928Google Scholar, 252b, 89a. and in Smith, R. Payne, Thesaurus syriacus I, Oxford, 1879, 587, 1356–7.Google Scholar

⁴⁹ c.f. The doctrine of Addai, the apostle, ed. Phillips, G., London, 1876, p. 1.3–4Google Scholar (referring to the courtiers of king Abgar of Edessa), and the passages from the Julian romance listed by Widengren, G., Der Feudalismare im alien Iran, Köln/Opladen, 1969, 140.CrossRefGoogle Scholar

⁵⁰ Moberg, A., The book of the Himyarites, Lund, 1924Google Scholar; c.f. Pigulewskaja, N., Byzanz auf den Wegen nach Indien, Berlin/Amsterdam, 1969, 233–5.Google Scholar

⁵¹ Drower, E. S., Macuch, R., A Mandaic dictionary, Oxford, 1963, 127b.Google Scholar

⁵² ibid.

⁵³ ibid., 69a.

⁵⁴ ibid., 127a.

^54a Numerous references in Blachére, R. et al. , Dictionnaire arabe-français-anglais, Paris, 1964 sqq., 2392–2400.Google Scholar

⁵⁵ Ṭabarī, , Ta'rīxu r-rusuli wa l-mulūk, ser. I, Leiden, 1879Google Scholar sqq., 956:13 = l-'Iṣbahānī, 'Abu l-Farăi, Kitāhu l-'Aghānī XVI, Cairo 1285 h., 75Google Scholar, 3 lines from the bottom; also quoted by Mas'ūdī, Ibn Hišām, 'Azraqī, etc. That the Yemen was, for a time, dominated by fāris al-'aḥrār is mentioned also in the much quoted lines supposedly inscribed on the gate of ẓafār; see the references cited by Tkatsch, J. s.v. “ẓafār”, EI ¹, 1186aGoogle Scholar and add the section from Bayrūnī's al-'Aθāru l-bāqiyah edited by Fück, J. W. in Documenta islamica inedita, Berlin, 1952, 77, with an additional reference in the footnote.Google Scholar

⁵⁶ Ṭabarī, 1036:3, 4Google Scholar = 'Aghānī XX 139:18Google Scholar = The Naḳā'id of Jarīr and al-Farazdak, ed. Bevan, A. A., Leiden, 1908–1912, II, 645:1Google Scholar; Ṭabarī, 1037:6Google Scholar = Naqā'id 645:16Google Scholar; Ṭabarī, 1037:18Google Scholar = Naqā'id 646:7.Google Scholar

⁵⁷ Ṭabarī, 2355:11–14.Google Scholar

⁵⁸ Dīwānu l-Farazdaq II, Beirut, 1960, 112:10, see also I 269:9, 307:2, etc.Google Scholar

⁵⁹ Dīwānu Baššāri bni Burd III, Cairo, 1957, 231:1.Google Scholar

⁶⁰ Beeston, A. F. L., et al. , Sabaic dictionary, Louvain/Beirut, 1982, 71Google Scholar. Very common in the inscriptions is the phrase “its (their) freemen and slaves”: Sabaic ḥrhw w^cbdhw, Répertoire d'épigraphie sémilique ( = RES) 3945: 6, 9, 12; Minaean ḥrs₁m wḥws₁m, RES 2771:7; Hadrami ’ḥrrs₁m w'dms₁m, Corpus des inscriptions et antiquites sud-arabes I, Louvain, 1977, 47.82/06. In a few inscriptions it might seem that hr is used in a less literal sense, e.g. when RES 4818:4–5 prays for “free, male, notable children (’wldm ’ḥrrm ’δkrwm 'lnym)”, or when RES 4912:1 mentions “Yd'’l Byn, king of Hadramawt, son of (or: of the clan) Rbs₂ms₁, son of the freemen (’ḥrr) of Yhb’r”. But it would be premature to postulate the meaning “noble” on the basis of such ambiguous references.Google Scholar

⁶¹ Dillmann, C. F. A., Lexicon linguae aethiopicae, Leipzig, 1865, 85–6.Google Scholar

⁶² Littmann, E., Hofner, M., Wörterbuch der Tigrē-Sprache, Wiesbaden, 1962, 65a: “Freiheit, Freier, frei”.Google Scholar

⁶³ da Bassano, F., Vocabolario Tigray-Italiano, Rome, 1918, 43: “liberta, libero”. The same word is entered again in col. 426 with the spelling (in Tigrinya are pronounced identically) and glossed “liberto, che è stato dichiarato libero dalla schiavitù”.Google Scholar

⁶⁴ ibid., 43: “libertà, stato libero e indipendente; p.e. di uno schiavo dichiarto libero si direbbe [ḥәrәnnät rä̆ibu] ha ottenuto la libertà.” Also 426 with the spelling [harsnndt] “libertà”.

⁶⁵ Guidi, I., Vocabolario Amarico-Italiano, Rome, 1901, 429–30: “(1) s. condizione di uomo libero, non ischiavo (2) a. libero, non ischiavo”.Google Scholar

⁶⁶ ibid., 430: “s. condizione di uomo libero, che non e schiavo”.

⁶⁷ Leslau, W., Etymological dictionary of Gurage (Ethiopic), Wiesbaden, 1979, III, 89.Google Scholar

⁶⁸ ibid., III, 328.

⁶⁹ Leslau, W., Lexique soqoṭri, Paris, 1938, 192.Google Scholar

⁷⁰ Jahn, A., Die Mehri-Sprache in Südarabien, Vienna, 1902, 192b.Google Scholar

⁷¹ Leslau, , Etymological dictionary of Gurage, III, 89, states that Dillmann and Littmann/Höfner “derive G[eez] ḥarawi ‘free’ from G. ḥara ‘army’. I wonder whether the Ethiopic root for ‘free, freedom’ is not borrowed from Ar. ḥurr ‘free’ and is not to be connected with G. ḥara ‘army’.” In fact, neither Dillmann nor Littmann/Höfner say anything like what is here attributed to them. Dillmann, indeed, says exactly the opposite, namely that ḥär(r)ā “primitus significat nobiles, liberos, quorum praecipuum munus erat cum rege in expeditionem proficisci”, op. cit., 85.Google Scholar

⁷² Guidi, , op. cit., 430Google Scholar, quotes an Amharic “nome di truppa” are bäṣär ḳädät with reference to Perruchon, J., Les chroniques de Zar'a Yâ'eqôb et de Ba'eda Mâryâm, Paris, 1893, 35. In fact, Perruchon's text has ore bäṣ₂är waďät, var. waďät; cf. 31: ṣ²ewa bäṣ₂ärwaďät, and also 45.Google Scholar

⁷³ See my article “Clan-names in ancient Ethiopia”, to be published in Welt des Orients.