CC BY
12
UDC 811.14
Philologia Classica. 2021. Vol. 16. Fasc. 1
"The Messengers are Called Sangandai": the Greek oayyavSn^ and its Relationship to
Rafal Rosol
Adam Mickiewicz University, Institute of Classical Philology, Fredry 10, 61505 Poznan, Poland; rafros@amu.edu.pl
For citation: Rosol R. "The Messengers are Called Sangandai": the Greek aayyavSr|<; and its Relationship to aa-yavSr|<;/aaTavSr|<;. Philologia Classica 2021, 16 (1), 40-49. https://doi.org/10.21638/spbu20.2021.104
This article examines the Greek noun aayydvSr|; 'messenger' which is attested in two lexica, dated to the Roman or early Byzantine periods: the Cambridge Rhetorical Lexicon by an anonymous author and Difficult Words in the Attic Orators by Claudius Casilo. In both works, aayydvSr|; appears together with three words of likely Iranian provenance: opoadyyri; 'benefactor of the Persian king; bodyguard', napaadyyr|; 'parasang; messenger' and ayyapo; 'messenger, courier; workman, labourer'. The word aayydvSr|; is analysed in comparison with daydvSr|;/daTdvSr|; 'messenger' occurring for the first time in Plutarch's works and closely linked to the Achaemenid administration. According to the hypothesis put forward in the present paper, both aayydvSr|; and daydvSr|; (with its secondary variant dardvSr|;) are connected to Manichaean Middle Persian/Parthian izgand 'messenger', Sogdian (a)zyand/(o) zyand/z(i)yant 'id.', Jewish Aramaic ' tzgadda 'id., Syriac izganda/izgadda 'id.', Mandaic asganda 'helper, assistant, servant; the Messenger, and go back to Old Persian *zganda- or to early Middle Persian/early Parthian *zgand- (or *zgand-) with the original meaning 'mounted messenger'. The reconstructed noun is derived from the Proto-Iranian root *zga(n)d- 'to go on, gallop, mount, attested in Avestan (Younger Avestan zgaS(/6)- 'to go on horseback, gallop') and in some Middle and Modern Iranian languages. The original form of the loanword in Greek was probably *aydvSr|; which then underwent certain transformations. Keywords: etymology, foreign words in Greek, Iranian loanwords, Greek lexicography.
The noun aayydvSnc; 'messenger, courier,1 which has not been previously investigated is attested in two Greek lexica, dated to the Roman or early Byzantine periods: in the so-called Cambridge Rhetorical Lexicon (Lexicon Rhetoricum Cantabrigense) by an anonymous author, and in the Difficult Words in the Attic Orators (Ta napa tot; AttikoT; prpropai (rfTOU^eva) by Claudius Casilo. The entries in both sources are almost identical (see Table 1).
The term aayydvSnc; appears three times in each text, but in Casilo's work two variants occur: once aayydvS- and twice aayydS-. The latter form is most probably an error like in some other words of the lemma. Apart from the errors in the sentences quoted above, Tic; instead of Ti and napaadyyoic; instead of napaadyyac;, there is also opoadyKai for opoadyyai and dyydpiov for dyydpeiov.2
1 LSJ gives two forms, aayyd5r|c; and aayydv5r|c;. However, only the latter seems to be correct (see below).
2 The manuscript is much better in the case of three other lemmas preserved, i. e., 'Opyd;, 'OarpaKiapou rp oho; and nsvsarcu.
© St. Petersburg State University, 2021
Table 1. The Greek term aayyérvônç in the sources
Lexicon Rhetoricum Cantabrigense, p. 22-23 (ed. by E. O. Houtsma = Lexica Graeca Minora, p. 82-83) Claudius Casilo, p. 397-398 (ed. by E. Miller = Lexica Graeca Minora, p. 243-244)
'Opoadyyn; Kai aayydv5r|c; Kai napaaayyn; Kai ay-yapo; Siacpspsi- (...). ZayydvSai 5s o'l dnoarsWo-[i£voi KaXouvtai. ZocpoKAxjt; 5s ¿v Toi; noi|isai Kai EupiniSn; ¿v EKupiai; napaadyya; aurou; K£K\r|-Kaaiv- ¿XP^v 5S sinsiv aayydvSa;- o yap napaady-yn; |£tpov ¿ariv. (...) Tic; (read: Ti) o opoadyyn;, Kai Ti aayydSn; (read: aayydvSn;), Kai Ti napaadyyn;, Kai Ti ayyapo;. (...) EayydvSai 5s o'l dnoaT£AA6|i£voi KaXouvTai-ZocpoKAxjc; 5S ¿v noi|isai, Kai EupiniSn; ¿v EKupiai; napaadyyoi; (read: napaadyya;) auTou; KSKX^Kaaiv- ¿Xpfjv 5S sinsiv aayydSoi; (read: aayydvSa;)- o'l yap napaadyyai |£Tpov ¿aTiv o5ou (...)
"Orosanges, sangandes, parasanges, and angaros differ from each other. (...) The messengers are called sangandai. Sophocles in the Shepherds and Euripides in the Scyrians have called them parasangai. They should use sangandai, for the parasanges is a measure unit." "What is orosanges, and what is sangandes, and what is parasanges, and what is angaros? (...) The messengers are called sangandai. Sophocles in the Shepherds and Euripides in the Scyrians have called them parasangai. They should use sangandai, for the parasangai are a unit for road measurement."
The word aayyâvônç bears a striking resemblance in phonetics, morphology and semantics to another Greek word for 'messenger, namely aaydvSn^/aaxdvSn^, and it is likely that they have the same origin. The latter word, known in two variants,3 is attested for the first time in Plutarch and, according to Greek sources, is closely linked to the Achaemenid administration. The form aayavSn; appears twice in all manuscripts of the Life of Alexander by Plutarch, although modern editors of this work correct the form to àaxavôn^ at both instances (this conjecture was accepted already in the 16th-century edition by H. Stephanus). Plutarch reports that Darius III was earlier a royal courier, i. e., àaxavôn^ (mss. àayavôn^) œv paaiÀiœc; "being a royal courier" (Alex. 18.7) and è^ àaxavôou (mss. àayàvôou) paatXeû; yevo^evoc; "having become a king from a courier" (18.8; ed. by K. Ziegler).
In turn, the form àaxavôn; is attested in all manuscripts of another work by Plutarch devoted to Alexander, namely, On the Fortune or the Virtue of Alexander the Great (De Al-exandri Magni fortuna aut virtute). We learn again that Darius III was earlier a royal courier: Aapeio; ^v aôv epyov, ov èK ôouÀou Kai àaxavôou paatA.éœ; Kuptov nepaœv énonça; (...) — "Darius was your handiwork: he who was a slave and courier of the king, him did you (scil. Fortune) make the mighty lord of Persia (...)" (Mor. 326e; ed. by W. Nachstadt, W. Sieveking and J. Titchener; trans. by F. C. Babbitt). Then, Plutarch writes the same about Arses (Oarses), the predecessor of Darius III: èni 0upa; aùxœ to ôiâôn^a x^; 'Aaia; ^À.0ev, œanep 'Oâpaq ôtà Bayœav, Kai axo\^v èKÔuaa^evo; àaxavôou neptéGexo x^v PaaiÀiK^v Kai ôpGonay^ Kixaptv; "Did the royal diadem come to Alexander's doors, as to Oarses through the machinations of Bagoas, who stripped from him the garb of a courier and put upon him the royal raiment and the tiara that ever stands erect?" (Mor. 340b-c; ed. and trans. as above).
The form àaxâvôn; also appears in the editions of Deipnosophistae (3.122a/94) by Athenaeus of Naucratis in a passage enumerating some Persian words in Greek. Howev-
3 Probably not in three; see below on the controversial gloss daKav5|c;.
er, we have here an emendation: rout; dar<dvSac; ^ dyy>dpouc; (edited and corrected by G. Kaibel),4 whereas all the manuscripts provide only rou; dardpou; without any hint at a meaning. We cannot exclude the possibility that dardpou; is simply an early distortion of dyydpouc; (a lapsus calami of AZTA- instead of ArrA- in majuscule is conceivable).
The Suda mentions daT&vSn; in two entries, namely a 4220 (ed. by A. Adler): AardvSai: ol ¿k SiaSox^; ypa^aTo^opoi. ol Se aurol Kal ayyapoi. ra Se ovo^ara nepaiKd. — "Astandai: the letter carriers in relays. They are the same as angaroi. Both are Persian words.", and a 165 (ed. by A. Adler; = Phot. a 95, ed. by Ch. Theodoridis): 'Ayyapoi: ol ¿k SiaSox^; ypa^aTo^opoi. ol Se aurol Kal dardvSai. ra Se ovo^ara nepaiKd. (...) — "Angaroi: the letter carriers in relays. They are the same as astandai. B oth are Persian words. (.. .)".5 A very similar definition of ayyapoc; is to be found in the reconstructed lexicon of Aelius Dionysius, i. e., a 16 (ed. by H. Erbse): ayyapoi- ol ¿k SiaSox^; ypa^aTo^opoi. ol S' auroi Kal dardvSai. ^ Se Xe^ic; nepaiKr (...) — "Angaroi: the letter carriers in relays. They are the same as astandai. A Persian word (.. .)".6
Moreover, Hesychius lists two related entries in his lexicon, i. e., a 7683 (ed. by I. Cunningham): fdaKavSr;- ayyeXoc; — "askantes: messenger", and a 7814 (the same edition): daT&vSn;- %epoSpo^oc;. ^ Kp&pparrov. ^ ayyeXov. Tapavrtvoi — "astandes: long-distance courier. Or pallet. Or messenger. Tarentians". In the case of the latter gloss, there are no difficulties with the beginning of the lemma, which is crucial for our study: dar&vSnQ ^^epoSpo^o;. In the second part of the definition, the case changes to accusative; for this reason we may presume that two words blended together, i. e., daTdvSn; 'messenger' and daKdvTn; 'pallet'. Therefore, it is likely that the ethnonym Tarentians does not refer to daTdvSn;, but rather to daravrnc;.7
As for the gloss daKavSr the editors mark certain difficulties with a crux philologorum (as quoted above), as well as by means of an asterisk or parentheses: *daKavSr|c;- ayyeXoc; (ed. by M. Schmidt), [aaKavSr;- ayyeXoc;] (ed. by K. Latte). The gloss is placed in correct alphabetical order between daKdvn and AaKavq, but the word aaKavSr; 'messenger' probably never existed.8 It might be a result of a misunderstanding of an earlier source where daydvSnc;/daTdvSnc; 'messenger' and daKdvTn; 'pallet', phonetically and morphologically similar, were compared.9 It is difficult to say exactly how it happened, but, no doubt, it was entirely possible (cf. the gloss daTdvSn; mentioned above, where the word Kp&pparroc; 'pallet' occurs in its definition).
The context and definitions in the lexica indicate that both daydvSnc;/daTdvSnc; and aayydvSn; might be of Iranian origin. As we have seen, daydvSnc;/daTdvSnc; in Plutarch is closely connected to the Achaemenid administration. Moreover, some lexica describe daTdvSn; (along with ayyapoc;) as a Persian word. ZayydvSn; appears together with three
4 J. Schweighauser in his edition from the beginning of the 19th cent. prints rou; daT<dv5a; Kal dYY>dpou;.
5 Cf. also EM 6.44 (ed. by Th. Gaisford): 'Ayyapoi: ol ¿k SiaSoxfjc; ypa^^aro^opoi. ol Se aurol Kal dardvSai.
6 Cf. Eust. Comm. ad Hom. Od., vol. II, p. 189. 5-6 (ed. by J. G. Stallbaum): Kupiw; Se Kara ATXiov Aiovuaiov ayyapoi ol ¿k SiaSoxfjc; ypa^^aro^opoi. ol S' auroi, 9^ai, Kal dardvSai nepaiKw;.
7 K. Latte in the apparatus criticus of his edition suggests that the Tarentine gloss was taken from Rhinthon (4th/3rd cent. BC). Cf. also Hsch. a 4554: dvapo; ayyeXo;. Tapavrtvoi.
8 Greek daKavSf|? or, more often, daKdvSn; is commonly treated as a real word without indicating any problems; see e. g., Frisk 1960-72, vol. 3, 41; Huyse 1993, 276; Chantraine 1999, 127; Beekes 2010, 149.
9 Cf. Hsch. a 7685 (ed. by C. Cunningham): &aKdvrr|c;- Kpdparo;, KXiviSiov eureXe;; Cyrill. a 1154 (ed. by U. Hagedorn): daKdvTn;: KpdparTo;.
other words of possible Persian or Median provenance: opoaayyn; 'benefactor of the Persian king; bodyguard', napaaayyn; 'parasang; messenger'10 and ayyapo; 'messenger, courier; workman, labourer'.
Since the end of the 19th century, scholars have juxtaposed dayavSn^/daxavSn; with similar words from Near Eastern languages.11 First of all, we have Manichaean Middle Persian and Manichaean Parthian izgand (written as jgnd) 'messenger'.12 There are also some Sogdian forms attested in different scripts and/or different types of texts, i. e., (a) zyand (written as zy' nt(t); Buddhist texts), (o)zyand ( ' zy' nt; non-Buddhist texts) and z(i) yant (zynt; Christian texts) 'messenger'.13 Moreover, the Iranian term was borrowed into some Semitic languages: Jewish Aramaic ' izgaddd 'messenger' and, without vocalization, ' zgd 'id.',14 Syriac izgandd and izgaddd 'id.',15 as well as Mandaic asganda 'helper, assistant, servant; the Messenger'.16
Concerning the etymology of Middle Persian and Parthian izgand etc., Huyse (1993, 276) is probably right in deriving them from the Iranian root *zga(n)d-.17 According to Cheung (2007, 473-474) its basic meanings are 'to go on, gallop, mount', but the author does not mention the words for 'messenger, and reconstructs the root as *zgad- without a possible variant containing an internal n. However, if we accept that izgand etc. belongs here, we should reconstruct *zga(n)d- as Huyse did.18 A possible variant with an internal -n- is probably also attested in the Sogdian form fi(a)zyand(-) (see below). The root *zga(n)d- itself has no related words outside the Iranian branch. It occurs in Young Avestan zgaS(/Q)- 'to go on horseback, gallop' (cf. also fra-zgaS- 'to go forth')19 and has well-established descendants in Middle and Modern Iranian languages, e.g. Khotanese ysgad- 'to mount' (attested with preverbs va- and ha-),20 Sogdian fi(o)zyaS- and fiozyast-(written as fizyS- and fiz'yst-; Buddhist texts), fiozyaS- (written as bzyS-; Christian texts), fi(a)zyaS- and fi(a)zyast- (written as fijySS- and fijyst-; Manichaean texts) 'to mount', as
10 The meaning of 'messenger' in the case of napaaàyyr|t; occurs only in these two lexica (= fr. 520 of Sophocles, ed. by S. Radt, and fr. 686 of Euripides, ed. by R. Kannicht). Besides, this meaning seems to be attested in a corrupted gloss of Hesychius n 658 (ed. by P. A. Hansen): napaaàyyr|- <ay>y£\w (cj.; ms. and M. Schmidt: napaaayyiXoyw)- o'l népaai tout; ÖiayysAAovrat; oûrw Aiyoum — "parasanges: messenger (dat.) (or parasangilogo): Persians call so messengers".
11 See Horn 1893, 29, note 3; Happ 1963, 98; Frisk 1960-72, vol. 3, 41; Huyse 1990, 95 and 1993: 276-277; Mancini 1995, 85, note 15 and 1995-96, 213; Chantraine 1999, 127; Schmitt 2002a; Ciancaglini 2008, 105-106; Beekes 2010, 149. It is surprising that Brust (2008) does not include àayàvÔn;/àaTàvÔn; in his study of Indian and Iranian vocabulary in Greek.
12 Durkin-Meisterenst 2004, 85.
13 Gharib 1995, 91, 458, 459; Sims-Williams 2021, 240.
14 Jastrow 1903, 46 ( ' tzgad and 'tzgadda); Sokoloff 1990, 43 ('zgd) and 2002, 112 ( 'tzgadda); Cook 2008, 7 ( 'zgd, attested only in the absolute plural form ' izgadîm).
15 Sokoloff 2009, 32-33; Ciancaglini 2008, 105-106.
16 Drower — Macuch 1963, 40.
17 Apart from Huyse, the connection of the words for 'messenger' to this ro ot is accepted by Gershevitch (1954, 51, § 342, note 1; he suggests the root *zgnd-) and Happ (1962, 198).
18 Note that such an ambiguity occurs in Iranian roots, e. g., *dra(n)j- 'to fix, fasten, hold' or *dra(n)c 'to oppress'; see Cheung 2007, 76 and 395-396.
19 Cheung 2007, 473; see also Kellens 1984, 21 (zgad- 'se précipiter (à cheval)'), 108 (zgad- 'aller à cheval'), and 1995, 69 (zgad- 'galoper'). However, not all scholars accept these meanings; cf. Bartholomae 1904, 1698; Narten 1986, 219-220; Hintze 1994, 285.
20 Bailey 1979, 355; Cheung 2007, 473.
well as ¡S(s)zyand(-) (written as ¡Sjynd-; Manichaean texts) 'mounting, riding' (all forms go back to *abi-zgad-),21 Pashto zyat- and zyas/st- 'to run'.22
The initial vowel in the Iranian words for 'messenger' is a prothesis, added before clusters consisting of a sibilant and a stop (however, this does not mean that d- in daydvSnc;/ daTdvSn; has the same origin; see below). This is a common feature in the Middle and Modern Iranian languages. It occurred especially in Manichaean Middle Persian and Manichaean Parthian as i- (written as ', rarely as ), but also as 3- (') in the Parthian inscriptions (see Table 2).23
Table 2. The prothetic vowel in Middle Iranian
Languages Forms24
Middle Persian (Zor.) spah (sp' h) 'army' skoh (skwh) 'poor'
Middle Persian (Man.) ispah ('sp'h) 'id.' iskoh ('skwh, 'sqwh) 'id.'
Middle Persian (inscr.) spah-bed (sp' hpt) 'army chief'
Parthian (Man.) ispaS ('sp'd) 'army' iskoh ('skwh), isko ('skw(w)) 'id.'
Parthian (inscr.) spaS-bed (spdpty and sp' dpty), 3spaS-bed ( sp dp[ty]) 'army chief'
In linking the Iranian words for 'messenger' to the root *zga(n)d-, we must deal with one more phonetic phenomenon. This is the change of the voiced alveolar fricative /z/ to the voiced postalveolar fricative /3/ (transliterated as z) before the voiced velar stop /g/ (or before the voiced velar fricative /y/ in Sogdian, where the change of voiced stops to voiced fricatives is a common feature). This phenomenon is found only occasionally in Iranian, but we must keep in mind that the occurrence of the cluster /zg/ is itself very limited.25 However, we can find confirmation for this phenomenon in the parallel cluster sk, where the voiceless alveolar fricative /s/ changes to the voiceless postalveolar fricative /J7 (transliterated as s) before the voiceless velar stop /k/. This development may be illustrated by the reflexes of Proto-Iranian roots *skap/f and *skarf in Middle Iranian languages (see Table 3).
Other Near Eastern words have been linked by researchers with izgand, asganda and/or daydvSnc;/daTdvSnc;, etc., but these comparisons are less certain.26 Some scholars
21 Henning 1936, 60; Gharib 1995, 103-104 and 121-122; Cheung 2007, 474; Sims-Williams 2021, 59.
22 Morgenstierne 2003, 101 (zyat- : zyas/st- 'to run'); Cheung 2007, 474 (zyastsl-/zyal- 'to run').
23 On prothesis in the Middle Iranian languages see Skjsrvo 1996, 531 and 2009, 201.
24 The data are taken from: Durkin-Meisterenst 2004, 86 and 92; Gignoux 1972, 33, 47, 63; MacKenzie 1986, 75 and 80; Nyberg 1974, 177 and 187.
25 Cf. Gray 1902, 232-233, § 933.
26 The following forms have been compared: 1. Akkadian (Neo-Babylonian) Asgandu (not asgandu), attested only as a family name; its original meaning of 'messenger' is not confirmed (it is based on a comparison especially with Mandaic asganda); see Brockelmann 1928, 9; Henning 1940, 35; Happ 1962, 199 and 1963, 98; Drower — Macuch 1963, 40; von Soden 1965-81, 80; Powell 1972; Macuch 1976, 38-39; Zadok 1976, 66 and 1979, 41; Mancini 1995, 85, note 15; Tavernier 2007, 512; Ciancaglini 2008, 105-106; Beekes 2010, 149; cf. Kaufman 1974, 38; rejected by Gelb et. al. 1956-2010, vol. A. 2, 427; Huyse 1993, 276, note 34; Sokoloff 2002, 112; 2. Jewish Aramaic ' ysqwndry (pl.) 'a game played with tokens (a kind of chess?)'; see Jastrow 1903, 57 (he gives three vocalized variants of the name: ' tsqundre, 'isqundre and squndre); on the Aramaic word see Sokoloff 2002, 122; 3. Armenian astandel 'to lead about, to lead up and down, to walk
have also adduced Iranian and Semitic words for 'messenger' to explain -t- in the form daravSn;: Manichaean Middle Persian and Manichaean Parthian azdegar 'messenger, herald',27 Sogdian dzdakre (written as ' z ' tqry; Christian texts) 'announcer, herald',28 as well as Official Aramaic ' zdkr' (emph.) 'inspector' (commonly treated as an Iranian word with the original meaning of 'herald')29 and Arabic askddr 'messenger, courier' (from Iranian, with a metathesis).30 This hypothesis assumes a rather improbable contamination of two words, *azgand- and *azddkdra, which would have resulted in the form *azdand-, rendered as daTdvSn; in Greek.31 The contamination would have taken place in an Iranian dialect.
Table 3. The phonetic change of /s/ to /J/ before /k/ in Middle Iranian
Languages Forms32
Proto-Iranian *skap/f 'to split, make a crack, crack' *skarf 'to stumble'
Middle Persian (Zor.) skdf- (sk p-) 'to split, burst' skarw- (sklw-) 'to stumble, stagger'
Middle Persian (Man.) iskdf- ('sk f-) 'id., iskafs- ('skfs-) 'to split up' iskarw- ('skrw-, 'sqrw-) 'to stumble, lurch'
Parthian (Man.) iskarfisn ('skrfysn) 'stumbling'
Sogdian (Buddh.) (g)skarufi or askruf ? ('skrwfi) 'trouble'
Sogdian (Man.) paskafs (pskfs-) 'to be split' (pa- < *pati-)
However, in the case of both Iranian words for 'messenger', the etymology was rather transparent to native speakers, and therefore a blending of this kind seems to be unlikely. What is more, the Greek words dayavSn; and daTdvSn; are so similar to each other that it is hard to believe that they might be of different origin. The consonant -t- must be explained rather as a miswriting.33 Accordingly, the lapsus calami was committed quite
about' or astandil 'to rove, to ramble, to wander'; cf. LSJ s. v. daTdvSn;; Beekes 2010, 149; rejected by Hub-schmann 1897, 109 and Huyse 1990, 95-96; cf. also Happ 1962, 200, note 3; on the Armenian astandel and astandil see Bedrossian 1875-79, 65; 4. Sogdian sstdnik ('st nyk; Buddhist texts) 'messenger' (treated as an etymon for Greek daTdvSn;); proposed by W. B. Henning in Gershevitch 1954, 248, Add. to § 997; see also Schmid 1962; Huyse 1993, 277; on the Sogdian word see Gharib 1995, 66.
27 Durkin-Meisterenst 2004, 85.
28 Gharib 1995, 93; Sims-Williams 2021, 47. However, instead of this word, Happ (1962, 200), following Eilers (1961-62, 225), gives "ztk'r; cf. Chantraine (1999, 127), which quotes an incorrect form: ztk r. This word has a different etymology and meaning: dzdtkdr or dzatkdr (written as ' ztk r in Buddhist and Manichaean texts and as ztq r in Christian ones) 'noble, nobility' (Buddhist and Manichaean) and 'soldier, military man, officer' (Christian); see Gharib 1995, 17 and 93.
29 Hinz 1975, 52; Hoftijzer — Jongeling 1995, 25.
30 Eilers 1961-62, 225-226.
31 See Happ 1962, 200 (he follows unpublished suggestions by W. P. Schmid); Schmid 1962; Huyse 1990, 95-96 and — not so decidedly — 1993, 277. Cf. also Chantraine 1999, 127.
32 The data are taken from: Cheung 2007, 344-347; Durkin-Meisterenst 2004, 86 and 92; Gharib 1995, 69 and 301; MacKenzie 1986, 80.
33 This is one of the explanations given by Huyse (1933, 277; following a suggestion by R. Schmitt). Cf. Happ (1962, 199), which is against this hypothesis; he emphasizes that the form daTdvSn; is too well attested in the sources.
early, so that it could spread in the Greek lexicographical tradition. We can theorize that the form daTdvSn; already existed in one of the influential ancient dictionaries, such as the comprehensive lexicon, consisting of 95 books, by Pamphilus of Alexandria (1st cent. A.D.). There is no doubt that the writing of AZT- instead of AZr- in the majuscule is pal-aeographically possible. An important cause of such an error may have been the fact that the Greek language does not have any other words beginning in day-, whereas there are many examples for dar-, e.g., dar^p 'star', aaru 'town', etc.
We can thus assume that both daydvSnc; and daTdvSn;, as well as aayydvSnc;, go back to an Iranian noun with the original meaning 'mounted messenger', based on the root *zga(n)d- 'to go on, gallop, mount'. The next question is when and from which language the word was borrowed. Admittedly, the Greek word is attested quite late, only in the Roman period; however, Plutarch and other authors may have followed some earlier sources. Therefore, it is possible that the term already existed in the Hellenistic or even in the Classical period. If so, the word may have been adopted either from Old Persian (alternatively, but less probably, from Median) or, later, from a Middle Iranian language, presumably from Middle Persian or Parthian.
The period and source language cannot be ascertained on the basis of the morphology. All three Greek forms, daydvSnc;, daTdvSn; and aayydvSnc;, belong to the same class of first-declension masculine nouns in This morphological feature is quite common in Iranian titles, names of functions, personal names and ethnonyms that were adopted by the Greeks from the Classical to the Early Byzantine periods. Especially nouns ending in -a or -a (nom. sg.), as well as — in the case of later borrowings from Middle Iranian — those without a final vowel, were adapted to this class; for example:
opoadYYn;, -ou m. 'benefactor of the Persian king; bodyguard' (since the 5th cent. BC): derived commonly from Old Iranian *v(a)rusanha- 'widely known';34 aaXdvn; in dSpaaraSdpav aaXdvn;, -ou m. 'a Sassanid title of a high civil and military official' (6th cent. A. D.): Middle Persian artestaran salar 'id' with the basic meaning of 'leader of soldiers, where aaXdvn; stands for Middle Persian salar 'leader, master'; Greek -Xdv- comes from -lar-, probably due to a dissimilation of l ... r > l ... n;35
Zwpodarpn;, -ou m. 'Zoroaster' (since the 5th cent. BC), also Za9pauarr|c; (hap. leg. in Diod. 1.94.2): Old and Young Avestan Zara9ustra- m. (nom. sg. Zara9ustra) 'id.', Middle Persian Zardu(x)st 'id.';36
Ssp^i, -ou m. 'Xerxes' (since the 5th cent. BC): Old Persian Xsayarsan- m. (nom. sg. Xsayarsa) 'id.';37
nsparii, -ou m. 'Persian' (since the 5th cent. BC): Old Persian Parsa- m. (nom. sg. Persa) 'id' as well as Parsa- (adj.) 'Persian.38
34 See Brust 2008, 494-497.
35 See Börm 2007, 145-146; Brust 2008, 45-48 and 569-571 (with an improbable claim that Greek aaXav- comes from the plural form *salaran, used in the function of the pluralis maiestatis, as a result of a haplological shortening); Sundermann 1986. The Middle Persian title artestaran salar is attested in Karnamag i Ardasir i Pabagan, i. e. Book of the Deeds of Ardashir, Son of Papak, and, in an Arabic transcription, in the History of the Prophets and Kings by al-Tabari.
36 Schmitt 1996, 91-98 and 2002b.
37 Schmitt 2000 and 2014, 285-285.
38 Schmitt 1996, 88-89 and 2014, 227.
The Greek words for 'messenger' may come either from an Old Persian noun with a nominal suffix -a- added to the root, or from a Middle Iranian form without a final vowel. Accordingly, we can reconstruct two possible source words: Old Persian *zganda-, with the nom. sg. *zganda,39 or Middle Persian/Parthian *zgand. In both approaches, the most difficult question is the presence of the vowel d-/-a- in the first syllable of daydvSnc;/ daTdvSn; and aayydvSnc;. If we had only the forms with d- in the initial position, we could assume that they originated in Middle Iranian *azgand with the prothetic vowel 9-, attested in the Parthian inscriptions (cf. aspaSbed 'army chief' quoted above). Incidentally, it is likely that this Middle Iranian form was borrowed into Mandaic as asganda.40
However, the form aayydvSnc; cannot be explained in this way. It seems more likely that both aayydvSnc; and daydvSnc;/daTdvSnc; go back to Old Persian *zganda- or to early Middle Persian/early Parthian *zgand- (or *zgand-, before the change of /zg/ to / jg/) without a prothetic vowel. It is reasonable to claim that the primary form in Greek was *aydvSn;. Since the initial cluster ay- /zg-/ was surely difficult to pronounce, a short vowel was added: either as a prothesis before the cluster ay- or as an anaptyxis between the consonants. Thus, two independent variants came into existence, namely daydvSn; and *aaydvSn;, which then underwent further transformations.
As discussed above, the variant daTdvSn; comes from daydvSn; as a result of mis-writing. In turn, the form aayydvSn; developed from *aaydvSn;. The distortion of aay- / sag-/ into aayy- /sayg-/ was probably caused by the influence of the following cluster -vS- /-nd-/. It may have been a kind of anticipatory assimilation: -ay- ... -avS- > -ayy- ... -avS- (/-ag- ... -and-/ > /-ayg- ... -and-/). However, we cannot exclude the possibility that it was, again, not a phonetic change, but only a miswriting in a source used by both the Cambridge Rhetorical Lexicon and the work by Claudius Casilo. The letter -y- /y/ could be added due to analogy to other words with -yy- (even with -aayy-) mentioned in the same entry: opoadyyn;, napaadyy^; and ayyapoc;.
To conclude, it is important to note that the Greek nouns daYdvSn; and aaYYdvSn; developed independently from the form *aYdvSn;, which goes back to Old Persian *zgan-da- or early Middle Persian/early Parthian *zgand- (or *zgand-) with an original meaning of 'mounted messenger' (derived from the Proto-Iranian root *zga(n)d- 'to go on, gallop, mount'). The initial d- in daYdvSn; is a prothetic vowel, while aaYYdvSn; received an an-aptyctic vowel -a- and, additionally, -yy- /-yg-/ instead of the expected -y- /-g-/ due to an assimilation or a miswriting. The form daTdvSn; is a secondary variant that comes from daYdvSnc, presumably because of an early lapsus calami of AZT- for AZr-. It is noteworthy that the Greek words exhibit some different features from their counterparts in Semitic: Jewish Aramaic ' izgadda and ' zgd, as well as Syriac izganda and izgadda, originated in Middle Persian or Parthian izgand (with the prothetic vowel i-), whereas Mandaic asganda was borrowed from Parthian (or, less probably, from Middle Persian) *azgand (with the prothetic vowel a-).
39 In the history of research, the Old Persian noun is reconstructed as *azganda- (Mancini 1995, 85, note 15; cf. Ciancaglini 2008, 105-106) or, more often, as *zganda- (e. g., Zadok 1976, 66 and 1979, 41; Tavernier 2007, 512).
40 Cf. Macuch 1976, 38-39; Mancini 1995, 85, note 15; Ciancaglini 2008, 105-106.
References
Bailey H. W. Dictionary of Khotan Saka. Cambridge, CUP, 1979. Bartholomae Ch. Altiranisches Wörterbuch. Strassburg, Trübner, 1904.
Bedrossian M. New Dictionary Armenian-English. Venice, Lazarus Armenian Academy, 1875-79. Beekes R. Etymological Dictionary of Greek. With the assistance of L. van Beek. Vol. I-II, Leiden — Boston, Brill, 2010.
Börm H. Prokop und die Perser. Untersuchungen zu den römisch-sasanidischen Kontakten in der ausgehenden
Spätantike. Stuttgart, Steiner, 2007. Brockelmann C. Lexicon Syriacum. Halle, Niemeyer, 21928.
Brust M. Die indischen und iranischen Lehnwörter im Griechischen, Innsbruck, Institut für Sprachen und Literaturen der UI, 22008.
Chantraine P. Dictionnaire étymologique de la langue grecque. Histoire des mots, avec un Supplément sous la
direction de A. Blanc et al. Paris, Klincksieck, 21999. Cheung J. Etymological Dictionary of the Iranian Verb. Leiden — Boston, Brill, 2007. Ciancaglini C. A. Iranian Loanwords in Syriac. Wiesbaden, Reichert, 2008. Cook E. M. A Glossary of Targum Onkelos. Leiden — Boston, Brill, 2008. Drower E. S., Macuch R. A Mandaic Dictionary, Oxford, Clarendon, 1963.
Durkin-Meisterenst D. Dictionary of Manichaean Middle Persian and Parthian. Turnhout, Brepols, 2004. Eilers W. Iranisches Lehngut im arabischen Lexikon: Über einige Berufsnamen und Titel. IIJ 1961-62, 5, 203-232.
Frisk H. Griechisches etymologisches Wörterbuch. Vol. 1-3, Heidelberg, Winter, 1960-72. Gelb I. J. et. al. The Assyrian Dictionary of the Oriental Institute of the University of Chicago. Vol. I-XXI, Chicago — Glückstadt, The Oriental Institute — Augustin Verlagsbuchhandlung, 1956-2010. Gershevitch I. A Grammar of Manichean Sogdian. Oxford, Blackwell, 1954. Gharib B. Sogdian Dictionary. Tehran, Farhangan Publications, 1995.
Gignoux Ph. Glossaire des inscriptions Pehlevies et Parthes. London, School of Oriental and African Studies UL, 1972.
Gray L. H. Indo-Iranian Phonology with Special Reference to the Middle and New Indo-Iranian Languages.
New York, CUP, 1902. Happ H. Zu àayàvÔn;, àaKavÔ^;, àatàvÔr|t; = "Bote". Glotta 1962, 40, 198-201. Happ H. Rec. Frisk 1960-72. IF 1963, 68, 95-99.
Henning W. B. Ein manichäisches Bet- und Beichtbuch. Berlin, De Gruyter (APAW 10), 1936. Henning W. B. Sogdica. London, Royal Asiatic Society, 1940.
Hintze A. Der Zamyad-Yast. Edition, Übersetzung, Kommentar. Wiesbaden, Reichert, 1994. Hinz W. Altiranisches Sprachgut der Nebenüberlieferungen. Wiesbaden, Harrassowitz, 1975. Hoftijzer J., Jongeling K. Dictionary of the North-West Semitic Inscriptions. Vol 1-2, Leiden, Brill, 1995. Horn P. Grundriss der neupersischen Etymologie. Strassburg, Trübner, 1893.
Hübschmann H. Armenische Grammatik. Part l: Armenische Etymologie. Leipzig, Breitkopf & Härtel, 1897. Huyse Ph. Persisches Wortgut in Athenaios' "Deipnosophistai". Glotta 1990, 68, 93-104. Huyse Ph. Von angaros und anderen iranischen Boten. HS 1993, 106, 272-284.
Jastrow M. A Dictionary of the Targumim, the Talmud Babli and Zerushalmi, and the Midrashic Literature.
Vol. 1-2, London — New York, Luzac — Putnam's Sons, 1903. Kaufman S. A. The Akkadian Influences on Aramaic. Chicago — London, UCP, 1974. Kellens J. Le verbe avestique. Wiesbaden, Reichert, 1984.
Kellens J. Liste du verbe avestique, avec un appendice sur l'orthographe des racines avestiques par E. Pirart.
Wiesbaden, Reichert, 1995. MacKenzie D. N. A Concise Pahlavi Dictionary. Oxford, OUP, 21986. Macuch R. Zur Sprache und Literatur der Mandäer. Berlin — New York, De Gruyter, 1976. Mancini M. A proposito di prestiti partici in mandaico: hambaga. AION 1995, 55, 82-95. Mancini M. Etimologia e semantica del gr. ayyapoc;. Glotta 1995-96, 73, 210-222.
Morgenstierne G. A New Etymological Vocabulary of Pashto, compiled and edited by J. Elfenbein, D. N. MacKenzie, N. Sims-Williams, Wiesbaden, Reichert, 2003. Narten J. Der Yasna Haptanhaiti, Wiesbaden, Reichert, 1986. Nyberg H. S. A Manual of Pahlavi. Part 2, Wiesbaden, Harrassowitz, 1974.
Powell M. A. Der neubabylonische Familienname Asgandu und die Urkundengruppe Nbn 314, TCL xii 122, Nbn 668. ArOr 1972, 40, 124-129.
Schmid W. P. Zu griech. dcrrdv5r|<; "Bote". Glotta 1962, 40, 321.
Schmitt R. Onomastica Iranica Platonica, in: Ch. Mueller-Goldingen, K. Sier (eds), AqvaiKa. Festschrift für C. W. Müller. Stuttgart — Leipzig, 1996, 81-102.
Schmitt R. Xerxes i. The Name, in: Encyclopedia Iranica 2000. Available at: https://www.iranicaonline.org/ articles/xerxes-1-name (accessed on 30 August 2020).
Schmitt R. Greece XII. Persian Loanwords and Names in Greek, in: Encyclopedia Iranica 2002a. Available at: http://www.iranicaonline.org/articles/greece-xii (accessed on 30 August 2020).
Schmitt R. Zoroaster i. The Name, in: Encyclopedia Iranica 2002b. Available at: https://www.iranicaonline. org/articles/zoroaster-i-the-name (accessed on 30 August 2020)
Schmitt R. Wörterbuch der altpersischen Königsinschriften. Wiesbaden, Reichert, 2014.
Sims-Williams N. A Dictionary: Christian Sogdian, Syriac and English. Wiesbaden, Reichert, 22021.
Skj»rv0 P. O. Aramaic Scripts for Iranian Languages, in: P. T. Daniels and W. Bright (eds.), The World's Writing Systems. New York — Oxford, OUP 1996, 515-535.
Skj»rv0 P. O. Middle West Iranian, in: G. Windfuhr (ed.), The Iranian Languages. London — New York, OUP 2009, 196-278.
Sokoloff M. A Dictionary of Jewish Palestinian Aramaic of the Byzantine Period, Ramat-Gan, Bar Ilan UP, 1990.
Sokoloff M. A Dictionary of Jewish Babylonian Aramaic of the Talmudic and Geonic Periods, Ramat-Gan, Bar Ilan UP, 2002.
Sokoloff M. A Syriac Lexicon. A Translation from the Latin, Correction, Expansion, and Update of C. Brockel-mann's Lexicon Syriacum. Winona Lake — Piscataway, Eisenbrauns, 2009.
Sundermann W. Artestärän sälär, in: Encyclopedia Iranica 1986. Available at: https://iranicaonline.org/ar-ticles/artestaran-salar-chief-of-the-warriors-a-high-ranking-title-in-sasanian-times-see-artestar- (accessed on 30 August 2020).
von Soden W. Akkadisches Handwörterbuch. Vol. 1-3, Wiesbaden, Harrassowitz, 1965-81.
Tavernier J. Iranica in the Achaemenid Period (ca. 550-330 BC): Lexicon of Old Iranian Proper Names and Loanwords, Attested in Non-Iranian Texts, Leuven — Paris — Dudley (MA), Peeters, 2007.
Zadok R. On the Connections between Iran and Babylonia in the Sixth Century B. C. Iran 1976, 14, 61-78.
Zadok R. The Jews in Babylonia during the Chaldean and Achaemenian Periods according to the Babylonian Sources. Haifa, HUP, 1979.
Received: September 29, 2020 Accepted: January 24, 2021
