Origin of the Mandaic Script

8/12/2019 Origin of the Mandaic Script

1/11

Iranian Scripts for Aramaic Languages:The Origin of the Mandaic Script

CHARLES G HXBERLDepartment of Near Eastern Languages and Civilizations

Harvard University6 Divinity Ave.

Cambridge, MA [email protected]

The unique cursive script still employed by the Mandaeansof Iraq and Iran, whichis unlike any other script found in the modern Middle East, may provide a clue to theobscure origins of their written literature and their emergence as a distinct religioustradition.Comparison with ancient scripts rom the regions where the Mandaeans rfound today indicates that the Mandaic script is a product of the late Parthianperiod andmore specifically the second century C.E. and has its closest affinities with a groupof scripts ranging from Anatolia and the Caucasus in the north to Characene andElymais in the south, all of which appear o derivefrom or to be heavily influenced bythe Parthianchancery script.The association of the Mandaeans with the later Arsacidsis corroboratedby their own legends and their textual tradition.

THE M ND IC SCRIPTearly every discussion of Mandaean originshas dealt with the issue of their peculiarscript. While some scholars have touted its

importance and others have questioned its relevance,no scholar of Mandaean origins has failed to take thescript into consideration. The reasons for this aremanifold. The Mandaeans themselves consider theirscript to be as sacred as their literature; according toone of their traditions, it predates mankind (Drower2002: 240). The fact that the script of the modernmanuscripts is not appreciably different from that ofthe earliest manuscripts 2 illustrates how faithfullythe Mandaeans have transmitted their sacred litera-ture across the centuries.3 As the vehicle of their re-ligious literature, inextricably tied to their traditions,

ICf. the Jewish tradition in the Sefer Yetzirah 1:2:10-12) thatthe 22 letters of the Hebrew alphabet were one of God s firstcreations and, with the ten Sephiroth, composed the foundation ofall things.2 The earliest manuscripts date to the eighth century C E ; seeNavch 1970: 33.

This is not the case for profane or magical texts, which arewritten in a later, more colloquial form of the language; seeNOildeke 1898: 143.

the script is one of the most characteristic featuresof the Mandaean community. As such, it can be fruit-fully compared with other epigraphic, historical, an dlinguistic evidence to provide us with a rough ter-minus ante quem for the formation of their commu-nity even if it is quite feasible that the Mandaeansadopted their script centuries after their emergenceas a religious tradition. 4 Once the age of the scriptis determined, a closer examination of the relation-ship of the script itself to its contemporaries can alsoilluminate the cultural and social circumstances pre-vailing at the time. of its adoption.

The Mandaic script is unique among the scripts ofthe ancient Near East, and it is precisely this unique-ness that makes it so difficult to contextualize. Evenrelatively basic data, such as its relation to the otherAramaic scripts or the date of its adoption by the

This is more often the norm than the exception. The twocultures to which the Mandaeans are most frequently compared,the Jews and the Zoroastrians, adopted their most characteristicscripts centuries after the emergence of their religious traditions.In fact, the Manichaeans were the first to adopt a particular scriptas a vehicle for their religious texts, which reproduce (presumablypreexisting) portions of the Mandaean liturgy; see Siive-SWider-bergh 1949: 137-62 and Colpe 1964.

53


2/11

CHARLES G. HABERL

Mandaeans, are unknown and have been hotly de-bated. The largest body of physical evidence for theMandaic script consists of a corpus of incantationbowls that have been discovered in various locationsin Mesopotamia, which has been dated to the periodcoinciding with the emergence of Islam. 5 Addition-ally, there is a smaller corpus of lead amulets, whichoffer the earliest examples of any form of the Man-daic script. The earliest of these were inscribed in thefourth or even third century, although they continueto be made even today. 6 Beyond these few examples,however, it is difficult to say when and where theMandaeans adopted their script. The Mandaic scriptis also unique by virtue of the fact that even its ear-liest inscriptions feature a fully developed systemof vowel letters. In this respect, it has no direct par-allel among the scripts of several Aramaic dialectsthat were in use in this region during the second andthird centuries C E 7

There are three major schools of thought about theorigins of the Mandaic script:I. The Mandaic script is derived from the Pal-myrene or Nabataean script, like other cursiveligatured scripts in the region, such as Syriac or

Arabic. This assumes that the Mandaic scriptoriginated in Syria and traveled west with theMandaeans. This is the oldest view, espousedby Mark L idzbarski and Rudolf Macuch amongothers (Lidzbarski 1909; Macuch 1971).

2. The Mandaic script is derived from the Elymaicscript, which is attested in the later coins of Ely-mais (modern Khuzestdn in Iran), from the sec-ond century C.E., when Elymais was a clientkingdom of the Parthian Empire. This vieworiginated with Joseph Naveh (Naveh 1970).83. The Mandaic script is derived from the Cha-

racenean script, which is imperfectly knownMost of the bowls that have been published were acquired

from dealers, and very few have any association with an archae-ological site, which makes their dating difficult. The bowls exca-vated at Nippur belong to the seventh century; see Montgomery1913: 13-14 and Hunter 1994: 605; one bowl, excavated at Ana,belongs to the eighth century, according to Hunter 1994: 607.

6 For further details about this important corpus of inscrip-tions, see Lidzbarski 1909: 349-73; Macuch 1967: 91-203 and1968: 34-72; Caquot 1972: 67-87; Naveh 1975: 47-53; and

Greenfield and Naveh 1985. The most recent discussion of thelead rolls is MUller-Kessler 1999.7 That is to say, the Mandaic vowel letters are not inatreslectiones see the discussion below. For this group of inscriptions, see Bivar and Shaked 1964.

from a small number of legends found on coinsminted in Mesene (southern Iraq) during thissame period. This view was first suggested byP. W. Coxon (Coxon 1970).

Both Coxon and Naveh have convincingly arguedthat Mandaic, Elymaic, and Characenean are east-ern Aramaic scripts 9 and cannot be derived fromNabataean. The argument for the western origin ofthe script rests upon two bases-the ligatures andthe superficial resemblance of the' in Nabataean andMandaic. One argument against the theory of a west-ern origin (which I have not seen raised in the litera-ture) is the fact that Mandaic makes no distinctionbetween final and non-final forms, which is a diag-nostic feature of not only the cursive western MiddleAramaic scripts (Nabataean and Syriac), but also theJewish (Hebrew) script as well. Furthermore, the Pah-lavi script is famously ligatured, and this feature is ab-sent from it as well. The presence of final forms is animportant diagnostic feature distinguishing the ductusof the western Middle Aramaic scripts from that ofthe eastern Middle Aramaic scripts.

ELYMAIC AND CHARACENEANNaveh's argument, that Mandaic is a more devel-

oped form of the Elymaic script, is less convincing,as he is unable to offer an example of the immediateprecursor of either script for the purposes of com-parison. The closest he can provide is the script of theAgoka inscription, which predates this material bynearly half a millennium. Furthermore, the earliestMandaic material is several centuries later than theElymaic material. For this reason, Coxon tentativelysuggests that the Characenean coin legends, which ap-pear later than the Elymaic material and more closelyresemble epigraphic Mandaic, are a better candidate.In the final analysis, neither Elymaic nor Charace-nean is a satisfactory candidate for the ancestor to theMandaic script, for precisely the same reasons thatthey attracted Naveh and Coxon-they are obscure,provincial, and late. These two were employed ex-clusively within the Parthian client kingdoms of Ely-mais and Mesene (Characene), respectively, and theiruse appears to be restricted solely to the local dynas-ties. Unlike the Mandaic script, these two scripts rep-resent a historical curiosity-an evolutionary dead

9 Or, more properly, are representative of the eastern ductus ofthe Middle Aramaic scripts.

54 BASOR 341


3/11

IRANIAN SCRIPTS FOR ARAMAIC LANGUAGES

end in the development of the Aramaic scripts in theeast.

Coins with Elymaic legends do not appear untilafter the local dynasty had been supplanted by a lineof kings with Parthian names, in the latter half of thefirst century. Prior to this point, the legends foundon the coins of the local rulers of Elymais, the Kam-naskirids, were in Greek alone (Gu6lpin 1965: 19-26).11 Greek appears to have been discarded through-out the Parthian Empire in favor of Aramaic an dParthian, a change that became effective by 53 C EBivar associates this switch with the revolt of the cityof Seleucia against Parthian rule between 35 and42 C.E. (Bivar 1967: 517). From the middle of thefirst century until the end of the Parthian Empire, thelocal rulers of Elymais were alternately vassals ofthe Parthian Empire and independent sovereigns. Itcannot be coincidental that, after Ardax9ir defeatedthe last member of this dynasty in 221 C.E. andseized Elymais, the Elymaic script disappears fromhistory.12 In short, the brief span during which thescript was used coincides with a period of Parthianclients in Elymais and an increase in the use of theAramaic language and the Parthian chancery scriptthroughout the empire, at the expense of Greek.

Mesene did not begin to issue coins with Aramaiclegends until much later, starting in the second halfof the second century C.E., following the reassertionof Arsacid influence in this region, which had brieflybeen occupied by Trajan and whose sovereigns wereoften in revolt against their Parthian suzerains. Fourof the published coins bear the legend ybygn'y mlk' King lbignai, whom Lidzbarski (1909: 87) dateswithin the period between 150 and 224 C.E. 3 Altheim

Sellwood (1983: 308-9) suggests that these kings may havebeen a junior branch of the Arsacid family. This hypothesis wasfirst advanced by Rostovtzeff (1936: 118), on the basis of the nu-mismatic evidence.

According to Hansman (1998), the title Kamnaskires ma ybe related to the Elamite title (qa-ap-nu-i9-ki-ra> kapnugir treasurer.12 Coxon (1970: 30) argues that it is quite feasible that the

Mandaeans could have adopted one of these scripts a century ormore after they are first attested, but I fail to see how this is pos-sible with the scripts under consideration here, which were ob-scure, provincial, and most importantly, obsolete after the fall ofthe Arsacids.

13 Bold letters indicate transliterations of Mandaic charactersand those in related scripts. The phonemes that they represent aretranscribed within /slashes/ (see below). The convention adoptedhere for transliterations of Mandaic characters is the one found inD)rower and Macuch 1963: xii.

suggests an even later date, on the basis of a passagein Ishodad of Merv s commentary on Genesis; accord-ingly, he claims that the Characenean script must havebeen created at some point between the accession ofArdaxgir-i Pdbagan in 208 C.E. and the defeat of theArsacids in 224 C.E., whereupon Ardaxir deposedthe local dynasty (Altheim 1969: 31-32). Sellwoodalso identifies the final series of Characenean tetra-drachms with the period around 200 C.E. (Sellwood1983: 313-14). The fact that the two scripts to whichthe Mandaic script is most frequently compared-namely, Elymaic and Characenean-are products ofthe latter half of the Parthian period is undeniablysignificant.

COMPARISON WITH THEPARTHIAN SCRIPT

According to Jorunn J. Buckley (2002: 4), thecolophons in the eft Ginza indicate that the oldestportions of the Mandaic textual tradition date sev-eral generations before the time of Mani, i.e., to thebeginning of the third century c.E. at the latest.' 4 Th eprimacy of the Mandaean tradition relative to theManichaean is confirmed by the Coptic ManichaeanPsalm-Book, which contains material adapted fromthe Mandaean masiqta or death mass, found alsoin the eft Ginza (Sdve-Soderbergh 1949: 163). Inthe absence of any evidence suggesting that theMandaeans preserved their religious literature in anyscript other than the present one, we must assumethat the colophons provide us with a terminus antequem for the composition of the sacred literature inthe Mandaic script. The evidence of the colophons iscorroborated by the Mandaeans' own historical tra-ditions, which maintain that they settled in the re-gion from parts farther north at the behest of a kingArdawan, whose descendants were subsequently de-feated by the Sassanians (Drower 1953: 3-14). It can-not be mere coincidence that both the Mandaeantextual tradition and their own historical traditionssituate them in the vicinity of Mesene and Elymaisduring the latter half of the Parthian Empire. Conse-quently, it seems only logical to seek the context ofthese three scripts within the Parthian Empire duringthis period.

14 Buckley has traced an unbroken line of copyists back to awoman by the name of Slama, daughter of Qidra, who lived sev-eral generations before the death of Mani.

2006 55


4/11

CHARLES G. HABERL

During the 1970s, when Naveh, Coxon, and Ma-cuch debated the chronology of these scripts, mostof the Aramaic material from the Parthian periodwas still poorly understood or as yet undiscovered(Skjxervo 1996). Yamauchi (1970: 76) briefly men-tions the Parthian ostraca found at Nisa (first centuryB.C.E.; Nisa is located on the Caspian Sea in modernTurkmenistan), but immediately dismisses them as ir-relevant to the question of Mandaean origins. Naveh(1997: 128-3 1) also mentions the Nisa ostraca in hisbook on the history of the alphabet, but his treat-ment of them is superficial. His omission is surpris-ing, considering that the script of the Nisa ostraca ismuch more proximate, in terms of both time andspace, to the bulk of the attested eastern Middle Ara-maic scripts than the script of the Agoka inscriptionto which he compares them.15 When one considersthe Iranian material (such as the Nisa ostraca or thethird-century C.E. Paikuli inscription from northernIraq), the affinities these scripts share become obvi-ous (table 1 :16

SThe Elymaic form of this letter is clearly a de-velopment of the form preserved in Parthian,with the addition of a ligature along the left sideof the character. Naveh (1970: 34) compares theElymaic ) to the ) of Hatra in northern Iraq.While the Mandaic ) is certainly more devel-oped than the Elymaic ', this does not neces-sarily indicate its descent from the Elymaic, asthis same form has been observed elsewhere.17b The line beneath the letter, which continues tothe right, is an innovation upon the originalletter found only in the Elymaic inscriptions.Mandaic does not share this feature.

g The Tang-e Sarvak g differs from the Parthiang through the addition of a horizontal stroke atthe top, not found in the Mandaic g.d The dot found beneath the Parthian d (added todistinguish it from r and 1) has been joined to

15 The bilingual Aramaic-Greek Agoka inscription was dis-covered near Jalalabad in eastern Afghanistan and dates to thethird century B.C.E.; see Naveh 1997: 137.16For the purposes of this discussion, Elymaic shall referspecifically to the inscriptions of Tang-e Sarvak and Shimb5ir; Parthian, unless otherwise identified, refers primarily to theinscriptions from the second and third centuries c.E., some ofwhich (such as the Paikuli inscription, 292 c.E.) date to the periodof the early Sassanians; and Mandaic refers principally to thescript of the lead amulet inscriptions. 17ost notably in the monumental Nabataean script; the simi-larity of the monumental Nabataean ' to the Mandaic ' was thebasis for many of the early arguments identifying the two scripts.

the body of the original character with a ligaturein the Elymaic script, just as it has at Armazi,at the other extreme of the Parthian Empire, inGeorgia (Naveh 1997: 142).'8 Neither the lig-ature nor the dot is found under the Mandaic d,which preserves the same form as the Nisa d.19t If the readings of this infrequently attestedcharacter are correct, the Elymaic t could onlyhave developed from the Parthian t. The twohalves of the Parthian t have been joined by aligature in Elymaic along the bottom of thecharacter. One of the attested forms of the Man-daic t found in the script of incantation bowlsalso appears to derive from the t of the laterParthian inscriptions. Unlike the Elymaic t,however, the ligature spans the top of the char-acter rather than the bottom. 20

y The Elymaic y, consisting of a single stroke oreven a simple dot, represents a simplification ofthe original two-stroke y, which survives inMandaic.21k The cursive Elymaic k represents a simplifica-tion of the angular k, the original form of whichsurvives in Parthian and Mandaic.m The Elymaic m has become simplified to an xin the Tang-e Sarvak inscription. Mandaic re-tains the original form.22n n in Elymaic and Mandaic are identical andappear to have developed independently fromthe same origin.It is highly unlikely that the angular, three-stroke I found in Mandaic developed from thetwo-stroke Elymaic C; both forms, however, arefound in the Nisa ostraca. Like its Parthian an-alogue, Elymaic ' resembles r and d, whereas

IS Note also the diacritic over the r at Garni. Naveh attributesthe diacritics at Garni to Syriac influence, but they are not attestedin the contemporary Syriac inscriptions. In fact, the earliest attes-tation of the dots distinguishing d from r in Syriac is found onlyin an Estrangela manuscript dated 411 C.E.; see Healey 2000.19If the two share the same origin, this suggests that the Man-daic script is more conservative than the others in this regard, hav-ing not adopted this convention. 21oxon (1970: 26) notes that no affinity pertains to theMandaic form of the letter, but adds that he is wary of conclu-sions drawn from the reading of this character in the Elymaicmaterial. Given that the t is uncommon in any Semitic language,his caveat need not apply only to Elymaic.21 It is not clear to me how the non-ligatured yod could bean offshoot of the Elymaic yod which is reduced to a dot, asNaveh (1970: 36) asserts.

Naveh concedes this point in his discussion of the Elymaicevidence (1970: 36).

56 BASOR 341


5/11

167:i .

M 211O 0 -N 1: -3> cn~j> '

Icdj )~ >N x U_S

c-

UU

cz

0

CO

Q

ca

uc;

q i-'-'fi 2S)0 .c1cJ I', it3

-n

cco ; >N ~

-

Qn 8n1~

S-i >E Z r

00 -

S.,

. .S..

-0

0 S : 7 -

-oI- *f

o

F-

00

a

Ln 0a n -%

-0 tO 7 E '- CA U) )V)

n f- JQ 0 M F)


6/11

CHARLES G. HABERL

the Mandaic has not been conflated with theseother consonants.p Coxon (1970: 28) notes that the full curve to theleft at the foot of the letter in the Shimbar in-scriptions is a feature of the Elymaic group and

is not repeated elsewhere.s The three-stroke Parthian s is clearly the fore-runner to the four-stroke s found in Book Man-daic and the later Elymaic inscriptions. The sfound in Bowl 079M in the British Museumcollection has the same number of strokes as theParthian s; Bowl 087M, on the other hand, moreclosely resembles the four-stroke Book Man-daic s. The final stroke in the Mandaic s riseshigher than that in the Parthian s, and curvesoutward, away from the character. In contrast,the s attested at Shimbar has the same numberof strokes as the Parthian; but the final strokedescends even lower before it curves back to-ward the character, a trend that becomes evenmore pronounced in the Tang-e Sarvak script.In the latter script, the s assumes a more cursiveand elaborate form than it does elsewhere.

q No indisputable example of q appears in theElymaic corpus.2 3 The Mandaic q resemblesthe Parthian q and those found elsewhere, 24 arectangular form with rounded corners.

r In the Elymaic material, the horizontal strokeof the r joins the perpendicular either at thetop, as in the d, or the center, making it re-semble the . This confusion is also typical ofthe Parthian material (in the later Iranian scripts,r and I eventually merge with w and even n) .The corners of the r are rounded, unlike theMandaic r. Mandaic shares with other earlierSemitic scripts a confusing similarity betweenthe r and the d, unlike Elymaic, Parthian, or thescript at Garni, all of which have adopted a con-vention to distinguish between the two.9 The form of the Elymaic , is perhaps the mostunique and difficult to explain of all the char-acters in the script. It certainly bears no resem-blance to the 9 found in Parthian, Mandaic, or

23 Naveh (1970: 36) rejects the sole example of q found in thecorpus on the grounds that it cannot be correct, and suggests analternate reading of the material which yields a q that bears aremarkable similarity to the Elymaic g.

4 Coxon (1970: 28) notes the similarity between the q inMandaic and epigraphic Syriac; the Characenean q, which is re-lated, agrees with incantation bowls. Naveh (1970: 35) comparesthe q, like the , with that attested at Hatra. Naveh's northernMesopotamian q is nearly identical to the Parthian q.

other contemporary scripts. The Aramaic 9gen-erally develops along two predictable axes-ei-ther a line along the bottom or (less frequently)along the left-hand side of the character. At firstglance, the Elymaic 9 appears to be upsidedown.One of the forms attested at Shimbar mayprovide a clue to the development of this un-usual form. Its central line is connected to theleft-hand vertical at its base. This suggests thatthe Elymaic 9 developed along the left-handvertical, which was joined to the other two ver-ticals by a ligature at the top. Gradually theseverticals detached themselves from the left-hand vertical, resulting in the form attested atTang-e Sarvak.The Mandaic 9 is no less unusual. Naveh(1970: 37) maintains that the two forms arerelated, but correctly notes (1972: 293-304)that the Mandaic 9 finds its closest parallels innorthern Mesopotamia, at Hassan-Kef in T5ircAbdin and Garni in Armenia, both in Parthianterritory and both dating to the second half ofthe second century, much like ) and q. It isimportant to note, however, that unlike the Ely-maic 9, none of these forms could have devel-oped along the left-hand vertical.

Most of the characters in the Elymaic script areeither identical with the forms found in contempo-rary Parthian inscriptions (such as w, z, h, I, r, andthe t at Tang-e Sarvak), or have clearly developedalong the lines of other cursive scripts found in theNear East during this period (to this category belongthe h and possibly the Shimbdr t, which have closeparallels within the extreme cursive Post-Herodianscript discovered at Wadi Murabba'at in the WestBank).25 In light of this data, a few features of theunderlying ductus of this group of scripts may betentatively identified:

The scripts are cursive in comparison withother Middle Aramaic scripts, with fewer sharpangles, a reduction in the number of strokes formost characters, and ligatures between individ-ual strokes in others, with the result that severalletters (particularly d, , and r) are conflated.

25 Jo Ann Hackett first indicated to me the similarity of theElymaic form with the Post-Herodian; see Yardeni 1997: 191.This same h also has a parallel in the final h in cursive Nabataean,albeit probably as the result of independent development.

58 BASOR 341


7/11


* Due to the confusion arising from the simplifi-cation of these characters, a diacritic (taking theform of a dot or a stroke beneath the d) wasadopted to distinguish between them. 26 * Ligatures feature prominently in these scripts,connecting the individual strokes of each char-acter and, in the case of the relative particle zy(Mandaic d), connecting one consonant withanother. * In contrast with the western ductus of the Mid-dle Aramaic scripts, such as Jewish (Hebrew),Nabataean, and Syriac, no distinction is madebetween the final forms and non-final forms ofthe characters.

What arises from the ensemble of these features isa discrete group of scripts arrayed in a broad geo-graphic and chronological range, from the region ofthe Caucasus and eastern Anatolia, to Elymais andCharacene in southern Mesopotamia. Members ofthis group are first attested during the second centuryC E and survive today in the form of the scripts em-ployed in the Mandaean and Zoroastrian sacred lit-erature. The dating of the members of this group inrelation to one another and the precise genetic rela-tionships between them are more difficult to ascer-tain from comparison alone. My first impression isthat the Mandaic script is one of the most conserva-tive members of this group. While it is more heavilyligatured than the Elymaic and Characenean scripts,it is not as heavily ligatured as Book Pahlavi; nor isit as cursive as the Book Pahlavi, Elymaic, or eventhe later Parthian script.27 It also does not participatein the innovation adopted by the others to distinguishbetween the d and the r.

It is true that the script of the Mandaic lead amuletsis superficially most similar to that of the Charace-nean coin legends. Nonetheless, a number of featuresdemonstrate the more conservative nature of the am-ulet script, despite the fact that it is chronologicallylater. Note that the heads of the b, the k, and the rhave been simplified to a horizontal stroke in the coinlegends, with the result that these characters resemble

6 See fn. 18 above. This feature is also found in Syriac, whichis not a member of this group, but not in Mandaic or Book Pahlavi,which are. This suggests that we are dealing with a conventionthat was adopted by the others at some indeterminate date ratherthan one that was inherited.

27 Coxon's suggestion (1970: 29) that the Elymaic has adheredto the archaic lapidary script more rigorously than the cursiveMandaic seems less likely after comparison with the Parthianmaterial.

one another, while each maintains its separate iden-tity in the amulet script. Intriguingly, these same fea-tures are maintained in the Mandaic book hand, whichclearly distinguishes between these characters. Thus itis impossible that either the book hand or the amuletscript could have developed from the Characeneanscript. Both Naveh (1970: 36) and Coxon (1970: 26)assume that the Mandaic developed from the trian-gular Characenean t but the Characenean t is so idio-syncratic, and so unlike the form of the Parthian t(which offers a much more suitable candidate for theancestor to the Mandaic form), that it is most likelythe result of independent development.Within the corpus of incantation texts, the scriptfound in the lead amulets more closely approximatesthat found in the manuscripts than the bowl script,despite the relative antiquity of the lead rolls in com-parison with either the incantation bowls or the ear-liest manuscripts. This is contrary to expectations, asthe process of engraving an inscription upon a leadroll is quite different from that of writing it upon aterracotta bowl or manuscript page, and consider-ing the fact that the bowls are much closer in dateto the earliest surviving manuscripts. Naveh (1970:33) noted that the cursive forms in the lead rolls andthe bowl texts could only be derived from the bookhand-despite the fact that our earliest manuscriptsdate from the eighth century and most of the bowlsare presumably pre-Islamic. 28 As the Mandaic bookhand has not changed appreciably in the interveningcenturies, it becomes clear that it is a very conserva-tive script, much like its cousin the Jewish (Hebrew)script.29 Considering that Mandaean attitudes towardtheir liturgical language and sacred literature parallelJewish attitudes toward theirs, it is not surprising thatthe Mandaic script has been passed down through thegenerations with the same degree of care. 30

The historical circumstances and the comparisonof the scripts suggest that the model for the ductus ofthis group of scripts was provided by the Parthianchancery script. Additional support for this argument

28A few of the bowls show Arabic influence, but the scholarlyconsensus is that the bulk of the corpus dates to the period imme-diately preceding the Islamic conquest; see Segal and Hunter2000:21-22.

29 Note that the script of the Dead Sea Scrolls is immediatelycomprehensible to any scholar familiar with the conventionalHebrew script, having not changed appreciably in the interveningtwo millennia.

30 The Mandaic bowl script, on the other hand, must representthe development of a contemporaneous cursive script, analogousto the Post-Herodian cursive script attested at Wadi Murabba'at.

2006 59


8/11

CHARLES G. HABERL

is provided by the ligatured zy in Elymaic, whichresembles the Mandaic relative pronoun d; Navehargues that it can hardly be doubted that Mandaic ddeveloped from Elymaic zy and not the other wayaround. This same ligature, however, is much morecommon than Naveh would have us believe; it isfound in the other Aramaic scripts adapted to Iranianlanguages, such as the Psalter Script and Book Pah-lavi.3 It is much more likely that this ligatured zyoriginated in Iran and spread from there to theseother provincial scripts, rather than the reverse. 32

IRANIAN AND MANDAICVOWEL LETTERS

Inscriptions in the Parthian language (as opposedto Aramaic inscriptions written by Parthians) are firstattested from the mid-second century C E onward.Among the orthographic features that distinguishthese inscriptions from their Aramaic counterparts(such as the confusion between h and .h) s a systemof vowel letters, which has supplanted the matreslectiones found in the Aramaic scripts. The vowelletters developed along lines remarkably similar tothose along which the Greek vowels developed. Theglottal stop ) and the voiced pharyngeal Care not partof the Iranian phonemic repertoire, and consequentlythe characters representing them were pressed intoservice to indicate either open vowels such as a and5 or closed vowels such as i, i, e, and e, respec-tively. Initially, only word-initial and word-final vow-els were indicated in this way, but these vowel lettersquickly came to be employed word-internally as well.Likewise, the semi-vowels y and w were employedto indicate the closed front vowels i, i, e, and e andthe closed back vowels o, 6 u, and U, respectively,just as in contemporary Aramaic inscriptions, albeitin a more consistent manner.

What distinguishes these characters from theequivalent matres lectiones in Aramaic is not merelythe consistency of their application, but the fact that

31 See examples in Skjxervo 1996: 523-24. The ligature ap-pears in Book Pahlavi only before enclitic pronominal suffixes.32 An interesting cross- inguistic parallel is found in Taiwan.Taiwan has historically had strong ties to Japan, despite the factthat its official language is Mandarin, which is written using amuch more traditional form of the characters used on the main-land. In Taiwan, the Japanese hiraganacharacter no is often sub-stituted for Chinese de, a particle that indicates possession in asimilar, although not identical, manner as the Mandaic relativepronoun d-; the hiragana yllabary is a cursive, simplified versionof the more traditional characters used by the Chinese.

they are exclusively used to indicate vowels and semi-vowels, just as in Greek, whereas the matres lectionesare consonants, first and foremost, which are occa-sionally em ployed to indicate the presence of a vowel.This distinction is subtle, yet crucial for the under-standing of the development of these scripts. All of theAramaic dialects attested during the Arsacid and Sas-sanian eras employ matres lectionesrather than vowelletters, with the sole exception of Mandaic.

The fact that even the earliest Mandaic inscrip-tions feature a phonetic orthography, with a fullydeveloped vowel letter system, should give mostAramaic scholars pause. Frankly, it is surprising thatscholars such as Macuch and N61deke-who wereboth Iranists as well as Semitists-never challengedthe assumption that the Mandaeans developed thissystem on their own. 33 Note that the letters a, i, u,and h, while formally derived from Aramaic conso-nants, 34 are nonetheless used only to indicate vowelsor semi-vowels, never consonants. Mandaic a is neverused to represent the glottal stop /?/ which is indi-cated by a separate letter C;Mandaic h is never usedto indicate /h/, which is represented by the letter h.In addition to indicating the third masculine singu-lar enclitic pronoun, which is its most common func-tion, h was also employed to indicate the indefinitemorpheme (Macuch 1965: 207) and the Iranian dzdall three of which, judging by their perennial confu-sion with the plural morpheme -ia in the incantationtexts, were most likely pronounced */i:/. 35 Note thatMandaic h is also never used to indicate final /a:/,unlike its equivalent in other Aramaic scripts; like-wise, a is used to indicate /a/ or /a:/ in any position,which is most certainly not the case elsewhere inAramaic, where it may indicate almost any voweland is found primarily in word-final position.

33 Note Blake 1940: 404: The Mandaeans, an Aramaic-speaking Gnostic sect of lower Babylonia (Iraq), have developeda complete system of independent vowel writing based on thematres lectionis but as the relation of the consonantal forms withthe other Aramaic alphabets is obscure, and we know nothingabout the time or manner of the introduction of the vowel system,it is impossible to speak with certainty about its origin. Blake ten-tatively advanced the hypothesis that the vowel system had beenborrowed from the Syriac writings of the Church of the East in theeighth century c.E., which, as we have seen, is untenable today.

4 The forms of these letters are derived from Aramaic 1,y, w,and h, respectively.35The last two morphemes were borrowed from the Iranianlanguages. The evidence for the use of this character to representthe ez fe comes from the corpus of Mandaic incantation texts; seeMontgomery 1913: 39, and examples on pp. 203-4 and 252. InParthian this letter is also used to indicate other final vowels suchas the feminine ending /a/; see Skjwervo 1996: 516.

60 BASOR 341


9/11


Further evidence of the influence of the Iranianwriting systems on the Mandaic is provided by theuse of the Mandaic to represent the prothetic syl-lable which breaks up initial consonant clusters, inmuch the same manner as Manichaean Middle Per-sian and Parthian (Skjxervo 1996: 531). The ensem-ble of these features shared between the Iranian andMandaic writing systems suggests that the two arerelated; logic dictates that the Mandaeans, who weresubject to the Arsacids and the Sassanians, followedthe latter in this regard. While the possibility that theMandaic writing system developed independently ofthe Iranian ones cannot be entirely discounted, it isnonetheless significant that it agrees with MiddlePersian in all of these particulars, and that all of con-temporary Aramaic scripts to which Mandaic is morecommonly compared, such as Elymaic and Charace-nean, follow the more traditional (i.e., etymological)orthography. 36

36 It must be admitted that the only clear word from theCharacene coin legends is mik king.

CONCLUSION

After comparison with the other Aramaic scriptsof the Parthian Empire, we can only come to the con-clusion that the Parthian chancery script influencedand perhaps even gave rise to new scripts for for-merly unwritten Aramaic languages such as Elymaic,Characenean, and Mandaic. If, as the evidence sug-gests, these three scripts derived from the Parthianchancery script, and their adoption followed the Ar-sacids' gradual abandonment of Hellenism from 53C.E. onward, then the Mandaeans must have adoptedtheir script at some point during the latter half of theperiod of Arsacid rule, and more specifically betweenthe second half of the first century and the end of thesecond century, the terminus ante quem for the com-position of Mandaic texts given by the colophons.While the written literature of the Mandaeans con-tinued to grow during the Sassanian era and even intothe Islamic period, its origins should be sought withinthe Arsacid era.

ACKNOWLEDGMENTS

The ideas presented here developed in tandem with mydissertation research into the Neo-Mandaic dialect ofKhorramshahr, Iran, conducted under the direction of Jo-runn J. Buckley, James E Coakley, Wolfhart P. Heinrichs,John Huehnergard, and Prods Oktor Skjxervo, and whichwould not be possible without the wisdom, resources, andgenerous hospitality of my consultant, Nasser Sobbi. Thispaper was first delivered in a preliminary form to the Har-vard University Semitic Philology Workshop in May of

2004 and benefited much from the ensuing discussion. Iwould particularly like to thank my advisors Jo Ann Hack-ett and Prods Oktor Skjxervo for their many helpful contri-butions on the matters of Aramaic and Iranian epigraphy,and my colleague Yuhan S. D. Vevaina for his keen eyeand unrelenting patience, which carried him through sev-eral drafts of this article, and which stimulated me to ar-ticulate my ideas further.

REFERENCES

Altheim,1969 Aramdiische Inschriften. Pp. 24-35 in Die Ara-

ber in der alten Welt Vol. 5, Part 2, by E Al-theim and R. Stiehl. Berlin: de Gruyter.

Bivar, A. D. H.1967 A Parthian Amulet. Bulletin of the School of

Oriental and African Studies 30: 512-25.Bivar, A. D. H., and Shaked, S.

1964 The Inscriptions at Shimb5r. Bulletin o theSchool of Oriental and African Studies 27:265-90.

Blake, E R.1940 The Development of Symbols for the Vowels inthe Alphabets Derived from the Phoenician.Journal o the American OrientalSociety 60:391-413.

Buckley, J. J.2002 The Mandaeans: Ancient Texts and Modern

People. New York: Oxford University.Caquot, A.

1972 Un phylact6re mandden en plomb. Semitica 22:67-87.

2006 61


10/11

CHARLES G. HABERL

Colpe, C.1964 Die Thomaspsalmen als chronologischer Fix-

punkt in der Geschichte der orientalischenGnosis. Jahrbuch lir Antike und Christentum7: 77-93.

Coxon, R W.1970 Script Analysis and Mandaean Origins. Journalof Semitic Studies 15: 16-30.

Drower, E S.1953 The Haran Gawaita and the Baptism of Hibil

Ziwa. Studi e Testi 176. Vatican City: Biblio-teca Apostolica Vaticana.

2002 The Mandaeans of Iraq and Iran. Reprint of2nd ed. Piscataway, NJ: Gorgias.

Drower, E. S., and Macuch, R.1963 A Mandaic Dictionary.Oxford: Clarendon.Greenfield, J. C., and Naveh, J.

1985 A Mandaic Lead Amulet with Four Incanta-tions. Eretz-Israel 18 Nahman Avigad vol-ume): 97-107 Hebrew), 69 English summary).

Gu6lpin, J. P.1965 A Contribution to the Location of Ta Azara, theChief Sanctuary of Elymais. Persica2: 19-26.

Hansman, J. F.1998 Elymais. Pp. 373-76 in Encyclopaedia Iran-

ica, Vol. 8, ed. E Yarshater. Costa Mesa, CA:Mazda.Healey, J. F.2000 The Early History of the Syriac Script: A

Reassessment. Journal of Semitic Studies 45:55-67.Hunter, E C. D.

1994 Two Mandaic Incantation Bowls from Nippur.BaghdaderMitteilungen 25: 605-18.Lidzbarski, M.1909 Ein manddiisches Amulett. Pp. 349-73 in Flori-

legium ou recueilde travaux d iruditiondidies i Monsieur le Marquis Melchior de Vogiij iil occasiondu quatre-vingtikme anniversairedesa naissance,ed. G Maspero. Paris: Imprimerienationale.

Macuch, R.1965 Handbook of Classical and Modern Mandaic.Berlin: de Gruyter.1967 Altmanddiische Bleirollen Erster Teil). Pp. 91-203 in Die Araber in der alten Welt Vol. 4, byF Altheim and R. Stiehl. Berlin: de Gruyter.1968 Altmanddiische Bleirollen Fortsetzung). Pp.34-72 in Die Araber in der alten Welt Vol. 5,Part 1 by F Altheim and R. Stiehl. Berlin: deGruyter.

1971 The Origins of the Mandaeans and Their Script.Journal of Semitic Studies 16: 174-92.

Montgomery, J. A.1913 Aramaic IncantationTexts from Nippur. Publi-cations of the Babylonian Section 3. Philadel-phia: University Museum.

Mfller-Kessler, C.1999 Interrelations between Mandaic Lead Scrollsand Incantation Bowls. Pp. 197-210 in Meso-

potamianMagic:Textual Historical,and Inter-pretative Perspectives, ed. T Abusch and K.van der Toorn. Groningen: Styx.Naveh, J.

1970 The Origin of the Mandaic Script. Bulletin ofthe American Schools of Oriental Research198: 32-37.

1972 The North-Mesopotamian Aramaic Script-Typein the Late Parthian Period. Israel OrientalStudies 2: 293-304.

1975 Another Mandaic Lead Roll. Israel OrientalStudies 5: 47-53.

1997 Early History of the Alphabet. Reprint of 2ndrev. ed. Jerusalem: Magnes.Nbldeke, T.

1898 Review of Inscriptionsmandaites des coupesde Khouabir,by H. Pognon. Wiener Zeitschriftfiur die Kunde des Morgenlands 12: 141-47;353-61.

Rostovtzeff, M.1936 The Sarmathae and Parthians. Pp. 91-130 in

The CambridgeAncient History, Vol. 11: TheImperial Peace, A.D. 70-192, eds. S. A. Cook,R E. Adcock, and M. P. Charlesworth. Cam-bridge: University Press.

Sd.ve-Sbderbergh, T1949 Studies in the Coptic ManichaeanPsalm-book,

Prosody and Mandaean Parallels.Arbeten ut-givna med underst6d av Vilhelm Ekmans uni-versitetsfond, Uppsala, 55. Uppsala: Almqvistand Wiksells.

Segal, J. B., and Hunter, E. C. D.2000 Catalogueof the Aramaic and MandaicIncan-tation Bowls in the British Museum. London:British Museum.

Sellwood, D.1983 Minor States in Southern Iran. Pp. 299-321 in

The CambridgeHistory of Iran, Vol. 3, Part 1:The Seleucid, Parthianand Sasanian Periods,ed. E Yarshater. Cambridge: University Press.

Skjwrv0, P. O.1996 Aramaic Scripts for Iranian Languages. Pp .

515-35 in The World s Writing Systems, eds.P. T Daniels and W. Bright. New York: OxfordUniversity.Yamauchi, E M.

1970 Gnostic Ethics and Mandaean Origins. Har-vard Theological Studies 24. Cambridge, MA:Harvard University.

Yardeni, A.1997 The Book of Hebrew Script: History,Palaeog-

raphy, Script Styles, Calligraphy Design.Jerusalem: Carta.

62 BASOR 341


11/11

COPYRIGHT INFORMATION

TITLE: Iranian Scripts for Aramaic Languages: The Origin of

the Mandaic Script

SOURCE: Bull Am Sch Orient Res no341 F 2006

WN: 0603204104006

The magazine publisher is the copyright holder of this article and it

is reproduced with permission. Further reproduction of this article in

violation of the copyright is prohibited. To contact the publisher:

http://www.asor.org/

Copyright 1982-2006 The H.W. Wilson Company. All rights reserved.

Origin of the Mandaic Script

Documents

Transcript of Origin of the Mandaic Script