You are on page 1of 15

A Newly Discovered Hebrew Version

of the Apocryphal "Prayer of Manasseh"


Reimund Leicht

The manuscripts found in the Cairo Geniza and now spread over numer-
ous libraries all over the world are still one of the main sources for
unexpected discoveries in all fields of ancient and medieval Judaism.
One of the newest findings is a Hebrew version of the apocryphal
"Prayer of Manasseh" (Oratio Mariasse, OrM) in a manuscript consist-
ing of three folded quires of parchment (Cambridge UL, T.-S. K 1.144,
T.-S. K 21.95P and T.-S. K 21.95.T), which probably had been part of a
larger codex. In the forthcoming second volume of P. Schäfer and Sh.
Shaked, Magische Texte aus der Kairoer Geniza,1 the whole fragment
will be edited and translated into German, but since this text deserves
more comment than is possible in an edition focussed on magic, it seems
to be in place to give a short presentation of some major aspects of this
new discovery.
The folios preserved (fol. la-6b) contain six prayers, two of them
fragmentary. The title of the first prayer unfortunately is lost, but it
becomes clear that it is an invocation of angels on behalf of a sick person
(fol. la/l-lb/9). The second and third prayers, attributed to Abraham
(fol. lb/10-18) and Jacob (fol.2a/l-17), remind us strongly of both,
well-known formulations from the Jewish liturgies and from Hekhalot-
literature. After the "Prayer of Manasseh, King of Judah, at the time
when he did repentance" (fol. 2a/18-3a/2), there follow two magico-mys-
tical prayers attributed to Elijah, a Seven Benedictions Prayer (fol. 3a/3-
6a/ 19)2 and the first two benedictions of a magical Eighteen Benedic-
tions Prayer (fol. 6b/l- 19). There does not exist any evident material

1 For a more detailed description of the whole manuscript cf. this edition; the first
volume of magical texts from the Cairo Geniza with a commented German translation
has recently been published: P. Schäfer and Sh. Shaked, Magische Texte aus der Kairoer
Geniza L Tübingen 1994.
2 This text was edited by P. Schäfer, Geniza-Fragmente zur Hekhalot-Literatur, Tü-
bingen 1984, pp. 140-151.

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
360 Reimund Leicht JSQ 3

or liturgical order in the collection of those six clearly heterogeneous


texts. But since most of them reveal a mystico-magical character, at first
sight OrM seems somehow odd in this context because it is a purely
pious and penitential prayer. This impression might also be due to the
fact that only a small portion of the original corpus has survived, but in
any case it seems difficult to learn much about those prayers, including
OrM, from the composition of the preserved part of the manuscript.
The discovery of a Hebrew text of OrM could raise hopes that this
could at last be the lost Hebrew original of this apocryphal prayer,
known from earlier sources only in Greek and Syriac versions and other
early translations. But as a closer analysis of potential sources, techni-
ques in translation and the language reveals, our text is a translation.
Not only the dependency of the Hebrew text on versions which are
attested only by relatively late manuscripts, but also internal textual
and linguistic features prove with little doubt that the Hebrew text can-
not be the original version of OrM.3 Although our text is therefore not
much helpful for the discussion about the original language and other
disputed textual problems of OrM,4 it is of major interest for the re-
search on the development of Judaism in the face of Christianity in
late Antiquity and the early Middle Ages.
The main purpose of the present article is a presentation and analysis
of the Hebrew version of OrM. Therefore the questions dealt with are,
firstly, the hypothetical textual sources and, secondly, the translator's
techniques and language. In an appendix to this article the reader will
find the Hebrew text and a translation with short notes dealing with
some additional textual problems.

3 A Hebrew text entitled "Prayer of Manasseh" is unknown in Rabbinic literature.


This of course does not necessarily mean that texts with this title did not exist in
Judaism in late Antiquity since a fragment of a "Prayer of Manasse" has been discov-
ered at Qumran (4Q381 33,8ff.). It is however totally independent of OrM; cf. E. M.
Schuller, Non-Canonical Psalms from Qumran. A Pseudepigraphic Collection, Atlanta
1986, pp. 146-162, and W.M. Schniedewind, "A Qumran Fragment of the Ancient
'Prayer of Manasseh'?", in ZAW 108 (1996), pp. 105-107, for a further discussion on
this text.
4 For some of the more recent surveys cf. A.-M. Denis, Introduction aux Pseudépi-
graphes Grecs d'Ancien Testament, Leiden 1970, pp. 177-181; L. Rost, Einleitung in die
alttestamentlichen Apokryphen und Pseudepigraphen, Heidelberg 1971, pp. 69-70; E. Os-
wald, "Das Gebet Manasses", in Jüdische Schriften aus hellenistisch-römischer Zeit,
vol. 4,1, Gütersloh 1974, pp. 15-27; L. Vegas Montaner, "Oración de Manases", in
A. Diez Macho (ed.), Apócrifos del Antiguo Testamento, vol. 3, Madrid 1982,
pp. 101-117; J. H. Charlesworth, "Prayer of Manasseh", in id., The Old Testament
Pseudepigraphia, vol. 2, New York 1985, pp. 625-637.

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
( 1 996) "Prayer of Manasseh " 361

1. The textual sources of the Hebrew translator

Apart from a few independent manuscripts containing OrM5 there are


four main strands of transmission, all of them in Christian literature.
The first strand of transmission are the Canticles appended to the
Psalms in many Bible manuscripts and liturgical psalmbooks. Originally
composed for liturgical purposes, this corpus of texts underwent many
changes from its presumably first appearance in the famous Codex Alex-
andrinus6 until the Middle Ages, a process analyzed in detail by H. Schnei-
der.7 Through the influence of the Greek Church, appendices with Canti-
cles are also found in a large number of early translations of the Bible and
the Psalms.8 Finally collections of Canticles containing OrM are attested
by patristic commentaries and glosses, such as Verecundus for Latin
North Africa9 and Hesych for Greek scholarship in Jerusalem.10
5 Two examples are: a Greek manuscript consisting mainly of exorcisms (Vat. cod.
gr. 1538, n. v.) and a Coptic homily by Demetrius on Jesaja 1,16-17 (W.E. Crum [ed.],
Theological Texts from Coptic Papyri, Oxford 1913, p. 56 [only fragmentary], and H.
deVis [ed.], Homélies coptes de la Vaticane, vol. 1, Copenhagen 1922, pp. 183-190).
The Greek OrM text of the Codex Alexandrinus was collated with two other manu-
scripts (minuscule 55 and Codex Turicensis) and compared to the Didascalia, Constitu-
tiones Apostolorum and Vulgata versions by A. Rahlfs (ed.), Septuaginta X, Göttingen
1931, pp. 361-363. In addition to this edition the following Greek versions have been
consulted: W. Till/P. Sanz, Eine Griechisch-Koptische Odenhandschrift (Papyrus Copt. Vin-
dob. K8706), Rome 1939, pp. 90-97 (Monumenta Biblica et Ecclesiastica 5); K. Treu,
"Zwei weitere Berliner Septuagintafragmente", in Münchener Beiträge zur Papy-
rusforschung und antiken Rechtsgeschichte 66, München 1 974, pp. 42 1-426; the manuscript
of a Greek-Latin psalmbook, Staatsbibliothek Berlin, Hamilton 552, fol. 189v-190r.
7 H. Schneider, "Die biblischen Oden im christlichen Altertum", in Biblica 30
(1949), pp. 28-65; "Die biblischen Oden seit dem VI. Jahrhundert", ibid., pp. 239-
272; "Die biblischen Oden in Jerusalem und Konstantinopel", ibid. pp. 433-452; "Die
biblischen Oden im Mittelalter", ibid., pp. 479-500; cf. also the earlier study by J.
Mearns, The Canticles of the Christian Church Eastern and Western in Early and Me-
dieval Times, Cambridge 1914.
8 A few of the most important editions follow Syriac: W. Baars and H. Schneider
(eds.), "Prayer of Manasse", in The Old Testament in Syriac according to the Peshitta
Version, vol. 4,6, Leiden 1972; Latin: J.-P. Migne (ed.), Liturgia Mozarabica (Patrologia
Latina 86), pp. 858-859; J.P. Gilson (ed.), The Mozarabic Psalter, London 1905,
pp. 153-154; Coptic: R. Tuki (ed.), Pi Com nte pi Psalterion nie David (Kitãb zabür
Dãwud), Rome 1744, pp. 445-449 (with an Arabic translation); W. Till/P. Sanz, ibid.;
Slavonic: V. Jagic (ed.), Psalterium Bononiense, Vienna 1907, pp. 731-733. For the
Ethiopie Psalter J. Ludolfs edition (Frankfurt 1701) was not accessible. Instead the
following manuscripts were consulted: Staatsbibliothek Berlin, orient, oct. 3032; orient,
oct. 3739; orient, quart. 749. The text of an Armenian version, in manuscripts generally
appended to the Psalms, can be found in H. Zohrapian's Bible edition (Venice 1805) in
accordance with the oost-Tridentinian Vuleata in an aooendix. d. 25.
9 Cf. R. Demeulenaere (ed.), Verecundi Iuncensis commentarii super cantica, Turn-
holt 1976, pp. 148-161.
Cf. V. Jagic, ibid., and id., òupplementum Psaltern Bononiensis, Vienna 1917,
pp. 314f; for the attribution to Hesych cf. H. Schneider, ibid., pp. 58-65.

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
362 Reimund Leicht JSQ 3

Some obscurity still surrounds the second strand of transmission,


which is of minor importance. Missing in older Latin Bible manuscripts,
OrM is appended to II Chronicles only from the medieval period on-
ward. The origin of this custom still needs to be clarified.11
Of much greater importance is the third strand, the Didascalia Apos-
tolorum. This text, originally written in Greek, has survived only in a
Syriac12 and a few other early translations.13 But already at an early time
the Greek text, now lost, had been revised and integrated into a work
called Constitutiones Apostolorum, which contains OrM as well.14
The fourth strand consists of various Horologia (prayerbooks for the
daily office) of the Orthodox churches. While OrM lost its central posi-
tion in the morning service in Constantinople, it has continued to be
used in the Great Compline (Apodeipnon mega) up to the present.15
This liturgical position of OrM has been adopted by Melchite Syriac
Christians, so that some of the extant Syriac versions of OrM are now
found in translations of the Greek Melchite Horologion.16 The same is
attested for the Ethiopian church.17 In the Armenian Church OrM is
recited in the Vespers.18
The manifold transmission in different literary contexts, numerous
translations and manuscripts makes a comprehensive analysis a task
which would clearly exceed the limits of the present study. As far as
the Hebrew version is concerned, we are fortunate enough to be able
to point out two textual sources which deserve closer attention: one of
the Greek versions and the Syriac translations of OrM. The majority of
the abundant textual variants found in the other versions (a phenomen-

11 Cf. H. Schneider, "Der Vulgata-Text der Oratio Manasse, eine Rezension des
Robertas Steohanus". in Biblische Zeitschrift N. F. 4 H960Ì. dd. 277-282.
12 There have been many editions of the whole text and of OrM. The newest one is
A. Vööbus (ed.), The Didascalia Apostolorum in Syriac, Louvain 1979, pp. 89-91.
13 Cf. e. g. Ethiopie: Th. P. Platt (ed.), The Ethiopie Didascalia, London 1834, pp.49-
51; Latin: E. Tidner (ed.), Didascaliae Apostolorum, Berlin 1963, p. 38 (only the last two
verses of OrM are preserved). I could not consult the Arabic version described by G.
Graf, Geschichte der christlichen arabischen Literatur, vol. 1, Rome 1944, pp. 564-569.
14 The newest edition is by M. Mercer (ed.), Les Constitutions Apostoliques, vol. 1,
Paris 1985, pp. 216-219, but the earlier one by F. X. Funk (ed.), Didascalia et Consti-
tutiones Apostolorum, Paderborn 1905, pp. 84-89, is still of much value.
15 Horologion to meza, Athens 1974, p. 172; cf. also H. Schneider, ibid., p. 259-260.
16 Cf. the edition by W. Baars and H. Schneider, ibid.
17 Cf. A. Raes, "Les Complies dans les Rites orientaux", in Orientalia Christiana
Periódica 17 (1951). pp. 133-145.
18 Cf. F. C. Conybeare, Rituale Armenorum, Oxford 1905, pp. 443-507.

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
(1996) "Prayer of Manasseh " 363

on which is not surprising considering the widespread use of this prayer)


cannot be a matter of discussion here.
The comparison of the Hebrew text with the other versions reveals a
larger textual congruity with only one of the consulted Greek versions,
namely the seventh-century Codex Turicesis (T) used by A. Rahlfs in his
edition of the Septuaginta. T reveals a number of peculiar features which
contradistinguish it from a number of older Greek manuscripts and
from the bulk of later ones: It omits the èrcoupávie ("heavenly") in v. 1
attested by the Codex Alexandrinus and leaves out the additional words
at the end of v. 7 which are missing in the Alexandrinus too, but are
found later in Rahlfs' tenth-century minuscule 55 (Rome, Bibl. Vat.,
Regin. greac. 1). In v. 10 T reads eíç tò 'ii' avaveöaai ttjv KecpaMjv
HOD ("so that I cannot lift my head towards you") instead of eíç tò
avaveöaai ne mcèp àjiapTicov |xoi) ("to raise myself because of my
sins") and replaces <xcí|aaç ß5eMynaTa Kai rcÀ,r|$6vaç rcpoaoÇiríanaTa
("putting up idols and multiplying detestable things") at the end of the
verse with |if| rcovnaaç tò ûkXi'[ià aou Kai 'xr' (piAáÇaç Ta npoaxày'iaxà
aoi) ("not carrying out your will and not preserving your laws"). In all
these cases the Hebrew version agrees with T, while all other Greek
versions, psalmbooks and Constitutiones-texts alike, differ from T in
one way or the other. Regarding the question whether the Hebrew text
in fact depends on a Greek version like T, it has to be noted that there is
no translation known to me which could be identified with T as its
textual source.
In two cases the Hebrew text is clearly different from T. In v. 7 there is
no Hebrew equivalent for ovj/icttoç ("most high") and at the beginning
of v. 9 the predicate of the sentence is not like in the Greek %apTOV ("I
have sinned") but im ("[my sins] have become more"). Both variants are
found in the Syriac Didascalia. This could lead us to argue in favour of a
textual source like the text of the Syriac Didascalia, but on the other
hand no Didascalia manuscript has the text of the Hebrew translation
in v. 7 and v. 10. Furthermore, there are many textual changes in the
Syriac Didascalia versions which do not occur in the Hebrew text.
The arguments in favour of a textual source similar to T or to the text
of the Syriac Didascalia are equally strong. Therefore none of the known
versions alone can be considered as the textual source of the Hebrew
translation. The investigation of the language, of the assumed textual
source yields a similar result.
Comparing the Hebrew text to the other versions, we find quite a few
cases where our writer has rendered the text rather freely. This makes it
difficult to point out linguistic patterns which might be influenced by

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
364 Reimund Leicht JSQ 3

another language. One formulation, however, is clearly influenced by the


Greek. This is the rendering of akoöuai Seóuevóç aoo in v. 13. The
Hebrew translator interprets this phrase in a most peculiar way as "|mK
-|t?p3K ^KW -p-iXK ("I ask you, you who I need") which can only be a
translation from the Greek where the verb Séouai has the meaning of
both, "to ask" and "to need". This interpretation is found in no other
version; all translate Séouai as a synonym of aîxoûuai ("to ask").
Since this is the only example illustrating Greek influence it is inter-
esting to see that cases of linguistic features close to the Syriac are much
more frequent.19
The most striking fact is the great number of etymological congruities
between the Hebrew and the Syriac versions. A short list might show the
most obvious examples:20
Hebrew = Syriac
(1) D^TSn DSnn = «m->.-' ' ^<'<a*i ' (plural!, SyrHor)
(8) **? nawn nnöW = >A «ftvciaLafc >ou» (SyrDid)
(10) "»JK «11ÖD1 = «Utf s.^cv (SyrHor)
(13) "naora 'naiui *C? = ><n',vi-> *-).uafc «l' (SyrHor)
(13) finn oinn nnn ^n^nn k^i = «iai^.-' a&^uki* > '-i»>^ «ü
(SyrHor, similar in SyrDid)
(7,8,13) Kin nnK = <'a> fcj* (SyrHor and SyrDid 7,13).
The plural adjective in v. 1 D'pHXH DJTlî, although perfectly logical,
contrasts with the singular in the Greek but is congruent with SyrHor.
In v. 8, •**? nawn nnötm and the Syriac _A i^koa^k ioi» are clearly the
same. The meaning of finn Dinn nnn *J3"nn ^X in v. 13 is not really
clear in Hebrew but this formulation can be explained on the basis of the
Syriac. Further influence of the Syriac versions (SyrHor) may be seen in
the use of the 2. pers. perf. in the first verses instead of the participle
found in the Greek text and possible in Hebrew too, and perhaps in the
plural D*ntP ni*ax for Syriac «-**** -*** *■». although this can be due
to the Hebrew language itself. The peculiar structure Kin nnK, found in
the Syriac version as a» *'-Jtf, is very striking as well.
19 My observations are based upon the Didascalia and the Horologion versions, as
edited by W. Baars and H. Schneider, ibid. I will use those two texts without paying
much attention to their respective textual problems and the relation between them. This
is justified because both versions are clearly dependent on each other and that the ear-
liest manuscript of the Didascalia is not older than the ninth century. The Horologion-
versions edited by Baars and Schneider might therefore preserve older readings. An-
other Syriac version, represented by the "Christian Palestinian Syriac Horologion (Ber-
lin MS. Or. Oct. 1019)", edited by Matthew Black, Cambridge 1954, reveals only very
few linguistic congruences with the Hebrew text and will not be considered here.
20 In the following, SyrDid stands for the Didascalia and SyrHor for the Horologion
version.

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
(1996) "Prayer of Manasseh" 365

The conclusions to be drawn from this evidence are not univocal.


Because of the similarities to the Syriac versions are so blatant with
respect to the choice of words, there can be little doubt that the trans-
lator did have a Syriac text in front of him, but we do not know what it
looked like, since the strong textual congruences with T (as opposed to
the Syriac versions) cannot be interpreted as merely coincidental and
remain to be explained. In fact, two hypotheses are possible to settle
this problem: a) The Hebrew translator might have used a Syriac text,
now lost, which was similar to T in v. 7 and v. 10 but close to our Syriac
Didascalia version in v. 9. This text also must have had the rendering of
ôeóu£voç as "to need", b) The Hebrew translator may have used both, a
Greek and a Syriac source. As these textual sources were probably not
identical, the Hebrew translation of OrM would be a combination of
two different versions.
It might seem to be a bit daring to take one single Graecism in v. 13
and the absence of an adequate Syriac version as sufficient indicators for
a dependency of the translation on a Greek textual source. On the other
hand, for the first hypothesis one has to suppose a whole Syriac version
of which we do not possess the slightest evidence. Although the known
manuscripts do not permit any definite answer (in fact, only the first
hypothesis could ever be proved to be true if a new manuscript turns
up), I am inclined to assume that the Hebrew translator did use a Greek
and a separate Syriac source rather than a highly hypothetical single
Syriac version.

2. The translator 's techniques and the language of the translation

It is a striking aspect of the Hebrew text that it contains a great number


of quotations and allusions to biblical verses. This is of course provoked
by the biblical style of OrM itself which has reminded also modern
scholars of certain verses in biblical texts.21 In many cases the translator
adapts the verses he borrows from the Hebrew Bible rather than render-
ing his textual sources very exactly, although the general meaning of the
original is generally conveyed rather confidently. This is the case in v. 5
where "pì/n ima *>lth TIB5P *a "pai22 reminds one of Neh 1,6 1ÖS7T ^S1?
Tl a SP "»a - "who can stand before his wrath", but the translation is not
literal. V. 7 is very similar to Joel 2,13, but it is remarkable that the
translator did not quote the verse literally although the second half of
21 Cf. e. g. J. H. Charlesworth, ibid., p. 628-633.
Cf. also the note on this verse.

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
366 Reimund Leicht JSQ 3

the verse in the Septuagint is practically identical with the Greek OrM.
The change from avftpdmcov ("human beings") to ùVTi ("the people") at
the end of the verse is probably influenced by Ex 32,12: nsnn ^V Drum
löy1? ("and regret the evil [you will bring] upon your people"). In v. 9 he
identifies the odd expression ''fá'i'io') da^áaar|ç ("sand of the sea")
with the common biblical expression D"»n Dot? bw Vin, replacing the
more biblical "ltPK by ~tP (in the Bible only in Jud 7,12). In a similar
way he does not bother very much about literality in v. 1 1 when he
renders the somewhat obscure kAívcd yóvu KapSíaç jioi)23 ("I incline
the knee of my heart") only with *aV TPW found in Ps 119,112. In
v. 12 the quotation of Ps 51,5 turns out to be a little longer than neces-
sary ("Pon mi TiXüm ...) and in v. 14 Ps 85,8 seemed to fit so well that
the translator not only used it for the rendering of the first part of the
verse but also for the words aœaeiç |ie ("save me") in the second part
and omitted the last five words completely.
Two additional examples may be at first sight a bit more complicated.
In V.6 n*? nv^ TnipTin is a quotation from Ps 36,11 which leads us
pretty far away from the original eXeoq Tx'q enayyeXiaq aou ("the mercy
of the promise"). Was enayyeXia ("promise") a word too Christian for
him or was he simply unable to find a more appropriate Hebrew equiva-
lent?
The additional quotation of Ps 40,13 in v.10 (*33ty ... 'XU'ttm) seems
to have no explanation in the original text, but it is not difficult to
imagine why one could associate this verse with the context of v. 9-10
in OrM. Both texts express repentance in a very similar way and our text
sounds indeed like a reworded version of this Psalm verse. Most ele-
ments such as the multiplicity of sins, the inability to look up and the
long-lasting evil are found in both verses so that the additional words in
v.10 are much less surprising. The translator evidently felt free to add
words he deemed appropriate.
If one singles out all allusions to biblical verses, the whole text looks
very much like a patchwork of quotations rather than a literal transla-
tion. But the Bible has probably not been the only source for his choice
of words and expressions. The whole vocabulary of the text is very close
to (penitential) prayers found in prayerbooks up to present times.
Should not e.g. nVrn1? HKJ "f?1 in v. 15 remind us of the words *f? *D
... HKJ in the Yishtebah of the morning prayer?
In spite of the translator's numerous extremely free renderings of the
text there is only one example of a clearly interpretative translation: In

23 The word noi) only in T.

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
(1996) "Prayer of Manasseh " 367

the Greek and Syriac texts of v. 13, Manasseh asks God not to let his
anger last forever. This was apparently not enough for the Hebrew trans-
lator. According to his interpretation Manasseh had to ask for forgive-
ness not only in this world but also in the world to come. Thereby our
translator reveals a knowledge of the Rabbinic discussions about Man-
asseh and his repentence. In mSan 10,2 the anonymous Rabbis declare
that Manasseh is one of the three Kings of Judah who have no part in
the world to come. Only Rabbi Yehudah is opposed to this view. The
following Gemara in the Talmud Bavli (bSan 102b-103a) explains and
specifies each of the opposing positions, but the Talmud Yerushalmi
(ySan 10,2; 28c) illustrates its approval of Rabbi Yehudah's opinion
with an Aggadic episode.24
The Hebrew translator of OrM evidently was aware of the fact that
the question whether or not Manasseh's repentance would be accepted
in the world to come, was a highly disputed topic. Therefore, in his
rendering, Manasseh, conscious of this Rabbinic dispute about him,
anticipates arguments against his repentance and asks for forgiveness
in this world and in the world to come.25
Although we do have the impression that the translator worked hard
in order to render his Greek and Syriac sources into good (biblical)
Hebrew, he was not always successful. That his skills in the Hebrew
language were somewhat limited becomes clear not only from the Syr-
ianisms mentioned above. In v. 10 the translator evidently intends to say:
"I myself am responsible for your wrath continuing against me", but the
expression he employs (IN TINn1?) actually means (at least in biblical
Hebrew) quite the opposite ("to delay [the outburst] of wrath"). If
D^riö in v. 15 is no scribal mistake this cannot be called a correct trans-
lation of the sources.26 Furthermore his distinction between 8*7 and *?K
in combination with the imperfect is not totally coherent (v. 7 and v. 13)
and the use of suffix pronouns is sometimes rather clumsy (v. 15, "|*7
iVrn1? nNJ). On the other hand it is a remarkable feature of the transla-
tor's Hebrew that he employs the imperfect and not the participle for the
present tense even where the Greek and Syriac versions use the latter
(v.7 Dnrn; v. 13 itnnmw).

24 Cf. for further Aggadic material on Manasseh L. Ginzberg, The Legends of the
Jews, vol. 4, pp. 277-281, and notes in vol. 6, pp. 360-376.
It should be noted that this interpretative translation is a fine example of Rabbin-
ic exegesis. There can be little doubt that the translator bases his exegesis on a close
reading of ur|ôè eîç xòv alcova unviaac xripfjaflc xà Kaicá uoi ("not in eternity may
[continue] your anger and do not retain the evil for me"). Why unvíaaç? In this world.
Why XT|pf|aiiç? In the world to come!
Cf. the note to the translation.

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
368 Reimund Leicht JSQ 3

The vocabulary of the translation is, where there was no biblical quo-
tation at hand, very close to Rabbinic Hebrew, mn ^ "PK (v. 9), ^ T»K
nxö (v. 10), TOT 'ant (v. 14), HKJ i? (v. 15) and the use of MWn in the
meaning of "repentance" (title, v. 8) are just the most prominent exam-
ples for such anachronisms. Nevertheless our translator did not lack a
certain feeling for the beauty of language as e.g. in those cases where the
Greek text repeats one verb and he uses two different roots (v. 9,12,13), a
stylistic feature typical for biblical and prayer texts.

Conclusions

It is impossible to discuss every single problem arising from the Hebrew


text of OrM, but some major aspects could be pointed out. The newly
discovered text is a translation from Christian sources. This is an inter-
esting fact in itself. Did the Hebrew translator know and use Greek
psalmbooks? Did he know and read the Syriac translation of the Didas-
calia or Horologia? Our text is of course not the first evidence of Syriac
liturgical texts from the Geniza,27 but here we have an example of a
direct Jewish (re-)adaption of a text from Christian sources, a process
which perhaps deserves more attention than generally attributed to it.
If we think about when and where this translation could have been
composed, we have little clear evidence. Egypt itself is one of the
possible candidates, but it is equally possible that it was done in
Syro-Palestine where Judaism co-existeded closely with Syriac and
Greek Christianity. Not much more can be said about the date of com-
position. It is rather evident that the version of T (seventh century) is
not the earliest stage in the evolution of the Greek OrM, but this does
not help us very much, even if we knew for certain that our translation
is dependent on this version. Nobody is able to say when the Syriac
versions were produced; the earliest known manuscripts stem from the
ninth century. This means that the only fixed date is the one provided
by the Hebrew manuscript itself. The tenth century serves as a vague
terminus ante quern.
Little more can be said about the background of the translation. As it
becomes clear from v. 13, the translator probably had some knowledge
of Rabbinic Judaism, but neither the text itself nor the character of the
manuscript tells us much about the actual use of the Hebrew OrM and

27 S. Brock, "East Syrian liturgical fragments from the Cairo Genizah", Oriens
Christianus 68 (1984), pp. 58-79; id., "Some Further East Syrian Liturgical Fragments
from the Cairo Genizah", ibid. 74 (1990), pp. 44-61.

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
(1996) "Prayer of Manasseh " 369

whether or not it was known to a wider public. The content and the style
of our text are extremely similar to other Jewish penitential prayers, so
that it could have been easily inserted into the Tahanunim for Monday
and Thursday, for Yom Kippur or any other feast, but there is no sign
that it really had such a function.
The main importance of the Hebrew version of OrM does not lie in
its value for textual criticism. For this purpose we possess other and
better versions. But it does pose new questions concerning the cultural
exchange between Judaism and Christianity in the period after the his-
torical split between Synagoge and Church.

Appendix: Text and Translation

The Hebrew text is identical with the edition in P. Schäfer/Sh. Shaked,


Magische Texte aus der Kairoer Geniza II (in print). The numbers on the
right side indicate the line on the manuscript, the numbers in paren-
theses on the left and the slash within the text indicate the respective
verses according to the Septuagint edition. Question marks stand for
illegible letters, a point on the top of them marks those letters of which
the reading is dubious. Suggested restorations for the lacunae (indicated
by brackets) are to be found in the footnotes to the translation.
In the translation three question marks stand for an untranslatable
word, words in parentheses indicate my additions, those in brackets
suggested restorations for the lacunae. The numbers in parentheses in-
dicate the verses.

Text

Fol. 2a
.miwn nwyw nyn mirr i^a nvaa nV»Dn 18
(1) ay-in aipsm pnsr omaa "nina •»jfnt la^iya ü^wn "•» I 19
(2) D*sñ? *7D hv y^xti nòS D^awn na nmtzwtp I D'pnam 20

Fol. 2b
(3) [ ]iirm nna?[ ]? trn na nmx inmm I 1
(4/5) t»ki I .Toh[ ]7rñs vani tf?iyn ^d I -pmtm 2
*75/ ias/n ima ^ nay* *a vki itim ni*n*? ^no1? Vo1» "»a 3
I a1? ntip^ Tnipixi TTon 100a vòi*) ipn K^ai I D^Konn 4
(6/7) nnx

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
370 Reimund Leicht JSQ 3

*?y onrm -pam D'a-n Tfon Deliro nay lonnt? Kin 5


mo ^y onrn íòi D^pnxn ti^k *« am nna I .oyn nyn 6
(8) D^pHsn
KBinn ^ ?âwn nnatm na won irtw aipim pnx* orna« 7
(9) •»ifra "mawa ?9?3i dvi notrr *?ytp Vina ia*i "maont? ^dö I 8
nwi ^ t»x 'f? ^n^yt^ '»•»niKün miai 'n^K ^*» '»'»maiy i^^i 9
^ t»ki •»'»niKon naai ^na ^aaa ^k ^loai I D*awa o^an1? 10
(10) nxa
Tim1? •»axy1? "»nana t^*ö "»rnay^ ^sn t»Vk '»wx-i onn1? 11
TPin Tnatrr k^i ^a^s-l "»n^y kVi t»3ö^ TiKom ^^y idk 12
•»iwn nnyt^a "»axy mxi^ "»n^ia*» k^t ^naiy ^irwm 13
ion naa tr^pa1? n^D1? sa^ •»rroj p ^y I •»aaîy "»aVi 14
(11/12) "»nxon I D^arm
(13) it^paK I i^an ma 'nKom yiK ^x '»"»y^s •»a ^Thx ™ "»n^yi 15
"maona 'naan ^ki ^y noim ■»rftK ^ian nmx ins ^Ktr^ 16
«an tí?w*7 "mKon ^aoa irán x1?! o^iya ^y idk n-in* ^i 17
dix '»aa1? D^an "»" Kin nnxtr^ pxn mnn nnn •»aa^nn ^xi 18
(14) ^ inn nyt^^ lion ™ ^xin I .ü^y-in on^yaa lüinn^ 19

Fol. 3a
1? •»a ^n •»a1» ^a nnat^xi I "»aytmn1? •»«aí ^•»x^ "»s ^y i« l
(15) D^ana
lax D^aViy ^a^iy1?! D^iy^ n^n1? niu ^i D^aur mxax ^a 2

Prayer of Manasseh, King of Judah, at the time when he did repen-


tance.28
(1) Lord, who rules in his world,
God of my fathers Abraham, Isaac and Jacob
and their righteous offspring,
(2) You, who made heaven and earth
with all W.29
(3) And with your words

28 The variety of titles attached to this prayer in the different versions is enormous,
ranging from the brief "Prayer of Manasse" to elaborate introductions on the historical
circumstances when Manasseh pronounced his prayer (especially in the Syriac ver-
sions). The remarkable differences in these introductions cannot be discussed here.
None of the different versions is especially close to the Hebrew text.
According to the Greek and Syriac text one would expect here the biblical ex-
pression DlttX *?3 ("all their host"; cf. Gen 2,1), but this reading is totally impossible.

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
(1996) "Prayer of Manasseh" 371

You have commanded the sea [ ] ?7?.30


And (4) the whole world
fears [you]31 and praises you.32
And [ ] is always afraid
[ ]your[ ].33
(5) And there is none
who bears to see your glory,
and none who endures before your power
and your anger against the sinners.
(6) And unexplorable and innumerable
are your merciful and righteous acts
for those who have an upright heart?4
(7) You are the one
with whom there is mercy,
and your mercy is great,
and plentiful is your compassion.
And you regret the evil
(which you bring) upon the people.35

Since the resh is dubious, D'Slv] ("their beauty") for the Greek KÓauxp aôicóv ("their
decoration") could count for a possible, though not very probable, reading.
The last word probably can be read as nn[ann] ("you have sealed"), so that the
lost text of the lacuna could have been something like: nnann lawa Dinnn TUO ("and
with your name you have sealed the depth").
31 Hebrew text: hllKT.
32 The two verbs "pmtm hJlICV are most probably the translation of the two
adjectives (poßepq) ("fearful") and èvôóÇcp ("glorified") in v. 3. It is difficult to say
how the change to a verbal construction happened because there is no reason why
the original could not have been translated into naiwani mían. In any case it seems
much easier to imagine such a change occurring on the basis of the Syriac text. In
Syriac the two adjectives follow the noun, which at least could explain how those
two words could have been combined with the following verse.
33 The Greek text suggests a translation like irVD '3Sa DHms ("it [i. e. the whole
world] is afraid before your power"), but this is impossible with respect to ("pn[ ]) m
fol. 2b/2. It is interesting to see that "|ITD "»3D1? is found in the following line, where it
does not fit the Greek text. Perhaps the translator or a copist has confused the text of
the two lines fol. 2b/2-3. In fol. 2b/2 it is possible to reconstruct T^Iön]. This would
fit the following verse v. 5 which, according to the Greek, should be TI8SP "pNl
inttîi inori "»3Ö1? ("none stands before vour wrath and aneer"V
34 Ps 36,11.
35 The Hebrew translation gives the text of v. 7-8 a new structure and interpretation.
In the Greek and Syriac, v. 7 is separate and describes God as one who regrets the evil
he might bring upon humanity. V. 8 introduces a new idea: God did not give, i. e. grant,
repentance to those who did not need it (Abraham, Isaac and Jacob), but to the sin-
ners. The Hebrew text is different. Here v. 7 and 8 form a logical unity: God regrets the
evil he brings upon men (v. 7), but does not regret the good (additional 31B in the
Hebrew text) he promised to the forefathers. This is the reason why he grants repen-

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
372 Reimund Leicht JSQ 3

(8) You are the Lord, God of the righteous.


You do not regret the good (you gave to) the righteous,
Abraham, Isaac and Jacob,
who did not sin against you,
and you gave repentance36 to me, the sinner.
(9) Because my sins have become more37
than sand on the seashore,38
my transgressions have become great,
my God,
and my offences have become great,
Lord, my God,
and because of the multitude of my sins,
which I have committed against you,
I am not allowed to look up to heaven.
(10) And I am bent by fetters of iron,
and my sins are heavy.
And I do not possess the insolence
to lift my head towards you.
And because I transgressed your commandments
I myself am responsible
for your wrath continuing against me.
And I have sinned before you,
and I did not carry out your will
and did not preserve your laws.
And my transgressions39 have caught up with me,
and I cannot see,
they have become more than the hair of my head,
and my heart has left me.40
(11) Therefore / inclined my heart41 before you
to ask for mercy and compassion.
(12) I have sinned and transgressed.
Lord, my God,
because I recognize my iniquities

tance to the sinners. It seems that for some reason the Hebrew translator did not
understand his sources and simply reinterpreted them.
30 Hebrew text: htawii.
37 Hebrew text: [)p[l]y
J8Cf. Jud 7,12 etc.
™ Hebrew text: "fi ah»
40 Ps 40,13.
411 Ps 119,112.

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms
(1996) "Prayer of Manasseh" 373

and I am always aware of my sins,42


(13) I ask you,
you whom I need,
take pity, my God, and spare me,
do not destroy me because of my sins.
And your wrath may not be
upon me in (this) world,
and do not bring my sins before me
(in view of) the world to come.
And do not pronounce me guilty
down (in) the depth of the earth,43
because you are God for human beings
who repent their evil doings.
(14) Show me your mercy
and give me your salvation,44
although I do not deserve to be saved.
(15) And I shall praise you
all the days of my life.
Because all hosts of heaven ask you for compassion,45
and you deserve to be glorified
forever and in eternity.
Amen.

42 Ps 51,5.
43 Cf. above on Syriac influence on the Hebrew text.
44 Cf. Ps 58,8; Tion and **? instead of írmn and 131? in the Masoretical text.
Thepfe/ of pn is unusual in Hebrew, the hitpa'el means "to ask for compassion".
The word does not make much sense in this context. According to the other versions
one would expect a word like "to praise". Therefore Diario could be a scribal mistake
for D^J-ia ("they sing your praise").

This content downloaded from 128.111.121.42 on Mon, 20 Jun 2016 05:42:56 UTC
All use subject to http://about.jstor.org/terms

You might also like