Ube '~niversttV of Cbtcago
A COMPARATIVE STUDY
THEODOTIONIC AND SEPTUAGINT
TRANSLATIONS OF DANIEL
A DISSERTATION SUBMITTED TO THE
FACULTY OF THE DIVINITY SCHOOL
IN CANDIDACY FOR THE DEGREE
OF DOCTOR OF PHILOSOPHY
DEPARTMENT OF NEW TESTAMENT AND EARLY
ALLEN PAUL WIKGREN
Private Edition, Distributed by
THE UNIVERSITY OF CHICAGO LIBRARIES
It is the purpose of the present investigation to contri-
bute to the study of the characteristics of translation Greek
which is part of a research project in the Septuagint recently
initiated by the Department of New Testament and Early Christian
Literature of The University of Chicago. More specifically, the
endeavor has been made upon the basis of a comparative study of
the Septuagint and Theodotionic translations of Daniel 1:2 2:4
and 8 to isolate some of the translation phenomena definitely
characteristic of these renderings, to indicate statistically the
departures of the texts from the Hebrew and their differences ij-
Aer S, and to test by a few examples the applicability of the
findings as criteria for the estimation of the quality of other
Greek, in particular that of the Gospels for which Semitic origi-
nals have been postulated.
The book of Daniel presents notable opportunities for such
an investigation because of the two complete, independent, and
differently executed Greek translations extant, and because ap-
proximately half of the Massoretic text is in Aramaic. The latter
fact is of special value in developing criteria for appraising the
quality of the Greek of the New Testament documents, and the rela-
tively literal Theodotion and paraphrastic Septuagint exhibit to-
gether a wide range of translation phenomena. This study is con-
fined primarily to the Hebrew of the sections already indicated,1
but a number of particular phenomena have been examined throughout
the book, and those for which the Gospels have been interrogated
are of a Semitic rather than solely Hebraic character.
1Professor J. M. Rife has already contributed an investi-
gation of the translations of the Hebrew of 1:1-16 and the Aramaic
of 2:4-22 together with a more extensive study of the article,
idioms of dating, and especially word order. See "Some Transla-
tion Phenomena in the Greek Versions of Daniel" (Ph. D. Disserta-
tion, The University of Chicago, 1951).
The investigation was beAun by a thorough comparison of
the translations with the Hebrew and with each other. In 1:1-2:4
all variants were first entered upon cards, the instances when
both translations agreed in literally reproducing the original
text being so few that almost the entire text was thus represented.
Upon the cards were noted the Greek and Hebrew, a literal English
translation of each, variant readings of other MSS, versions, etc.,
as provided in the texts of H. B. Swete1 and Holmes and Parsons2
for the Greek and R. Kittel3 for the Hebrew, and a brief descrip-
tion and attempted explanation of the variants, These were then
classified and the divergences of chapter 8 entered directly un-
der the resulting categories, textual variants only being noted if
significant. Some characteristic literalisms were investigated
throughout Daniel and also sought for in the Gospels and in eight
volumes of Greek papyri of the Ptolemaic period.4 Finally the
data of each category were statistically summarized and evaluated.
Variants due to a different Hebrew text or to mistranslations were
noted but excluded in the evaluation of the comparative literal-
ness of the two Greek versions.
1ghe 01 Testament n Greek According to the Septuagint,
vol. 114 (Cambridge, University Press, 1925).
2yetus Testamentum Graecums j variis lectionibus, vol.
III (Oxford, University Press, 1798).
Biblia Hebraica, Pars II2 (Stuttgart, Priveleg. Wurtt,
4Ulrich Wiloken, Urkunden Ur Ptolemaerzeit. Parts I-III
(Berlin und Leipzig, Verlag von Walter De Gruyter & Co., 1922-27).
B. P. Grenfell and A. S. Hunt The ibeh Papyri, Part I (Oxford,
University Press, 1906); The Amherst Papyri, Part II (Oxford,
University Press, 1901). F. G. Kenyon and H. I. Bell, Greek aU.R-
ji I th British Museu Part I (London, Wm. Clowes & Sons, 1893);
Parts II and III (Oxford, University Press, 1898 and 1907).
METHODS OF TRANSLATION
To provide for an appreciation of the differences in method
exemplified by the Greek translations a number of passages are
given below together with the Hebrew original. Some 45 such pas-
sages were particularly noted in which the LXX paraphrased or
otherwise freely translated the Hebrew and 8 rendered it with a
high degree of literalness. The opposite situation was noted only
5 times (in 2:3,4; 8:5,13,19), one of which in 8:3 is highly to be
suspected.1 These do not include instances of vocabulary in which
6 is occasionally less exact than the LXX.2 Such a preview should
render the subsequent particularized statements and statistics
more intelligible and meaningful, for the latter cannot, as these
examples do, exhibit the syntactical and grammatical relationships
involved in departures from the Hebrew. Some of the more interest-
ing and instructive examples follow.
1:1 'Enl BaartiE 'Ia 'Ev Eits rp(r*y TrF
xsl~ TC 'Iousoaao PaotXEa(ac 'IoKxel
ETouc tp(Tou potitc 'Io68-a
nt Do1 oBni niva
nT1in 10 w0r inS
cola )IN lox c'l
ntD *311El CHOD 13
tlopv Cn nz 1VXt
1:4 vsav(oxouC &pi-
pouc xal esoe8sec
vtav(cKouC o0 oc6x
Eortr airoI & iSoc
xal xaXobu rT 5ys(
xat oTc &otrv IOTxC
ev abTooTc( iivas
1There is weighty textual support in each translation for
the directly opposite reading. The usual symbols, LXX for the
Septuagint, 8 for Theodotion, and MT for Massoretic Text will be
2See pp. 6f.
1:8 xal 1vseugen
AavtXA 9v T)
1:17 xal ToiT vsav(oxotc
ISuxev 6 xuptoc
2:1 ouvrti Sic 6p&paTa
xal lvdsvit ip-
Rneosv Tbv pooaiia
elTov xptbv Eva
8:4 1C fe0sEX
8:27 xol &vaathe 7npay-
xal e19so aavti?-
adl Th xapb(av
l3, fp 'j civ il'1
xal Ta nati6pta anpanx nsn c'Vl'nl
raora; oi riofapEc en)ita aDn ins
aTro1c 6 9ebc
hvuvt&a cr Napou-
xal f~pa robc- 6Geuax-
god( gou xal 16ov,
xal tSob xpibc sIe
XarT ob T60eXa a0toO
xal Avgornv xal
Inofouv T1 Epya toC
The extreme freedom which often characterizes the LXX
translation sometimes makes it difficult to say when a particular
rendering is or is not really a mistranslation. In compiling a
list of these, however, the translators were given the benefit of
any doubt. Yet some 26 indubitable errors were noted (LXX-15,
8-11, coinoide-5), most of which were due to misunderstanding the
meaning of the Hebrew. So, e.g., in 1:5 rpaonfre (8), 1:10 Tpa-
XAX( (LXX), 1:15 icrupol (6), and 8:4 PeTi 68 raTra and 6Cn[a(LXX)
the original text is mistranslated.1 The force of the Hebrew con-
struction is missed by 6 in 1:12 (xal
8:6 (ior~), by the LXX in 8:10, and by both in 8:4 where 6 reads
a future and the LXX an aorist for the Hebrew potential imperfect
ipoy' aS. Five other renderings, 3 by 8 and 2 by the LXX, were
very inadequate though not enough so to be called mistranslations.2
10thers occur in 2:1; 8:2,17,20,21,22,24,25,26. Two or
three of these may be due to corrupt or difficult text.
2E.g. in 8:10 and 18.
noln 131Z331 Cn
nniMt 'I'V KVDl
Top inx "K nntI
It is evident that when such partial departures as occur in para-
phrasing are disregarded the mistranslations are relatively few
in number and unimportant. Again, while 8 is ostensibly engaged
in obviating differences from the Hebrew sense, he has as many as
the LXX in these sections.
In connection with inadequate translation we should mention
transliteration, a favorite device of 8 even at times when there
could have been no doubt about the meaning of the Hebrew.1 Five
instances, one shared with the LXX in 8:13, occur in these chap-
ters as follows: 1:3 popeoipev, 1:11,16 'Aps~Xa8, 8:2 Tb BpEt,2
8:2,3,6 Op6)x, and 8:13 peXiouvel (LXX s)kpouvl). They seem to be
all due to ignorance of the Hebrew meaning.3 Other instances occur
in 11:38 (pawoeTv), 11:41,45 (oapasfv), and 12:6f. (pa88s(v), all
of them by 6.
About two dozen phrases or expressions were recorded, al-
so, in which both translations deserted the Hebrew for idiomatic
renderings, almost half of which were identical. A typical assort-
ment of these is given below together with the original text.4
LXX 8 MT
1:10 rbv ExTcEavT' Same In iw
rva p) I8 pAi ROTSE (n rnXI not irn
auv'rpEqpdvouC 6iTv Th ouv4AXxo "'pjv C3'33 lux
2:1 Wn' auToC Same 51,tp
8:6 rbv rN xipaTa TroO b xpIarpa 0'1ipn tpy
8:7 ilt r~v yv Same 66s3I
8:23 5 pnpoupivvv Same CnnI
1According to J. Reider in Prolegomena _o A Greek-Hebrew
ag Hebrew-Greek Index D9 Auila (Philadelphia, 1916) Theodotion
transliterates more than Aquila.
2The LXX mistranslates this word.
3For discussions of these see R. H. Charles, A Critical
and Exegetigal Commentar 2y ok of panel (Oxford, The
Clarendon Press 1929).
4See also 1:5,8,9,14,15,18.20 8:7 10,1,1,16,20,21.
5Aquila translates literally eI aL 6v
6So in 8:10, but in 8:12 both translate Xapml*
It will be noticed that these do not represent such radical and
significant departures from the original as did the paraphrastic
renderings of the LXX. Theodotion's extravagances amount usually
to no more than the harmless alteration of a word or two. One
favorite expedient seems to be the use of the article as a rela-
tive to translate -ix, hardly to be described as free translation.
The favorite literalistic device of representation, or of
unvarying translation of a Hebrew word or phrase by the same Greek,
though frequent in 8, is far from being universally applied. His
tendency in this direction is indicated, however, in many places.
The verb 1jU, e.g., occurs 5 times in the eighth chapter1 and is
invariably translated by gsyakova, but in the LXX it is thrice
rendered by 6Vp6 and twice by xOKTaX.d. e and the LXX each use
5vep6xno 24 times in Daniel'. 9 so translates UNx (or Zix) 21
times, but the LXX usage is distributed over .3a (7), sti (7),
cx (4), 'x (4), 'n (1), and six 13 (1). 6 translates 15 occur-
rences of 133 by A&vp whereas the LXX renders 14 occurrences 7
times by &vfp and 7 by 6vepuxoc. Yet neither translator availed
himself of the ideal opportunity for representation as well as
more accurate translation afforded in 8:16 by the occurrence of
both 13i and oIK, for 6 renders each by &vip and the LXX each by
8vOpwnoc. Likewise in 8:26 two entirely different words for
"vision", nxDi and 1i1n, are translated by the same word, rb 6poKa
in the LXX and h Zpacri in 9. Examples of this kind could be mul-
tiplied. The great deal of variation that exists in the trans-
lation of the various parts of speech will be presented later.2
In general the vocabulary of the LXX is richer and more
varied than that of 8, and exhibits more idiomatic usages. 9 in
attempting to render a Hebrew word by the exactly corresponding
Greek root is sometimes less accurate than the LXX in reproducing
the actual Hebrew meaning. It seems, too, from the limited obser-
vation afforded by the sections studied that the LXX translator
had a slightly better knowledge of Hebrew than 6. Some examples
2E.g., the idiomatic '3E, translated in a half-dozen
different ways by 8, although lvcntov predominates. See pp,9f.
of differences in vocabulary follow.
LXX 0 MT
1:4 8t6Xextov fxaooav l1
1:9 TtpIv x6ptv Esov oixTrstpPv con'o jIn
1:12,16 T'v boxpfav rv a onspa6v D'piyn
1:18,22 intIroEv slnev oW
8:4 ipbc Buoj&C xaOT e6xaaoav no'
8:6 Ev Oeug iv 6pvi ncn
8:9 bvsqdver EEiX06v K
8:17 8eopupferv ltaepevjv ,rpaV
8:18 hystpt pi x6l Eo'Tna~ v gst 'eTo'
In all but the eighth of these e is translating by the
root meaning of the Hebrew. The LXX does so in none of these
instances, but seeks rather to employ more idiomatic terminology
or to find a meaning more suitable to the context. Yet only one
of the usages of the LXX that of 1:9, is inaccurate, whereas
at least the third and the last three are more accurate than the
renderings of 8.
SOME CHARACTERISTIC TRANSLATION PHENOMENA
One of the most noticeable features of the translations,
especially that of 6, is the extreme frequency of xal and of para-
tactio construction. A worthy example is presented by 6 in 8:27
which reads xal ly& Aav?)X Ixop~ijieTv xal 1Ca)axCfaenv xal &via-rTv
xal lnxoouv r, ESpya Tro PaauX9.c, xal Ja6vacov rTv ipaoLv, xal
oCx fiv 6 auviwv. Other such instances are to be found in both at
1:19; 2:1; 8:4,7,12,16, and 24, and in e at 1:12; 2:2; 8:2f.,17f.,
etc. The LXX, as exemplified in the paraphrases previously cited,
adopts a hypotactic rendering a few times, especially by altering
a finite verbal form to a participle. So, e.g., in 2:2 it reads
xal napayev6psvoLt taoav for Theodotion's xal hXeav xa1 floramv.
Other instances may be seen in 1:1; 8:2,3,18, and 27. It also
employs a genitive absolute 5 times (in 8:1,2,18,23), that of 8:23
being shared by 6, The scarcity of this common Greek usage, of
the familiar pgv 85 construction, and of adversative particles
are characteristics which contribute to the parataxis and profusion
of xaf in both translations. The idiomatic piv 81 occurs through-
out the entire book of Daniel only 8 times in the LXX1 and twice
in 62 in the translated portions. Simple 86 is frequent in the
LXX but occurs only 7 times in 6,3 each time in the Aramaic chap-
ters. 'AXX) and y&p are oonspicious by their absence. It is of
interest to note that in the much less extensive Greek additions
of Susanna and Bel and the Dragon 68 is found 20 and 14 times
It is true that some at least of these phenomena are met
11:7; 2:24, 3, 41, 42; 3:1,25; 12:2.
52:6,15,24,30; 3:15; 4:15; 5:17.
with in the KoIvA Greek but hardly in such proportions. Perhaps
the number would be increased if more non-literary papyri of nar-
rative content existed. Expressions like xal h&v roXpocoart xal
xaTrool1 occur and xif is frequent, but consistent parataxis is
absent. Genitive absolutes, for example, are overwhelmingly fre-
quent in the papyri. In a writer like Arrian xaf is about half
as frequent as in Daniel, and in Josephus the occurrences are even
fewer. In a passage of 212 words taken at random from the latter
only 5 were used, whereas there were 30 in the first 215 words of
the eighth chapter of Daniel by 0, 28 by the LXX.2 Josephus'
average, however, seems to be about twice that of the section in-
dicated. It is apparent that parataxis with its profusion of the
coordinate conjunction and lack of subordinating idioms and adver-
sative particles is an outstanding characteristic of translation
Greek. It is, in fact, a feature of the LXX in general and is
directly proportionate in amount to the literalness of translation.
There is also a noticeable frequency in places e.g. here
in chapter 8, of the personal pronoun, particularly in the geni-
tive to denote possession. Profusion of personal pronouns in the
oblique cases is, however, a characteristic of the non-literary
Kotv- in general3 and is therefore inconclusive alone as evidence
of translation. Such evidence must be limited to a pleonastic
and often unidiomatic ty6 for '3~ or 'nIN as in 8:1 by e, GOpe
npbc pi, ly& bav&tX, x.Tr.. Other instances of this occur in
1:10; 8:5,15,27; 9:2, etc.
In addition to these more general phenomena a number of
specific literalisms were noted which, although in most instances
possible or perfectly legitimate Greek, were because of the origi-
nal Hebrew overly frequent or used in lieu of more idiomatic ex-
pressions. 9 is fond of iv6xtov as a translation of 'vEf and
1U P Z 64, 1. 10.
2In lists 8 and the LXX generally conform to Greek idiom
by adding conjunctions before all items or by omitting that before
the last which Hebrew idiom retains. So in 1:6f.,11,19, etc.
3E.g. in U P Z 79, 11. 15-17, OfSTo avePpwxov Xytv pot
tips b6 6ipipo 1 too6< cou xal ly 4 aofo To &fpi a ToO no586
iou. 'EY6 seems to be used here for emphasis.
related usages of C'e, using it 33 times as compared to 4 by the
LXX in the translated portions. The LXX seems to prefer IvavT(ov
but also uses a dozen other locutions. A literal translator would
be prone to overuse vSnitov because of its correspondence to the
A similar situation exists in the use of nap& and less
frequently Onip with the accusative in expressions of comparison
to translate the idiomatic Semitic usage with ic and Py. The Greek
is of course perfectly permissible. Liddell and Scott cite usages
from Herodotus and Thucydides; Moulton and Milligan and others
give examples from the papyri. But the frequency of the con-
struction in Daniel coupled with the almost total absence of the
common Greek genitive of comparison marks it here as a literalism.
Of 6 to constructions (1:10,15; 2:30; 8:3; 11:2,13) 8 translates
4, the LXX 3, by nap& and the accusative, e 2 and the LXX 1 by
aip and the accusative, and the LXX 1 by the genitive of compari-
son. Two occurrences of the parallel yV usage (1:20; 3:19) are
rendered once each way by the LXX and once by the nap& construction
in 8. 8 has a different text in 3:,19. The following examples
amply illustrate the various usages in both the Greek and Hebrew.
LXX 8 MT
1:15 f ietc tou apaITOC E(oupol Talc o6pEtv Ica 'I'a1i 31
xpeaoowv T@v SXXuv Intp T& noata6pta Cl'n 3 to
1:20 aoqfttpouC Sexa- 8exanxao(ovac naoph 3 py ni' icyp
Rkao(oC unIp tobc 1i&vTxoc ool T'owinn
11:13 ouvoyuhv Ipef(Cova XXov noAbv unep lirtnn Io 1 lion
nap& ThV Xp6tv Tbv 7p6sOpov
It is interesting to note that the LXX uses the nap6 construction
twice (1:13,20) where there is no expression of comparison in the
Hebrew. In the Greek additions, too, it is found twice in the
LXX and once in 8.
But of. Aquila's even more literal np xpoalxov.
Chapters 8 ana following contain a noticeable use of Ev T;
with the infinitive corresponding to the Hebrew construction with
z or, less frequently, 3 and the infinitive. It occurs here 10
times in 81 and 8 in the LXX,2 all but the 3 in 8:8,19 and 10:9
representing the a usage. Of the 3 instances where e does not read
with the LXX 2 are apparently due to text3 and 9:27 is an addition.
In the remaining 4 differences 10:9 and 12:7 are textual and 8:8
is rendered by a os clause and 8:18 by a genitive absolute in the
LXX. The translators clearly found it difficult to avoid the in-
finitive usage when confronted by it in the Hebrew. On the other
hand, with the exception of the LXX in 9:27, they never use it
elsewhere in the Hebrew chapters although it is acceptable Greek.
The literalism of the occurrences in Daniel extends also in every
instance but one to the expression of a personal pronoun as the
subject of the infinitive, corresponding to the pronominal affix
of the Hebrew.
'I6od occurs with fair frequency in Daniel and is used in
the Hebrew sections with one exception as a translation of .nn.
8 and the LXX coincide on 10 uses5 and the former has one, the
latter two others. In 8:3 the LXX omits and in 10:8 it adds the
expression. In 9:4 it is translating'Inx, omitted by e. There
are as many more occurrences in the Aramaic, employed to translate
1iu or tX. The literalism of the expression6 should be extended
to include the xoa which is generally prefixed to it and the par-
enthetic form which it assumes. Without xa( it occurs only 8
times out of the total of 41 in both translations, and of these
at least 5 are without the conjunction in the original. Yet 4 of
the 6 occurrences in Bel and the Dragon and Susanna omit the xa(.
'E.c is also unduly frequent, occurring 21 times in 67
18:8,15,17,18; 10:9,11,15,19; 11:34; 12:7.
28:15,17; 9:27; 10:11,15,19; 11:2,4.
3z for 2 in 11:2 and 4.
4See J. H. Moulton, A Grammar of New Testament Greek, I
(Edinburgh, T. and T. Clark, 1908), 14, 215, 249.
DIn 8:5,15,19; 9:21; 10:5,10,15,16; 11:2; 12:5.
6See Moulton, or. git., p. 11 for its Greek character.
71:21; 8:6,7,8,10,11,13,14; 9:25,26,27; 10:3; 11:10,24,
35,45; 12:1,4 (twice),6,9.
and 13 in the LXX1 in the Hebrew chapters, all but one2 being
translations of Iy, whose conjunctive, adverbial, and preposition-
al functions this particular Greek word was fitted to fulfil. The
fact that over half the occurrences are prepositional, however,
makes the overusage especially un-Greek, and the LXX translator
has evidently exerted himself to avoid it. The various usages and
translations are illustrated by the following examples.
8:6 xal eX8ev int Tbv
8:15 ewc T(voc
11:55 iwC xaspoU ouv-
11:36 Lwc av auvT~Osoei
xKl hXe8v ieW ToD
EWC xatpoO IepaO
RIXP o o0ouvrEXroefa
The Hebrew construction p'Ki with a following participle
is responsible for an unidiomatic literalism in the combination
of a form of sevat in the third person singular with a participle
usually arthrous as subject or as attributive to an unexpressed
subject. Such occur in 8:4,8,27; 10:21; 11:16 and 45, all of them
read by both 8 and the LXX. In the Hebrew 8:8 alone has the vari-
ant reading n'n KSl. The construction is best described as well
as understood by means of some actual examples.
8:4,8 xal oOx lvv 6 pu6-
10:16 Kal oUx EGtot 6
10:21 xal oCueEe v 6
xal o6x fotav atpod
xal oix aoTyv EIc
11:21; 8:10,11,15,14; 9:27; 10:3; 11:55f.; 12:1,4,9.
2@ so renders 3xM in 8:7.
rp ny ip
'Kn IVy X1'1
ptnno inx p1,
An interesting literalism is to be found in the intransi-
tive use of notstv for the Hebrew verb nap, usually in conjunction
with another verb as, for example, in 8:12, xol eno(srs xal eso6 -
O1.1 Other instances are found in 8:4,24; 11:3,7,30,32,36 and 35
in both translations. Again it is the frequency rather than the
usage itself that is un-Greek.
A solecistic redundancy of expression occurs a few times
in attempts to reproduce completely the subordinate clause which
is introduced by Irm and uses other words to define its nature
and relationship to the main clause. There are 4 of these in e,
2 in 1:4 and 1 each in 9:7 and 18, the last two also oeing read by
the LXX. Those of 1:4 have been given on page 3 in illustration
of LXX paraphrase. The other 2 are reproduced below.
LXX e MT
9:7 eic &; 8tEcx6pntras oL 8tic eipac ab6Tou Cz cnnnn i~n
aCrobt xeT IxKE
avoi6 Cou in' arT( avog6 gou in' 0a6iTC nwkp
That the LXX should twice adopt these readings is not startling
since, as we have seen, it often lapses into the most abject
The physical expressions characteristically employed in
Semitic languages to convey ideas are prominent in both of the
translations. The LXX substitutes more idiomatic phraseology in
only a few instances, e.g., in 1:8 and 8:33 as already indicated
on page 4. A few other outstanding occurrences are here added.4
LXX 8 MT
1:4 elvat iv Tr olx ieoT&vat iv tr. oixi l7cn ':rin Ir3 p
too paCtxu( TOU paoi\Ey<
e1 and the LXX for annlSni nnzpi.
2The construction is not without Kotru parallels. I noted
a partial one in several U P Z documents, e.g. No. 8, 11. 9f.,
Tb ... 'AaapTeov, iv Truyx6vw iv Ti1 xoroxrt yeyovic, K.T.X.
3Cf. 10:5 where the LXX reads with 8.
4See also 1:7; 8:5; 9:12,18,19; 10:5,16, etc.
LXX 6 MT
10:12 I8wxac 'rb op6o'nx6v 58wxac Trv Kapblav >rlz 13 Am na nn
cou b6avorv~vat cou roU ouvvevat
10:15 i8wxac T b Ap6ou6v Same n3Ix 'e 'nn
aou itl t~h Yv
A relative infrequency of compound verbs and adjectives
and of adjectives in the comparative and superlative degrees is
also characteristic of the translations. The LXX, however, uses
almost twice as many compound verbs as 0. In chapter 1 it has a
total of 23 as against 14 by 6 It also has more adjectives than
0, a total of 68 to 41 in the 2 chapters exclusive of substantive
usages. But approximately 1/3 of these are numerals, and among
them, also, the LXX has only 5 and 6 3 compound forms, 2 of which
in each are numerals. Among the remainder the LXX has 4 compara-
tives and 0 2, 3 superlatives and e 1. All 4 of the latter are
forms of pp:~roc. It is clear that 8 is satisfied with reproducing
the adjectives of the Hebrew, whereas the LXX often introduces
them in place of other constructions, particularly that of the
attributive genitive. On the other hand there is an unmistakable
tendency in both translations toward simple verbs and simple
adjectives of positive degree.
The matter of word order in Daniel constitutes an out-
standing and important literalism, especially in 0. It is not
treated here because a full presentation of it may be found in
the study made by Prbfessor Rife.1 His findings indicate that
translation Greek abandons the hypotaotio and inteopolative word
order of natural Greek for the order of the Semitic original.
This he has statistically stated for the relative order of subject,
verb, and object as exhibited in the Hebrew, Aramaic, and Greek
translations of Daniel compared with composition Greek of every
variety from classical to modern.
1Q2. Ag., chap. 4.
SUMMARY OP DIFFERENCES
Further marks of literal translation are revealed by a
mere comparative survey of the differences between the two trans-
lations in their departures from the Hebrew. The divergences in
the two sections under consideration, as they consist in omissions
and additions and in alterations of nominal and verbal forms, are
statistically summarized and discussed below. The tabulation of
additions and omissions is exclusive of the article, preposition,
and conjunction, which are treated separately.
Apart from these and from all omissions which are textual,
i.e., either not in the translator's text or due to the vicissi-
tudes of transmission, there are 28 words omitted entirely, 11 by
6 and 18 by the LXX. The relatively high number in 8 is due to
the inexplicable omission of 8 words in 8:2, at least 2 of which,
however, may be textual.1 Otherwise these omissions are all at-
tributable to free and idiomatic rendering of the Hebrew. A good
number of them consist in no more than the omission of pronominal
affixes. There were actually more omissions than these, especial-
ly in the LXX, where the Greek rendered the Hebrew meaning ade-
quately in fewer words, but such were not included in the count.
Yet it is here that the chief difference between 8 and the LXX
lies. The former endeavors with few exceptions2 faithfully to
represent every word of his text in Greek; the latter is often
content to render the meaning irrespective of the number of words
or their exact correspondence to the original.
This difference between the two is even more pronounced
1Added by V a122 Compl Aid Alex Chrysbis Arab SlavOstrag
(according to Holmes and Parsons) and marked as an addition on
the other hand by the Syro-Hexaplaric and Chigi USS.
in the matter of additions of words and phrases. In a total of
451 8 has 13 and the LXX 42, 10 in chapter 8 coinciding. Theo-
dotion's 3 independent additions (1:4,20; 8:21) and 2 shared by
the LXX in 8:2 are merely forms of esvaL implied in the Hebrew.
Of the remaining 8, 3 others are also implied in the text (8:8 and
22 x4para; 8:14 71ipct) and the other 5 are each additions of one
word necessary to the Greek, all shared by the LXX.2 In the LXX,
in addition to the 5 noted as common to both translations, 3 others
in 8:5,20, and 21 are forms of .evat implied in the original, 3
in 1:2,12, and 8:19 are perhaps textual, and 6 are apparently glos-
ses.5 ToUTo in 8:17 is probably a rendering of the article of
nirnn with its original demonstrative force. At least 24 are due
to free and paraphrastic translation. A translator like 8, how-
ever, clearly refrains as much as possible from making additions
that are not absolutely necessary.
The additions and omissions of the article would be about
equal in number were it not for the special factor of the Hebrew
definite construct forms which generally require its presence in
the Greek. Of a total of 104 different additions,5 of which 74
coincide in the two translations, there are 73 due to constructs,
57 of them coinciding.6 Of the remaining 31 (8-23, LXX-25, coin-
oide-17), 12-are added before proper names,7 4, of which 3 coin-
cide, make the noun definite in Greek,8 6, of which 5 coincide,
are due to feeling for Greek idiom or its demands,9 4 have generic
force,10 and 3 are for 7 (8) and 2 for z (both) before infinitives
1Chap. 1-22 (8-2, LXX-20); chap. 8-25 (6-11, LXX-22).
28:8 a6'ou(1),17 and 19 iT t, 19i Opa Xc (LXX ilevs), 23 auQtv.
5In 1:10,17 (bis),20 (bis); 8:16. The last 5 are so marked
in the Syro-Hexaplar US.
4E.g., in 1:3,5, 12,13,15,17,18,21; 2:3; 8:2, 3,4,5,26,27, etc.
5Not including those incidentally made within larger addi-
tions or glosses. The same applies to omissions. Inclusion of
these would greatly increase the totals for the LXX.
Chap. 1-45 (6-45, LXX-40, coincide-40); chap. 8-28 (8-23,
SBoth in 1:7,9,10 and 8:21; 8 in 1:2; the LXX in 1:2,3,11,
19. The nN in 1:2 and 9 is perhaps responsible for those of 8.
8E.g., 1:15, "the ten days". In 8:7 rip is really definite.
9E.g., 2:1, Tr 86surpy.
10E.g., 8:4, T& eptla (both).
in 2:2,3; 8:7,15, and 17. The high degree of coincidence which is
exhibited by these additions suggests that most of them are due to
the translator's sense of idiom.
Again the omissions are much fewer in number, especially
in 8. They total 7 in each chapter, of which e has 7 and the LXX
12, 5 coinciding. Two of these, in 8:2 and 5, together with 4 of
the others,1 were unnecessary to Greek idiom and usage. Four are
due to the freedom of the LXX. Both translators transliterate the
article of iDon in 1:11 and 16 as part of the noun. The LXX reads
KGp.toc for D'ntn in 1:9, and its omission in 8:5 is probably a
In the treatment of prepositions, however, omissions are
of much the greater frequency. This is attributable especially to
the constant Hebrew usage with nouns in expressing case relation-
ships, a function often adequately discharged in the Greek by the
case endings. In many instances the translators could have been
more literal by adopting prepositional phrases,3 but more often,
as in the expression of the indirect object, such procedure would
have resulted in unidiomatic and even impossible Greek. This fact
is attested by the high number of coincidences in the omissions,
49 in a total of 62. The total number for 8 is 54 and for theLXX
57.4 The preposition heads the list with 39 omissions. Among
these, 18, of which 14 coincide, are rendered by the dative, in
most cases of the indirect object. Eight in both 6 and the LXX
are omitted from before infinitives. Of the remainder, 8 are
rendered by the accusative of the direct object, 4 by a possessive
genitive, and 1 by an adverbial accugative. There are 7 omissions
of 3 (dative 3, genitive 2, adverbial accusative 1, para-
phrase 1), 7 of to genitivee 6, accusative 1), 4 of X (all
datives), and 1 each of cp, nN, ), p, and ip. Of these last 5
1@ in 1:17 and 8:3; the LXX in 8:9.
2Charles believes this theoriginal reading.
3E.g., in 8:17 'x is translated ati by the LXX, but npbc
pi by 8.
4Chap. 1-43 (0-38, LXX-40); chap. 8-19 (8-15, LXX-17).
These are exclusive of a number within larger omissions, e.g., in
1:5; 8:1,8,24, etc.
the first 2 are translated by datives and the others occur within
paraphrases. These statistics indicate that the translations are
not far apart in their number and treatment of omissions, although
there is a good deal of variety in the case translation of a given
Hebrew prepositional phrase.1 Where, also, the preposition itself
is translated there are, notwithstanding a tendency towards repre-
sentation no less than 27 divergences between 8 and the LXX in
these 2 chapters.
In spite of the profusion of prepositions in the Hebrew
the translators have added others to the number of 18, 15 by the
LXX and 8 by 6.2 All of those in chapter 8, however, are due to
the He directive of the Hebrew3 and the remainder occur in free
renderings by the LXX4 or in deference to Greek idiom.5 Three of
the last are due to accusatives of the end of motion where the
Hebrew may omit the preposition.
In the matter of conjunctions, that is to say xal, omis-
sions also outnumber additions two to one. Here it is the ubiqui-
ty of the Hebrew i that is responsible. Exclusive again of those
occurring within more extensive ones, there are 10 additions6 and
20 omissions.7 Although the additions are about equal in number
in 0 and the LXX, there is a wide difference in total omissions,
8 having only 5 and the LXX 17. Of the additions 6 are due to
Greek idiom, 4 of them in a series of proper names, and the others
are probably textual. Of the omissions 2 or 5 may be textual, 7
are due to altered constructions,8 and the remainder to Greek idiom
or free renderings. There are at least 7 of the last variety, all
of them in the LXX. Theodotion's variants, on the other hand, are
almost all attributable to the demands of idiom.
The translators have both rendered the Hebrew nominal forms
1See p. 19.
2Chap. 1-11 (8-4, LXX-10); chap. 8-7 (8-4, LXX-5). These
are exclusive of a half-dozen more within extensive additions.
3E.g., in 8:4,7, and 10.
4Seven in 1: 12,15,16,18, and 2:1.
5Four in 1: 1,2, and 20.
6Chap. 1-8 (8-6, LXX-6); chap. 8-2 (8-1, LXX-2).
7Chap. 1-8 (8-3, LXX-7); chap. 8-12 (6-2, LXX-10).
forms. 8ive are substitutions of participial for finite verbal
with a fairly high degree of accuracy. Yet they have altered the
number of at least 15 nouns, 12 of them from singular to plural.
The LXX has 9 and 6 7 of these divergences, most of them due to
free or idiomatic translation.1 Another item of interest is the
fact that the translators differ at least 28 times in their case
translation of the same nouns. Less than 1/4 of these are attri-
butable to the use of different prepositions with the nouns. In
most of the others the LXX has adopted paraphrastic readings, and
in general its renderings are more idiomatic and less literal than
those of 8.2
Verbal forms have been treated with less pious regard. A
total of 43 different deviations from the Hebrew were detected,
20 in mood,3 4 in tense, 10 in voice, 7 in number, and 2 in person.
There are 10 substitutions of participial for finite forms, 8 by
the LXX4 and 3 by 8,5 1 in 1:10 coinciding. Since those of 1:11
and 8:26, however, are probably merely due to different pointing
of the Hebrew text, this alteration is practically confined to the
LXX. Infinitives are read for finite forms 4 times in the LXX,
1 being shared with 8 in 1:5. In 1:12 the LXX renders the Hebrew
waw conjunctive of the imperfect after an imperative to express
purpose by the infinitive with Trs where 8 mistranslates. The
other 2 occur in paraphrases in 1:5 (8isoaeOa) and 2:1 (rapaxn8-
vai). Finite forms occur for the same participles in both trans-
lations at 1:10,16 and 8:19.6 The LXX has another in 8:8 for
moJy. Both in 8:23 read participles in a genitive absolute for
z with the infinitive. In 1:15 the LXX reads a subjunctive (pavM)
for an imperative.
Theodotion is chiefly responsible for the departures in
tense, only 1 of the 4 being shared by the LXX. This is as it
1Akurv (1:22) and axt(&Tv (1:23) may be textual.
2Eg., in 8:22, xar-& dv tlaXv am6tv for tv r' tI~iI aut~v
(8 for in ). But of. 2:4, npbc Tbv paa&Xia for T9 paais.X (1o0).
3Inoluding for convenience the infinitive and participle.
4In 1:1,10,11; 2:2; 8:18,22,27 (bis).
5In 1:10; 8:19,26.
6For Fn' inp, and 9r'V0I.
should be, for as far as the Hebrew is concerned the variants are all
really mistranslations. In 8:4 6 has an aorist (Iolnot v) for a
waw conversive of the perfect.1 In 8:17 he reads a present(nnLTu)
for a waw conversive of the imperfect and in the following verse
the same for a nifal perfect.2 In 8:4 0 reads a future (aTacov-
ar) and the LXX a simple aorist (la-raov) for a potential imper-
fect. There are also a few other differences between 8 and the
LXX which involve no recognizable departures from the Hebrew. A
kal infinitive in 8:7, e.g., is translated by a present (pXsa6eat)
in the LXX and by an aorist (6kOaev) in 0. In 8:13 the LXX reads
an imperfect (hxouov) and 0 an aorist .(?xouaa) for a waw conver-
sive of the imperfect. The rendering of the LXX here, in view of
the following piel participle, is perhaps the better.
All of the 10 actual variants in voice3 are by the LXX,
only 1 in 8:12 being shared by 6. Actives for passives occur in
8:1 (slbov and 6e.lv), and passives for actives in 1:18 (a4aixOe-
aav), 2:2 (AtoevsXevs vaL), 8:10 (!pp&X86 and cuaTnnacTc6), 8:12 (Ap-
p.Qly, 6-ap(il), and 8:25 (VeO7eTas~t). In 8:15 the LXX reads an
intransitive Oswpelv where the Hebrew followed by 0 has an active
infinitive with a direct object. The LXX has also altered the
text in 1:15 by reading an active imperative (Xpoat) with an ob-
ject (troc naict cou) where the Hebrew verb is intransitive and
followed by a prepositional phrase.4 The variant in 8:12,however,
is probably due merely to a different pointing of the Hebrew.
The variants in number are almost equally divided, the LXX
having 1 more than 8. The LXX reads plurals for singulars in 1:6
(haav) and 18 (seaiXnaa)v) and 6 in 1:10'(auvs7iE1cav),5 1:15
(up&6raav), and 1:19 (supiE0ieav). The LXX has singulars for plurals
in 1:12 (boOre ) and 13 (qpav). It also has both of the variants
in person, reading a first for second in 1:10 (xtvsuvs6eC) and a
first for third in 8:1 (Tb lbsTv ge).6
1The LXX reads InoCte.
2The LXX reads LnEav and exaitjvnv respectively.
3I.e., exclusive of mere forms.
46 reads literally nonaoov get tzv 6aifwv cou.
e translates T v 6Opseaav goL.
Among these verbal variants there were only 8 shared by 6
and the LXX. Almost all of the 38 by the latter occurred in free
or idiomatic translation, whereas at least 5 of the 13 by 6 were
due to misinterpretation of the Hebrew. There is no consistent
practice of alteration discernible in either, the only variant of
any frequency being the substitution of participles for finite
verbal forms by the LXX The aorist and future for the Hebrew
perfect and imperfect respectively are the favorite tenses of the
translators. There are a few imperfects and presents, the latter
mostly participles. Perfects are rare, the LXX having 3 and 6
only 1, all of them participles. This diminished usage is another
feature of the Kotv4, and yet Susanna, which is about equal in ex-
tent to the 2 Hebrew chapters, uses perfects 11 times. Subjunc-
tives are also infrequent, the LXX having 3 and b 4 in chapter 1,
3 of them coinciding. There are none in chapter 8.
Finally, there is a group of divergences in which Hebrew
words have been rendered by entirely different parts of speech.
There were some 31 instances of this, 24 by the LXX and 9 by 6.
The only ones worth mentioning as recurrent are the substitution
of adjectives for attributive genitives of the noun, 8 times in
the LXX,1 and the use of the article to translate the relative
Irm, 4 times in each.2 The slightly extended use of adjectives
in the LXX reveals, as we noted, another endeavor to escape from
1In 1:.5,4,5,10,,15; 8:2,27.
2In 1:11 and 8:21 (LXX); 8:19 and 26 (6); 1:10 bis (both).
As revealed by this study, then, the literalisms and the
translation characteristics of Daniel may be allocated to the fol-
1. Reproduction of original word order.
2. Representation of each word.
4. Systematic representation
5. Overuse of certain good, or possible but uncommon
6. Infrequency or absence of certain common Greek usages.
The order of these phenomena might be taken to represent roughly
their relative importance if for the LXX number 4 be placed after
6 and the positions of 2 and 3 reversed. The LXX departures from
the Hebrew, as we have seen, are generally more significant and
extensive as well as more numerous than those of 8. Yet it also
lapses into literalism continually, while 0 is inexplicably free
in places. The variants of the latter, however, are mostly due
to feeling for Greek idiom rather than to free translation.
Characteristic overusages noted were those of xal, per-
sonal pronouns, t8o6, 6v T with the infinitive, literal rendering
of physical expressions idiomatic to Hebrew, expressions of com-
parison by a prepositional phrase, Avuinov, EtC, intransitive xot-
a-v, unidiomatic representation of p'K with a participle, and a
redundant translation in irx clauses. Infrequent or absent were
subordinating idioms and conjunctions, particles, subjunctives,
perfects and imperfects, compound verbs and adjectives, and adjec-
tives of comparative or superlative degree.1
lCompound nouns were also infrequent.
In connection with the probable validity of conjectural
emendation of a Greek text on the basis of a hypothetical Semitic
original the departures from the Hebrew are significant. Exclu-
sive of those occurring within large additions or omissions and of
errors and transliterations there are at least 307 differences in
the LXX and 242 in 6 in the two chapters. The fact that 167 of
these coincide would seem at first to indicate that in 8 at least
deviations are extensively of a nature which might be predicated
from the Hebrew. But this can only be affirmed of the addition of
articles for Hebrew constructs and the omission of certain prepo-
sitions as already indicated. These two account for 57 and 49 of
the coincidences respectively. Yet there are 16 differences in
the treatment of constructs and 13 in that of these particular
prepositions. In relatively few, also, of the remaining coinci-
dences does the departure from the Hebrew extend to translation by
Greek that is identical in vocabulary or construction. We have
seen that the translations differ 28 times in their case usage of
nouns and 27 in their choice of a preposition when they are ren-
dering the same Hebrew. From the figures given above their total
minimum number of differences would amount to 382, but in reality
they are even more extensive. Except, also, for a few inevitable
commonplaces, neither of them exhibit any invariable or wholly
consistent lexical representation.
THE SEPTUAGINT AND THE NEW TESTAMENT
Professor Swete in his introductory volume to The Old Tes-
tament in Greek makes the following noteworthy statement:
The Septuagint is not less indispensable to the study of
the New Testament than to the Old. But its importance in the
former field is more often overlooked, since its connection
with the New Testament is less direct and obvious, except in
the case of express quotations from the Alexandrian version,
These, as we have seen, are so numerous that in the Synoptic
Gospels and in some of the Pauline Epistles they form a con-
siderable part of the text. But the New Testament has been
yet more widely and more deeply influenced by the version
through the subtler forces which shew themselves in countless
allusions, lying oftentimes below the surface of the words,
and in the use of a vocabulary derived from it, and in many
cases prepared by it for the higher service of the Gospel.1
In view of this kinship in thought and diction the translation
characteristics of the LXX should be admirable criteria for use in
evaluating the Greek of New Testament documents alleged to be trans-
lations from the Semitic. The extant non-literary KoLv has been
interrogated for New Testament usages alleged to be Semitisms.2
The New Testaments documents should be examined for phenomena that
have been established as translation characteristics from a study
of actual translations. The fact that the proponents of the trans-
lation hypothesis have employed neither of these criteria is an
indictment of both their methods and results.3
2Recently, e.g., the case for an Aramaic original of the
Fourth Gospel based upon lists of alleged Aramaisms has been de-
molished by Professor E. C. Colwell in Th Greek 2f theourth
Gospel (Chicago, The University of Chicago Press, 1931) through
a study primarily of Epictetus and the papyri.
For a comparative statement on methodology see Professor
D. I. Riddle, "The Logic of Translation Greek", The Journal g
Biblial Literature, LI (1932), 13-30.
The examination of the Gospels with respect to a few of
the translation characteristics of Daniel which were Semitic in
nature proved to be an experiment interestingly productive in re-
sult and highly suggestive of the applicability and effectiveness
of the method. The findings can be briefly summarized. Parataxis
is not nearly so pronounced as even in the LXX of Daniel.1 Subor-
dinate clauses and genitive absolutes are much more in evidence.
r&p and & ha, almost non-extant in Daniel, are very frequent. The
giv 86 construction occurs 20 times in Matthew, 6 in Mark, and
8 in John. Ai itself, which is found only 7 times in the entire
book of Daniel, occurs 9 times in the first 2 chapters of Mark and
is 2 or 3 times as frequent in Matthew. Adjectives and compound
verbs and nouns are more numerous. Comparatives and superlatives
are more frequently used, and when the former occurs it is almost
always followed by the genitive of comparison. There are, e.g.,
5 comparatives and 3 superlatives in Matthew's Sermon on the Mount,
4 of the former followed by the genitive. In the Fourth Gospel
there are 13 occurrences of gClewv alone, 12 of them with the geni-
On the other hand a number of literalisms are absent or
rare. The use of a prepositional phrase to express comparison is
found in the entire New Testament only in Luke (4 times) and He-
brews (8 times).2 Similarly lvntuov occurs nowhere in Matthew and
Mark and only once in John (20:33), but in Luke it is found no less
than 23 times and in Revelation 34 times. The statistics for tbo6
reveal a somewhat different situation. Matthew uses it 62 times,
Mark 6, Luke 56, and John 4. Other usages which are really only
Hebraisms and occur, as is well known, frequently in Luke but rare-
ly in the other Gospels, cannot legitimately be used as criteria
for judgment here.
The examples given suffice, however, for illustration.
They confirm the impression that Luke is consciously imitating the
style and vocabulary of the LXX. This would be natural for a
1See E. C. Colwell, 2a. ag., pp. 10ff., for statistics
on synde on and parataxis in the Fourth Gospel.
tAi with nap&.
Gentile and a Greek, but conscious or unconscious imitation of his
native speech idiom would be more natural for a Jew. Is this not
what we find, for example, in Matthew's t6o6 and T6TE and other
such overusages of expressions corresponding to Semitic idioms?
The presence of many Greek usages rarely found in translation Greek
would indicate that the Gospels as translations must be extraor-
dinarily free and paraphrastic, much more so than the LXX version
of Daniel. If this be so it would be a somewhat precarious proce-
dure to reconstruct hypothetic Semitic originals on the basis of
the Greek text. More significance should be attached to the ab-
sence in the Gospels of usages which can be established as trans-
lation characteristics than to a capricious overusage of certain
"literalisms" that are good Greek and can be easily explained as
due to the author's native idiom,