Vetus Testamentum X X X V I , 4 (1986)

T H E KETIB/QERE IN T H E ARAMAIC P O R T I O N S OF
EZRA A N D DANIEL

by

WILLIAM S. M O R R O W and ERNEST G. CLARKE

Toronto

T h e study of the K./Q(Ketib/Qere) readings in the small corpus of

Biblical Aramaic (BA), found in Ezra iv 8-vi 18, vii 12-26 and
Daniel ii 4-vii 28, is of particular importance to scholars who are in­
terested in the history of the development of the Aramaic language.
This article is concerned with the date and character of the Aramaic
language represented by the Tiberian vowel points in the biblical
text (Q). By way of contrast, most previous work on the date and
provenance of Β A has concentrated on the consonantal text (K). 1
At an earlier date, scholars studying the question of dialects in Ezra
and Daniel drew mainly on Old Aramaic (OA) texts which come
from the 8th-5th centuries B.C.E. or on Egyptian Aramaic, so-
called Imperial or Official Aramaic (OfA). Currently, the outcome
of this research is to suggest that the Aramaic of Κ in both Ezra and
Daniel antedates the emergence of Middle Aramaic dialects with
distinctive Western or Eastern features. The Κ appears to belong to
the stratum of OfA, albeit, in the case of Daniel it may belong to the
very end of that period. 2
Opinions vary regarding the identity of the Q with the Aramaic
represented in the K. Although it is often assumed that the
language represented in the vocalization of BA antedates the
3
language of targumic Aramaic and cognate literature at the pres­
ent time the linguistic provenance and character of the Q remains

1
See the summary of past research in Ε Y Kutscher, " A r a m a i c " , in Τ A
Sebeok (ed ), Current Trends in Linguistics 6 (The Hague, 1970), § 4 6
2
J A Fitzmyer, A Wandering Aramean (Chico, Calif , 1979), pp 6 1 , 77 η 32
3
G Dalman, Grammatik des judisch-palästinischen Aramäisch (Leipzig, 1905 2 ,
reprinted Darmstadt, 1960), pp 39-40, H Bauer and Ρ Leander, Grammatik des
Biblisch-Aramäischen (Halle, 1927, reprinted Hildesheim, 1962), § 1 j
 THE KETIB/QERE IN EZRA AND DANIEL 407

disputed. 4 Overshadowing this debate is a larger methodological

question. Although texts in O A and OfA have some comparative
value here, researchers in the area have typically appealed to the
language of the Targums to support their positions. However, it
must be emphasized that such a procedure is not without its dif­
ficulties, since the date and provenance of targumic literature re­
main more or less uncontrollable. As targumic literature is
available to us exclusively from mediaeval manuscripts, it is always
possible that these texts have been re-worked in the course of their
transmission just as is so often the case with texts in Mishnaic and
Mediaeval Hebrew. Hence, some doubt must always remain about
the date and dialect involved when targumic texts are being appeal­
ed t o . 5
In our opinion, this methodological impasse has been somewhat
relieved by the recent discoveries of a large number of texts in
Aramaic which are both datable and come from a known
geographical area i.e. Palestine. They are of major importance for
studying the dialect not only of the Κ but also of the Q.
J . A. Fitzmyer and D. J . Harrington 6 as well as K. Beyer 7 have
provided Aramaicists with an extensive corpus of texts in Aramaic
found in Palestine written in Early Palestinian Aramaic (EPA) from
approximately 200 B.C.E. to 135 C . E . , 8 the period of the final
redaction of Daniel. The numbering of the EPA texts quoted in this
article is from the Manual of Fitzmyer and Harrington. There is a
large degree of similarity between the vocabulary of these texts in
EPA and Β A so that a comparison is justified. More importantly,
some of the extra-biblical Aramaic texts are in the vernacular of the
language. For instance, the Bar Cochba letters can be assumed to
reflect the spoken language of the day. In addition to these texts in
vernacular Aramaic there are two other groups of texts more for­
mally literary from the same time and place which are valuable for

4
cf Kutscher, ρ 403, it is understood that Kutscher was never able to publish
the study in which he wished to advance the thesis of an identity between the Q and
Babylonian vocalization and Eastern Aramaic
5
A D York, " T h e Dating of Targumic L i t e r a t u r e " , JStJ 5 (1974), pp 49-62
6
A Manual of Palestinian Aramaic Texts (Rome, 1978) T h e citations from the
Manual are normally only those which are not reconstructed
7
Die aramäischen Texte vom Toten Meer (Gottingen, 1984)
8
For the purposes of this article the authors accept Fitzmyer's time constraints
placed on the corpus "texts from the last two centuries Β C and the first two
centuries A D (roughly from 200 Β C to A D 135)" (ρ xi)
408 WILLIAM S. MORROW AND ERNEST G. CLARKE

this study. First, texts from Murabba c at and the Synagogue In­
scriptions sem to reflect a more formal Aramaic than that in the Bar
Cochba letters, but were presumably understood by the people who
read them. Second, there are literary texts such as the Genesis
Apocryphon, the J o b Targum, and items such as the description of
the New Jerusalem with language and many ' 'historical" spellings
which could be considered influenced by BA. In fact some want to
call such instances archaizing. 9 Of primary significance for the pres­
ent study, however, are those examples of texts in Aramaic from
Palestine which deviate from the Κ of Β A towards the Q of Β A.
Beyond the dialectical question we are also concerned to address
the continuing discussion about the nature of K/Q notes in the
Masoretic tradition. The intention is to classify all the K / Q
readings and to account for the non-dialectical as well as the dialec­
tical features. In the following discussion the citations of the Κ
always precede the Q. No matter whether the vowels are added to
Κ as in some modern editions (BHK3 and BHS) or added to Q in
the margin one is expected to read only the Q . 1 0
Prior to our study R. Gordis 1 1 in 1937 published his valuable
study on the K/Q question, arguing that the Q w a s not a correction
or improvement of the Κ as much as it " w a s the preservation of the
text as it reached the Masoretes" (p. XIV). Gordis believed that
some K / Q provided a substitute for the divine name, and for inde­
cent terms. Secondly, he argued that some K/Q were intended to
aid the reader before the consonantal text was vocalized. Thirdly,
he suggested that some K / Q reflect variant manuscript readings. A
basic point made by Gordis is that K / Q means " w r i t t e n - r e a d " . In
a real sense the purpose of the operation, in Gordis's opinion, was
to safeguard the correct pronunciation of the main text i.e. Κ (pp.
X I V , 3-4, and lists 1-8).
12
Η. M . Orlinsky, in 1959, suggested that the K / Q system was
9
For instance, S A Kaufman, " T h e J o b T a r g u m from Q u m r a n " , JAOS 93
(1973), pp 317-27, esp ρ 325, " t h e literary Aramaic of the intertestamental
period an artificial, literary Aramaic primarily a conscious attempt to imitate a
'classical' language "
10
I Yeivin, Introduction to the Tibenan Masorah, translated and edited by E J
Revell, (Chico, Calif , 1980), §§ 93, 97
11
The Biblical Text in the Making A Study of the Kethib-Qere (Philadelphia, 1937,
reprinted New York, 1971), Gordis re-issued the book with a new introduction in
which he took note of the Hebrew material from Q u m r a n
12
" T h e ongin of the kethib-qere system a new a p p r o a c h " , SVT 7 (1960), pp
184-92
 THE K E T I B / Q E R E IN EZRA AND DANIEL 409

an almost mechanical collating of three manuscripts. Where there

were two manuscripts in agreement over against the reading of a
third one, that one became the Q . O n e can see that there was no
common ground here with Gordis's position.
Both of these theories raise a number of questions which became
the subject of the article by J a m e s Barr. 1 3 H e notes that except for
those K / Q which are substitutions for the divine name or for an in­
decent term most K / Q represent the change of only one consonant
(p.25). Barr's emphasis on this point is supported by the results of
our study and has broad implications. Given our understanding of
variant readings in other literature the minimal change in K / Q can­
not be explained on the basis of variant readings alone. Barr sug­
gests that " t h e Κ is the consonantal graphic tradition accepted for
the M T , the Q i s the oral reading tradition" (p.27). Furthermore,
the oral reading tradition may be very old if one accepts Barr's ex­
planation for the K / Q tradition in the text of Samuel which often
results in the Q of Samuel being the written form in Chronicles.
Barr suggests a different approach to the subject from that of
earlier researchers based on the " s o r t of difference made by the
transition from Κ to Q " (p.28). H e establishes five categories: (1)
instances where there is neither semantic nor phonetic change but
only " a change of spelling convention''; (2) instances where there is
" a real morphological or syntactic c h a n g e " ; (3) instances of a
phonetic but no semantic change; (4) instances of a semantic but no
phonetic change; and, (5) instances with both a semantic and
phonetic change. A second major category Barr notes consists of
those instances where " i t is clear that certain aspects of the K Q
system are connected with language c h a n g e " (p.32).
A major difficulty in any study of the K / Q is that the total
number of K / Q is not the same in all manuscripts nor, for that mat­
ter, in BHK3 and BHS. T h e source for the K / Q designations is the
corpus of marginal notes contained in the Masora Parva (Mp) to
the biblical text. Previously, Gordis had provided scholars with a
full list of K / Q notes based more or less on C. D. Ginsburg's edi­
tion of the Rabbinic Bible and Masorah. However, it must be noted
that Ginsburg's work was of an eclectic nature and the Masorah he
compiled is not homogeneous. 1 4 Latterly, scholars have had access
13
" A New Look at K e t h i b - Q e r e " , Oudtestamenfische Studien 21 (Leiden, 1981),
pp. 19-37.
14
A. Dotan, Thesaurus of the Tibenan Masorah (Tel-Aviv, 1977), p. X I I .
410 WILLIAM S. MORROW AND ERNEST G. CLARKE

to a much earlier complete set of Masoretic notes found in Codex

Leningrad Β 19A (L) copied in 1009 C.E.
Consequently, the list of K/Q used in this study is taken directly
from the M p of L. 1 5 As such, although our list is similar to the list
previously published by Gordis, certain omissions and additions
with reference to Gordis's work should be noted. At the basis of this
article's divergence is the decision to undertake the analysis on as
formal a basis as possible. Hence, excluded from the study are those
cases of K / Q which are recognized in the Rabbinic Bible and BHS,
but are not specifically noted in the M p of L. 1 6 The list which we
have compiled is the result of the observation that L has two types
of K / Q notations. First are those cases where qop appears as an ab­
breviation for qere in the M p , either beside a notation of the reading
it indicates or in a phrase of the kind he lamed qop ( = he not read; e.g.
D ii 31, 37). Second, it is also apparent that notes containing the
word ytyr (' 'excessive/otiose") indicate K / Q readings. This is clear
from D ν 8 where ytyr occurs beside a marginal note with qop as a
kind of secondary indication of a K/Q. Another example is in D vii
19 in which the same K / Q situation is marked once by qop (snyhlsnh)
and once by ytyr (tpryh/tprh); (see also D iii 5/15; D iv 16/E iv 12; D
vi 14/E vii 18). In terms of intention we also find no distinction
drawn in the M p between K/Q situations marked by ytyr and
recognized by Gordis and by the tradition of the Rabbinic Bible,
and others overlooked by both (D ii 21, iii 5,7, vii 7; E iv 12, 18, ν
17, vi 15). Therefore, the analysis below seeks to classify and
discuss all examples of notes with qop and/or ytyr found in the M p of
L, understanding each to be an indication of a K/Q.
In basing our study on the M p of L it would be remiss to ignore
the fact that this body of material has its own textual problems. For
example, the notation in D ii 29 (lamed = unique), might indicate
that the Masoretes were aware of a K / Q situation for the parallel
lexeme in D ν 10, as indicated in the apparatus oí BHS In D ii 35,
the notation lamed wytyr ^alep on the word mit shows that the

15
For a facsimile see D S Loewinger, Codex Leningrad Β 19A (Jerusalem, 1970)
16
As is shown by the variations found m the manuscripts and in the editors'
notes in BHK3 and BHS, found in Gordis but not in L D n 29, 38, m 2, iv 13 (sic'
read 15), 16, 22 (2x), 29, v i l , vu 4, 7, E i v 11, vi 15, 16, 18, vu 13, 24, found in G
E Weil's reconstruction of the M p of BHS but not in L D n 16, 38, m 2, 3, 29, iv
22 (2x), 29, vi 1, 16, E iv 8, 11, 13, vi 14, vu 21, special attention should be given
to D ii 16 and E iv 13 where the vocabulary of K / Q notes in L is used by Weil to
create pseudo-K/Ç) notes
 THE K E T I B / Q E R E IN EZRA AND DANIEL 411

Masoretes were aware of a reading in the Κ of mPt which is not

present in the text of L. Other examples of various textual corrup­
tions will be added below especially in section 1.6. Another puzzling
instance concerns the variant notes on the verb khl " t o be able' ' (see
1.3). Clearly, the M p of L has had its own complex transmission
history. In order to proceed on as formally a controlled basis as
possible, we have found it advisable to restrict the commentary to
those cases marked by the K / Q notes we have identified. For the
purposes of this study, therefore, we ignore inconsistencies in the
M p and unmarked situations and concentrate on those examples
where a K / Q situation is actually indicated in the M p of L. Future
scholars in this field ought to take note that the M p of L appears in
its most accurate form only in the edition of BHK3. It is a matter of
no small disappointment to note that the expanded form of the M p
contained in BHS renders that edition substantially useless for the
kind of study undertaken here. 1 7
Even though Barr's article served as an initial inspiration for this
study, we have not accepted his categories completely, although
recognizing his observation that, in the main, K / Q involves
minimal change. Consequently, this study has established only two
major categories of Q : (1) orthography and (2) morphology. Even
within these two general categories it would be possible to arrange
specific K / Q readings differently without in any way changing the
basic thrust of our analysis. With regard to the first category it
could be argued that examples such as nhyrVnhwr·*, ^nwP/^nP, and
slhlslw should be discussed under morphology. However, when one
examines the K / Q readings listed under the second category it is
clear that this deals with patterns of phonetic shift. While those
readings included in category one could, in some instances, be due
to scribal error or confusion, those in category two clearly reflect
changes in the language itself.
17
E Wurthwein, The Text of the Old Testament ([Grand Rapids and London,
1979 2 ], E tr of Der Text des Alten Testaments [Stuttgart, 1973 4 ]), pp 28-9, writes
" although the text of B H S reproduces manuscript L with the greatest felicity,
the editor of the Masora, G E Weil, is much freer with it T h e notes of the M p in
the margins of B H S are still based on the M p of L, but its terminology and ab­
breviations are made consistent in a standardized form, and its references are filled
out where the manuscript itself is incomplete T h e larger part of the M p in B H S ,
then, was supplied by the editor " Weil, himself, in BHS pp X I I I - X V I I I , ex­
plains his method of integrating and standardizing the M p notes O n e should also
note the following inaccuracies in the M p of BHS (D m 28 gsmhyn for gsmhwn, ν 8
psrh without the mappiq in the he, ν 21 °/Ä for QPh
412 WILLIAMS MORROW AND ERNEST G CLARKE

I T h e first group of K / Q to be studied reflects alternative spell­

ing conventions of the day, exclusive of language shift, which can
be further classified as (1) proper spelling, (2) historical spelling, (3)
Hebrew versus Aramaic forms, (4) vowel variation, (5) foreign
words, (6) purely textual, (7) gender corrections
I I A number of K / Q make clear the proper Aramaic spelling
With regard to D iv 5 ^hrynPhrn, the Κ could be taken as a plural
but the Q without yod shows that it is an adverb In the extra-
biblical evidence from EPA the word-form without the yod appears
consistent (see %nz 8 1 ι 18, ^hrn> 8 1 n 7, ^hrnyn (pi) 10 4, 12 4,
52 4, 60 3, ^hrny^ 7 8 4) In Egyptian Aramaic also the word is
18
without t h e j W In contrast, note the analogous unmarked situa­
tion in D vi 7 ( ^ln) All the other instances of this word m Daniel (n
40, 44, vi 3, vn 17) are spelled Hyn Clearly no Q i s needed for "*lyn
as no one could confuse the pronunciation The examples in EPA
texts reflect both spellings (cp Manual, ρ 308 Άη 12x and Hyn lx)
The same possible confusion between sere and hireq is likely respon­
sible for the notes on mhqym in D n 21 and for hqym in D m 5,7
where the M p notes that the yod is otiose By contrast this K / Q
situation for hqym in D in 2,3 is unmarked
Another spelling convention which K / Q notes concerns the verb
HI D iv 4, ν 8 Hlyn/Hyn, and D ν 10 HitIHt All the evidence from
EPA supports the Q, {Manual, pp 332 3) Also Egyptian Aramaic
confirms Q (Cowley, 7 8, 16 6) There are no examples in O A
In D vn 10 rbwnlrbbn is an attempt to write a feminine plural ab­
solute noun of a double cayin root which in the singular has an
abstract ending The Q signals the proper pronunciation of the waw
by reading beth
In D ν 8 psPlpsrh the ^alep has a dagesh and a Q exists indicating a
3 m s suffix O n the other hand, in D iv 15, 16 psP has no Q
although the M p says that the form is unique and, although the
vocalization of the Κ with a sere suggests a 3 m s suffix The context
could just as easily tolerate an emphatic state as reflected by the Κ
I 2 A number of the K / Q signal what the conventional spelling of
the Aramaic form could be
The first set of K / Q concerns the quiesence of the \lep which was
part of the root, (in D iv 16,21 mr^ylmry and in E iv 12 Pysthlbysth)

18
A E Cowley Aramaic Papyri of the Fifth Century Β C (Oxford 1923 reprinted
Osnabrück 1967) pp 274 5 under \r and \rn
 THE K E T I B / Q E R E IN EZRA AND DANIEL 413

T h e Q mry is well attested (see the Assur Ostracon 6; but mr^y

7,8,17; and Manual 12x but mr^y 23:2.3). T h e same observation ap­
plies to Ptrlbtr in D vii 7. 1 9 T h e Q i s attested in all instances in EPA
{Manual, p. 313; 6x) except one (64:1.2). Also for sysyVsysy (E vi 15)
the Q i s attested in EPA (14:29.2 sysy[ and 19:1.5 s]ysy).
In D vii 10 ypym (which along with the word Hmym in 29B:2.7;
20.13; 21.10, 12, could be an Aramaic plural) 2 0 has a Q Hpyn, the
plural in Aramaic acceptable to the Masoretes.
In D ii 9 the partial assimilation of taw to zayin in
hzmntwnlhzdmntwn is reflected in the Q. Evidence is lacking to ex­
plain the unusual spelling of the K. In examples from EPA one
finds support for Q (see 27:12;29B:21.25).
In thirteen instances of ^nthPnt in Daniel (D ii 29,31,37, iii 10, iv
15[2x],19; ν 13,18,22,23, vi 17, 21), the Q reflects an historical
spelling found both in Of A and in E vii 25 where \t is the written
form. In EPA there are three forms of this personal pronoun: *nth
predominates ( l l x ) as opposed to \t (49:1.9) and Η (60:3).
In Ε ν 15 (%/•>/) the Q a l s o reflects an historical spelling found in
OA.21
In D ii 39, does the form in the Κ of V°/Vc reflect an historical
spelling? Both the longer and shorter forms are attested in EPA for
the adverb " a b o v e " : mmcP 42:13, and the shorter mmH 6:1[2].
Another aspect of the group of K / Q related to historical spelling
concerns the spelling convention for foreign words which could be
expected to be the one in vogue at the time the Q w a s established.
In the Aramaic section of Ezra and Daniel there are five foreign
words occurring in some sixteen instances. O n e of these words con­
cerns the Q for the name of King Artaxerxes in Ezra. There are
fourteen occurrences of Artaxerxes, six in the Aramaic section (E iv
8, 11, 23, vi 14, vii 12, 21) and eight in the Hebrew section (E. iv 7,
vii 1, 7, 11, viii 1 ; Neh ii 1, ν 14, xiii 6). The Κ has a final \lep in all
fourteen instances and the Q considers it paragogic except in the

19
cp. D vii 6 where Ptr has no Q and D ii 39 where the prepositional form is
btrk.
20
J. A. Fitzmyer, The Genesis Apocryphon of (Qumran Cave I (Rome, 1971 2 ), ρ 84,
considers it a " H e b r e w e n d i n g " , but contra, E. Y. Kutscher, Gahlaean Aramaic
( R a m a t - G a n , 1976), pp.60 ff; G. Svedlund, The Aramaic Portions of the Pesiqta de Rab
Kahana (Uppsala, 1974), pp.23-4.
21
H. Donner and W. Röllig, Kanaanaische und Aramäische Inschriflen {ΚΑΤ)
(Wiesbaden, 1962): ΚΑΙ 202 A 9, 16; Β 8 (cp. R Degen, Altaramaische Grammatik
[Wiesbaden, 1969], ρ 59); JL4/214 29.
414 WILLIAMS MORROW AND ERNEST G CLARKE

Hebrew section in E iv 7 In EPA, generally, the final ^alep

represents a long vowel and not a shewa 2 2 Interestingly, there is no
Q for the interchange of sin and samek in any occurrence of this
word T h e spelling in both the Hebrew and Aramaic sections is
mixed It has been noted that the shift from sin to samek had already
taken place in some words 2 3 If Q i s a reading tradition there would
be no need for a notation even though sin and samek are different
consonants
I 3 This example of a spelling convention seeks to distinguish
between Hebrew and Aramaic and concerns the verb khl " t o be
able' ' In D ν 16 (2x) twkl has two Q notes {tkwl and tykwl) but in D
n 10 the Kywkl has no Q, although the M p says that the Κ does not
exist in the Aramaic language, and one Ms has a Qj, ykl In D ν 16
both Q readings stress the Aramaic pronunciation This is em­
phasized by the reading tykwl which, in addition, intends to clarify
the quality of the short vowel found in the first syllable By contrast,
the written formykl in D in 29 reflects the Q reading In EPA all the
instances lean to the Q form ykl 29B 20 17, yklw 29B 20 20, ykwl
29B 20 22, ykwlwn 29B 20 19, and tkwl 5 37 4
I 4 This sub-group of spelling conventions concerns vowel varia­
tions In D η 22 nhyrVnhwP the Κ may appear to be corrupt if one
refers to examples in EPA where it is always written as Q nhwP, for
example, 5 23 7, 5 29 6, 20 24 1, 23 2 6, 26 13 O n the other
hand, the Κ may reflect a noun form related to the abstract nhyrw in
D ν 11,14
In D iv 13,14 ^nwPPnP suggests that Q i s a clarification or even a
harmonization with the same word as it appears elsewhere in BA
Both forms appear in EPA but ^nsPnP predominates (see Manual
ρ 309, \s 22x and \ws 3x)
In D m 29 the Q f o r slhlslw is confirmed in D vi 5 and E iv 22, vi
9 where the written form is slw
In E vn 26 srswlsrsy indicates the Q's preference for yod ending in
the verbal noun formation while Κ reflects the waw possibility for
abstract nouns F Rosenthal in discussing noun patterns notes
24
both yod and waw as abstract noun endings

22
Fitzmyer Genesis Apocryphon ρ 196
23
M Sokoloff The Targum to Job from Çhimran Cave XI (Ramat Gan 1974)
pp 14 15
24
A Grammar of Biblical Aramaic (Wiesbaden, 1974), §§ 55 57
 THE K E T I B / Q E R E IN EZRA AND DANIEL 415

I 5 This sub-group of spelling conventions concerns the other

foreign words in Daniel Two words in the list of musical in­
struments in D in are subject to Q The first qytrws/qtrws (D m 5, 7,
10, 15) has a patah in the first syllable where the Greek has an iota
The Syriac is qitar The T a r g u m of Isaiah ν 12 reads qatrôs J F
Stenning notes that there are a number of examples of the inter-
change between patah and hireq 25 Clearly the patah vowel of Q in-
dicates the acceptable pronunciation
Is there a confusion in pronunciation between " a " and " i " in
Aramaic when reproducing a Greek iota? Two examples from EPA
show an inconsistency when reproducing a Greek vowel ktwn
(5 14 9) is from κιτών and qtwP (5 1 1 9 ) is from κοιτών Both are
written in this Aramaic without an indication of the first vowel Is it
possible, because both words are without a.yod that the pronuncia­
tion in EPA at this time was with an " a " as is reflected in the Q o f
qtrws?
T h e other musical instrument in D in subject to Q is from the
Greek συμφωνία It occurs three times with two different spellings
swmpnyh in D m 5,15 and sypnyh in D m 10 Although Q takes note
of the mem pe issue it may well have been the yod in the spelling of
the word in D in 10 which triggered the Q i n order to stress the " u "
vowel in the first syllable
The final two examples of foreign words with a Q
(ptysyhwn/ptsyhwn, D in 21 and hmwnP/hmnyP, D ν 7, 16, 29) are
probably based on Persian words There are no known parallels in
EPA
I 6 This sub-group concerns purely textual changes, for exam­
ple D η 43 dy/wdy, E iv 12 swry ^skllw/swry^skllw, and D iv 15, 16, ν
8 prP/prsh which has been discussed above Textual transmission
error also seems suggested for ^rkwy in E iv 9 where the anomalous
singular form is in contrast with the Q {^rkwy7) of the expected
plural Both the L X X and the Syriac support Q Carelessness is
also suggested in the variation for yod and waw in D m 19 where
again the L X X and the Syriac support Q in reading ^stny as op­
posed to the grammatically more unusual ^stnw In E iv 9 the alter­
nation between dhw·* and dhy·* may also have its origins in a similar
copying error The L X X reads the Κ while the Syriac reads the Q
which, with its present pointing, must have considered dhy·* as

25
The Targum of Isaiah (Oxford, 1949), ρ xxin
416 WILLIAM S. MORROW AND ERNEST G. CLARKE

another plural gentilic in the list 2 6 Textual considerations also

come into play in the variation between singular and plural for the
readinggsmyhwnlgsmhwn in D iii 28. It should be noted that the con­
sonantal form gsmyhwn is printed in BHS as opposed to the editorial
decision in BHK3 to print gsmhwn 2 7 The Q is consistent with
gsmhwn in D iii 27. Finally, the K/Qswylswyw in D ν 21, although
both forms are plausible in context, may be due to a haplography of
waw in the K.
A most difficult textual situation occurs in D iv 19 with the K / Q
rbytlrbt M a n y commentators assume that rbt represents a 3 f.s.
perfect in contrast to the 2 m.s. of the Κ (Gordis, p. 199 η . 20).
However, this makes no sense in the context, and it is doubtful
whether the Masoretic tradition ever understood this lexeme other
than as a 2 m.s. perfect. O u r opinion is buttressed by the inter­
pretation found in Saadia Gaon's commentary, which was
presumably based on the Tiberian tradition. Moreover, a
manuscript from a non-Tiberian tradition, with Babylonian point­
ing 2 8 clearly displays the required 2 m.s. vocalization. Since the
form of Q i s unique (no similar form has been found in EPA, for in­
stance) it is unlikely to be a dialectical variant. Hence, we are at a
loss to describe it other than as a reflection of some older corruption
which has worked itself into the reading tradition.
1.7 A final sub-group concerns corrections towards the 3 f.pl.
gender in personal suffixes and verbal forms. For example:
= Κ bynyhwn = Q bynyhyn D vii 8 (see also D ii 33 [2x], 41 [2x],
42 [2x], vii 19)
npqw npqh D ν 5 (see also D iv 9, vii 8, 20)

There is evidence for the 3 f.pl. verbal form in EPA (yldn 5:32.2;
ypsn 5:32.3; ydbqn 5:36.2; ypqn 5:36.2,7; slm> 29B:22.28; yhrbn
28:2.14). Since a 3 f.pl. suffix is written male in mwldhyn (5:32.2)
this suggests that in the forms Hyhn 5:32.4; 5:37.5; bnyhn 5:32.2;
kwlhn 29B:20.7 the suffixal ending hn represents hyn also. Q may

26
M Zar-Kabod, Ezra u Nehemyah (Jerusalem, 1980), ρ 30
27
In Lyod is written above the line with no space between mem and he to accom­
modate it
28
Published by S M o r a g in The Book of Daniel (Jerusalem, 1973), according to
M o r a g (p vin), the m a n u s c n p t is to be dated to the 14th century C Ε , however, its
textual features reflect the work of a Masoretic School which flourished some cen­
turies earlier
 THE K E T I B / Q E R E IN EZRA AND DANIEL 417

arise from deliberate editorial activity here in order to " i m p r o v e "

the grammar of K.

II. The second major group of K / Q according to our schema con­

cerns morphological changes in the language.

II. 1 A major sub-group in the category of morphological change

concerns the loss of the plural inflexion on certain forms of suffixed
masculine plural nouns and prepositions which govern their suf­
fixes in an analogous fashion. This feature occurs throughout Β A
and involves the 3 f.s., 2 m.s., and 1 c.pl. suffixes. Examples are:
= Κ rglyh = Q rglh D vii 7,19 (see also D iv 14, ν 21, vii
5, 6, 19 [2x])
qdmyh qdmh D vii 7,8,20
c c
bdyk bdk D ii 4 (see also D ii 26, iii 12, 18, iv
16 [2x], 24, ν 10, 23 [2x], vi 23; E vii
18 [2x])
Hyk Hk D iii 12, iv 24, 29, ν 14, 16, vi 14
ytyn^ ytn^ D iii 18 (see also E iv 18, ν 17)
The form of Q c a n n o t be explained as a phonological shift, since the
only shift apparent in BA results in the reduction of the diphthong
ay to ê not qames or patah. The situation here is rather to be explained
on the basis of analogy: singular suffixal patterns have been
substituted for the original plural patterns. 2 9
No good evidence for the phenomenon can be found in the EPA
material. At best, a few examples can be established which suggest
that writers of EPA texts were not always consistent when applying
suffixes to prepositions which use the plural pattern as opposed to
those which do not. 30 By and large, however, scribal conventions in
EPA represent the diphthong ay accurately and one must conclude
that the change under discussion had not emerged as a significant
linguistic trait in the period.
II.2 Another sub-group shows the dissimilation of yod to \lep
when preceded by the vowel qames and followed by another vowel.

29
J Barth, Die Pronominalbildung in den semitischen Sprachen (Leipzig, 1913,
reprinted Hildesheim, 1967), ρ 46
30
44 1 3, 52 6 {Hh) where the antecedent is a masculine singular noun, cp Hwhy
in, for example, 5 24 8, also, 39 8 {Iqwblyk) and 51 13 {Iqwblk, which in the editto
princeps is read Iwqblk), even in OfA this preposition takes suffixes using the
singular pattern
418 WILLIAM S. MORROW AND ERNEST G. CLARKE

The shift from äyä to Pä is found in singular nouns which are in the
feminine absolute state and in the masculine determined state. This
feature occurs throughout Β A where the basic form ends in ay. Ex­
amples are:
= Κ rbyQyh = Q rby^h D ii 40, iii 25, vii 7,23
ksdy* ksd^h D ν 30; Ε ν 12
See also D ii 39, iii 26, 32, iv 14, 21, 31, ν 18, 21, vi 29, vii 25; E v i
17. It is noteworthy that this shift is not wholly regular in BA. It has
not taken place in infinitival forms with a similar pattern: cp.
Ihhwyh (D ii 10, 16, 27), Ihsnyh (D vi 9, 16, vii 25; E vi 12), and Ihytyh
(D iii 13, ν 3). 3 1
The other locus for this phonological shift is in the plural gen­
tilles. Here again one encounters a mixed situation. While the form
ksdy^ regularly appears in Q as ksd^y (D ii 5, 10, iv 4, ν 7),the shift
does not take place in other examples of this pattern; see for exam­
ple, the list in E iv 9. It is difficult to account for this alternation in
forms. Evidently one must date this Q tradition to an era in which
the dissimilation was occurring but had not yet taken hold com­
pletely. In the light of this mixed situation, the evidence from EPA
is of some value, for one is able to adduce examples which reflect a
similar state of affairs.
In the case of singular noun forms in -ay, one can construct the
following table:
=K V 23:2.6; 29B:2.4
tlyty* 8:i.i.3
bryi 8:i.ii.2
c
nyh 145C:1
tdmryh 53:3
=Q V 5:14.6
VkP 108:1
We assume that the examples listed as corresponding to Q are
determined adjectives in which the dissimilation has taken place but
which lack a vowel letter to indicate the second qames The context
of these lexemes certainly suggests this conclusion.
As regards plural gentilics the following table can be constructed:
=K zwmzmy^ 29B:21.29
V y 29B:21.29
31
S Morag, "Biblical Aramaic in Gaonic Babylon", Studies in Egyptology and
Linguistics in Honour of H f Polotsky (Jerusalem, 1964) = IEJ 14 (1964), pp 117-31,
esp ρ 124
 THE K E T I B / Ç ) E R E IN EZRA AND DANIEL 419

hwry* 29B.-21.29
rhmyh 58:2
=Q >mwr» 29B:21.21
The evidence above suggests (1) that the dissimilation in question
was already appearing in the spoken speech of EPA communities,
and (2) that some of the readings in EPA texts which resemble Κ
may be historical spellings. 3 2 The Q tradition in Β A seems to sug­
gest a gradual displacement of an older conservative form of pro­
nunciation preserved in K, by the younger vernacular. T h e mixed
situation in Q w i l l reflect the freezing of this process at a given time
when the reading tradition became firmly established. Although we
are unable to date the time exactly, it should be reiterated that the
situation observable in EPA is commensurate with a text where
historical spelling traditions have begun to be supplanted by the
vernacular.
II.3 T h e third sub-group dealing with morphological change con­
cerns the dissimilation of \lep to yod in the masculine plural of the
participles in medial weak verbs:
=K d?ryn =Q dyryn D ii 38, iii 31, vi 26
d?ry dyry D iv 32 (2x)
See also D iii 3, ν 19, vi 27; E vii 25 (spelled d^ynyn in L). This
group is striking because it appears to be a reversal of the dissimila­
tion of yod to *alep discussed above. In fact, this dissimilation has
already occurred in these middle weak participles. The Κ reflects
the dissimilation of an original *qayëm etc. to qa^em, the typical form
attested in BA (Bauer-Leander § 58 d). The form of the participle
with original jw¿/ is still observable in Of A texts, while the dissimila-
tion to \lep may be dated as early as the 3rd century B.C.E. 3 3 By
the time of the composition of the text of Daniel this shift had taken
place throughout these middle weak verb forms.
The Q , therefore, would appear to be a restoration to the original
root letter. However, the shift t o j W i s likely not an intended return
to the older form. 34 Rather, it is likely related to a shift of stress with
the yod replacing the \lep specifically in those cases where the sec-

32
Note the additional evidence for Q in the Manual's text of Megillat Tacanit
rwm^y (150 6) andyhwd^y (150 12)
33
S Segert, Altaramaische Grammatik (Leipzig, 1975), § 5 7 6 5 2
34
Compare the treatment of verbs with ^alep as the middle radical in Targumic
Aramaic, see W Β Stevenson, Grammar of Palestinian Aramaic (2nd edn, Oxford,
1962), § 23 9
420 WILLIAM S. MORROW AND ERNEST G. CLARKE

ond syllable loses the stress because of inflexion. T h e retention of

the \lep in the form q^my^ (D vii 16) is possibly due to the shift of ac­
cent two places from the basic position, leaving a remnant of secon­
dary stress in the second syllable (Rosenthal § 137).
The forms nby^h/nby^ in Ε ν 1, vi 14 and nby^yVnbyy^ in Ε ν 1, 2
show a similar dissimilation which can also be related to the loss of
stress from the final syllable of the root word.
None of the Q readings in II.3 can be observed in the consonan­
tal writing of EPA texts studied. We are forced to conclude that
they reflect the influence of a later form of pronunciation on the
reading tradition in the M T .
The discussion above sets out evidence which reflects not only on
a probable date and provenance for the reading tradition, but also
on the origin and meaning of K / Q notes in the M p of L.
So far as the origin of the K / Q notes is concerned, it is clear that
the divergences from the consonantal text which they record have
arisen for a number of different reasons. In some instances
deliberate editorial activity is indicated (I.2(?); 1.7) while others are
likely the result of language change (II. 1-3) or even mechanical
transmission error (1.6). It does not seem likely that activity entail­
ing textual collation or correction will explain any of the K / Q
readings in BA. Clearly, grammatical or morphological
divergences are involved in a number of categories (1.2, 3, 4, 7;
II. 1). However, it is noteworthy that the Masoretic notes involved
merely draw the reader's attention to the differences, much as they
would for an unusual form or spelling. This evidence suggests that
it is advisable to draw a sharp distinction between the processes in­
volved in the creation of the reading tradition and the making of the
K / Q notes proper. The origin of those notes in BA can be explained
by assuming that the Tiberian Masoretes were in possession of a
received, fixed reading tradition as well as a received consonantal
text as Barr has suggested. We submit, therefore, that the
Masoretes are best perceived as recorders of the reading tradition
rather than as having a direct part in its creation. In the process of
bringing these two traditions together, discrepancies were noted
where the reading tradition did not mesh with the vocalization sug­
gested by the consonantal text. The divergences were the source for
the present K / Q notes in the M p .
Regarding the purpose of the K / Q notes, we find it difficult to ac­
cept Barr's judgement (pp. 36-7) that, like the majority of
 THE K E T I B / Q E R E IN EZRA AND DANIEL 421

Masoretic notes, the K / Q notes were also meant simply to protect

the consonantal text. For instance, some notes such as those in 1.1
which deal with the distinction between sere and hireq male would
seem to be chiefly of interest to those concerned with the correct
pronunciation of the text. Another example is found in D vii 10
{rbbn) which is ascribed to Ben Asher himself. Furthermore, con­
sider the distinction between the notation qop and ytyr in section
II. 1. It seems more natural to considerjy/yr as a commentary on the
pronunciation of the text rather than on its orthography. O n e
wonders if the Masoretic activity can be said to have as its only con­
cern the preservation of the consonantal text. If this were true, it
would be difficult to understand why it was necessary to record fully
vocalized texts simply for the purpose. O n the face of it, it is likely
that the Masoretes also had an interest in preserving the reading
tradition which they had received. In this light, it seems possible
that the purpose of the K / Q notes was to avoid mutual contamina­
tion of the Κ by the Q, or the Q b y the K. In a situation where the
consonantal text was the authoritative liturgical entity, the
possibility that the received reading tradition might be compromis­
ed by a divergent vocalization suggested by the Κ must also be
entertained.
After examining the data collected from EPA, we conclude that
the reading tradition of BA can be considered to be a dialect related
to early Palestinian Aramaic. This position is based on the observa­
tion that there is a congruence between the orthography of the
Aramaic in the Q notes of the M p and that of EPA.
O n e of the major assumptions underlying this thesis is that the
Tiberian Masoretes would most likely have spoken a dialect of
Aramaic current in Palestine. Moreover, there is no reason to
doubt that the orthography used in creating the Q notes would have
reflected spelling conventions alive in their own dialect. This
assumption, that the Q notes reflect Palestinian orthography, has
been validated by comparing the orthography of the Q notes in the
c
M p and that of EPA materials. T h e double ayin verbs in 1.1, the
cases in 1.2-5 with analogies in EPA, the gender indications in 1.7,
as well as the spelling in II.2 all can be said to reflect orthographic
patterns present in EPA.
For this reason, one may argue that those cases where Q notes in
the M p reflect EPA orthography can also serve as a reliable guide to
the vocalization of the corresponding forms in EPA. T o what extent
422 WILLIAM S MORROW AND ERNEST G CLARKE

can this observed congruence between the Q notes of BA and EPA

be use to argue for a Palestinian provenance of the entire reading
tradition? Simply put, the Q n o t e s reveal no major exceptions to the
rules governing the pronunciation of the reading tradition
throughout BA Even where language shifts appear (II 1-3), they
can be shown to be derivative of an older state of affairs which, in
the case of II 2 and II 3, has not been completely obliterated in the
reading tradition Moreover, all these cases (as well as those involv­
ing historical spellings in I 2) show only partial divergences from
the vocalization of the lexeme assumed by the K, usually involving
only the inflection of one syllable (ι e Barr's idea of " m i n i m a l
c h a n g e " ) No grammar, as far as we know, has ever questioned the
integrity of the basic vocalization patterns of these lexemes with the
rest of the language of Q Consequently, we find no linguistic
grounds for separating the Q readings assumed to be congruent
with Palestinian Aramaic from the rest of the reading tradition
So far as dating this dialect is concerned, it must be noted that
EPA evidence for the language shifts of II 1 and II 3 has not been
uncovered Therefore, although one can assume that the Q
readings in these two categories come from the same linguistic en­
vironment as II 2, one can only conclude that the present form of
the reading tradition requires a terminus post quern somewhat later
than 135 C E Establishing a terminus a quo for the dialect of the Q i s
more difficult for two reasons First, one cannot be sure how early
the text of the Κ became stabilized For instance, the following
cases may reflect influence of the Q o n the transmission of the Κ (*nt
E vn 25, ykl D m 29, yhwd^yn D m 12, ksd^yn D m 8, ν 11) They
cannot be used to argue that the Κ form is early evidence for Q
Secondly, datable comparative materials are lacking One can only
hazard an educated guess based on the assumption that the Q w a s a
reading tradition with a fixed form by the time it was received by
the Masoretes
At present, it is thought that the T i b e n a n Masoretic activity
began within the period of 600-800 C E (Yeivm § 18) We suggest,
therefore, that the reading tradition for Β A most probably reflects a
dialect of Aramaic spoken in Palestine some time between 200-600
C E
 ^ s
Copyright and Use:

As an ATLAS user, you may print, download, or send articles for individual use
according to fair use as defined by U.S. and international copyright law and as
otherwise authorized under your respective ATLAS subscriber agreement.

No content may be copied or emailed to multiple sites or publicly posted without the
copyright holder(s)' express written permission. Any use, decompiling,
reproduction, or distribution of this journal in excess of fair use provisions may be a
violation of copyright law.

This journal is made available to you through the ATLAS collection with permission
from the copyright holder(s). The copyright holder for an entire issue of a journal
typically is the journal owner, who also may own the copyright in each article. However,
for certain articles, the author of the article may maintain the copyright in the article.
Please contact the copyright holder(s) to request permission to use an article or specific
work for any use not covered by the fair use provisions of the copyright laws or covered
by your respective ATLAS subscriber agreement. For information regarding the
copyright holder(s), please refer to the copyright information in the journal, if available,
or contact ATLA to request contact information for the copyright holder(s).

About ATLAS:

The ATLA Serials (ATLAS®) collection contains electronic versions of previously

published religion and theology journals reproduced with permission. The ATLAS
collection is owned and managed by the American Theological Library Association
(ATLA) and received initial funding from Lilly Endowment Inc.

The design and final form of this electronic document is the property of the American
Theological Library Association.