Grace Theological Journal (1985) 67-89.
[Copyright © 1985 Grace Theological Seminary; cited with permission;
digitally prepared for use at
KNOWN TO EZEKIEL
JAMES D. PRICE
Extensive evidence from ancient Near Eastern .texts and from
normal Hebrew syntax supports the view that wxr is a toponym in
Ezek 38:2, 3; 39:1. The syntactical support involves a detailed examina-
tion of instances where some scholars posit a break in a construct
chain. These hypothetical breaks are not convincing for several rea-
sons. Therefore, wxr in Ezek 38:2, 3; 39:1 should be translated as a
proper noun ("the prince of Rosh, Meshech, and Tubal" [NKJV]),
not an adjective ("the chief prince of Meshech and Tubal" [KJV]).
* * *
AMONG Bible expositors, controversy continues over the translation
of the phrase lbAtuv; j`w,m, wxro xyWin; in Ezek 38:2, 3 and 39:1--
should the translation be "the chief prince of Meshech and Tubal"
(AV), or "the prince of Rosh, Meshech, and Tubal" (NASB)? The
controversy centers around the Hebrew word wxro is the word a place
name (Rosh) or an adjective (chief)?
There are two principle arguments denying that wxro is a place
name: a philological argument and a grammatical argument. The
philological argument states that the primary meaning of wxro, is
"head" as a noun, and "chief" as an adjective,1 and that the word is
unknown as a place name in the Bible, Josephus, and other ancient
literature. J. Simons, a noted authority on ancient geography, wrote:
That in one or more of these texts a people of that name whose
home was in
but it is at any rate rendered improbable by the fact that the same
name can be discerned only very doubtfully in other (Assyrian) docu-
1 BDB. 910-11.
2 J. Simons, The Geographical and Topographical Texts of the Old Testament
(Leiden: E. J. Brill, 1959) 81.
68 GRACE THEOLOGICAL JOURNAL
The grammatical argument states that the absence of a conjunc-
tion between wxro, and j`w,m, precludes wxro, from being a noun. William
Gesenius stated the applicable grammatical principle: "Contrary to
English, which in lengthy enumerations uses and to connect only the
last member of the series, in Hebrew polysyndeton is customary.”3
This means that Hebrew uses a conjunction between every word in a
series. On the basis of this grammatical rule Simons concluded, "The
reading j`w,m, (not j`wem,U) in both texts argues against a tripartite enu-
meration of peoples or countries.”4
These arguments have been convincing to many scholars and
have resulted in the retention of the AV reading in a number of
modern versions (RSV, NIV, NAB). Ralph H. Alexander represented
the typical response when he wrote, "The author does not consider
the word ros [sic!] to be a proper name in light of the syntax of the
Masoretic text and the usage of the term throughout the Old Testa-
ment and extra-biblical literature.”5
But on the other hand, many authorities accept wxro, as a top-
onym, and regard the grammatical problem to be of no consequence.
Among these are C. F. Keil,6 C. L. Feinberg,7 D. J. Wiseman,8 T. G.
Pinches,9 and standard lexicons.10 Also, several modern versions
translate the phrase "prince of Rosh, Meshech, and Tubal" (ASV,
NASB, NEB, NKJV, Harkavy); and some even recognize the land of
Rosh in a reconstruction of the difficult Masoretic text of Isa 66:19,
"Meshech, Rosh, Tubal, and Javan" (NASB,
arguments against this translation may not be as convincing as some
Those who support the view that wxro, is a toponym observe that
this use of ros is not entirely unknown in the ancient literature.
Pinches pointed out that the LXX translators must have known the
place, because they transliterated the word as a place name. He also
3 GKC, 154a.
4 Simons, Geographical and Topographical Texts, 81.
5 Ralph H. Alexander, "A Fresh Look at Ezekiel 38 and 39," JETS 17 (1974) 161,
6 C. F. Keil, Biblical Commentary on the Prophecies of Ezekiel, trans. James
7 Charles L. Feinberg, The Prophecy of Ezekiel (Chicago: Moody, 1969) 219-20.
8 Donald J. Wiseman,
"Rosh," The New Bible Dictionary
Rapids: Eerdmans, 1962) 1107.
9 T. G. Pinches, "Rosh," International Standard Bible Encyclopedia (ed. James
I0 BDB, 912; William Gesenius, Hebrew and English Lexicon of the Old Testament,
trans. Edward Robinson (Boston: Crocker and Brewster, 1849) 955; William L. Holla-
Concise Hebrew and Aramaic Lexicon of the Old Testament (
Eerdmans, 1971) 329.
PRICE: ROSH: AN ANCIENT LAND KNOWN TO EZEKIEL 69
noted references to the
Sargon.11 Opponents of the view discount these references as insig-
Also, those who support the place-name view point to a much
more serious grammatical problem involved with regarding wxro as an
adjective--the adjective intervenes between the construct noun xyWin;
(prince of) and its genitive nomen rectum lbAtuv; j`w,m, (Meshech and
Tubal). This is a syntactic anomaly. Opponents of the view dismiss
the problem by observing that broken construct chains do occur in
Biblical Hebrew. Simons discounted the problem by stating, "The
translation of Eze. xxxviii 2.3 and xxxix 1 by “Gog, chief prince of
Meshech and Tubal' is grammatically difficult but cannot be said to
be impossible."12 But is is very doubtful that this problem can be
brushed off so lightly and that the ancient references to the land of
Rosh can be ignored.
This article demonstrates that Rosh was a well-known place in
antiquity as evidenced by numerous and varied references in the
ancient literature. The article also demonstrates that in Ezek 38:2, 3;
39:1 the absence of the conjunction with j`w,m, is inconsequential and it
is syntactically improbable that wxro, is an adjective. A logical explana-
tion is offered for the origin of the interpretation of wxr as an adjec-
tive. The conclusion is drawn that the best translation of Ezek 38:2, 3;
39:1 is "prince of Rosh, Meshech, and Tubal."
Rosh was a Well-Known Place
Rosh has not been recognized among the place names of antiquity
because scholars have failed to take into account the well known
phonetic shifts that occur within the Semitic languages. When differ-
ences in pronunciation are taken into account, I found the name
Rosh (or its phonetic equivalents) twenty times in five different
ancient sources without an exhaustive search.
Variant Pronunciations of Rosh
The word that means "head" as a noun and "chief" as an adjec-
tive is common to most of the Semitic languages, but its pronuncia-
tion varies. Due to the phonetic phenomenon known as the Canaanite
shift13 the word is pronounced ros in Hebrew and the Canaanite
11 Pinches, "Rosh," 4:2623.
12 Simons, Geographical and Topographical Texts, 81.
13 William S. LaSor,
Handbook of Biblical Hebrew (
1979) , The Semitic /a/ shifted to /o/ in the Canaanite dialects.
70 GRACE THEOLOGICAL JOURNAL
dialects,14 but in the other Semitic languages it is pronounced as rasu
(Arabic),15 res (Aramaic),16 ris/resu (Ugaritic),17 and resu/rosu (Akka-
dian).18 The final vowel (u) is the nominative case ending; alternative
final vowels supply the genitive (rasi/ resi) and the accusative (rasa /
resa). Wherever the Semitic word for "head/chief" was used as a
place name, it is expected that it would follow the pronunciation and
orthography of the language in which it was used. That was true for
most place names that were derived from the meaningful Semitic
Rosh was a Name
The word wxro, (ros or its phonetic equivalent ras / res) was not.
used exclusively as a common noun or adjective in the Semitic
languages. The word also was used as the name of persons and
places, and in compound names of persons and places. The use of
ros as the name of a specific land is demonstrated in the next sec-
tion. Rosh was the name of a son of Benjamin (Gen 46:21), and
Resh was the name of an Akkadian temple.19 Also, the word is found
in compound place names such as Resh-eni;20 and in modern Arabic
place names such as Ras Shamra, Ras Naqura, Ras el-Ain, etc.
Additionally, the word is found in many compound personal names
of antiquity, such as Rashi-ili,21 Resh-Adad king of Apishal,22 Resh-
beli father of Tubalit-Bini,23 Resh-Dumuzi,24 Resh-Ea,25 Resh-ili son
14 BDB, 910.
15 Ibid.. 910; the Semitic /s/ shifted to /s/ in Arabic.
16 Ibid., 1112.
17 Cyrus H. Gordon, Ugaritic Textbook (Rome: Pontifical Biblical Institute, 1965)
18 Theo Bauer, Akkadische Lesestucke (
338, 342, 344, 345.
20 David D. Luckenbill, "Bavian Inscription of Sennacherib," Historical Records of
wood, 1968) 2: 149.
21 Mentioned three times by David B. Weisberg, Texts from the Time of Nebuchad-
nezzar, vol. 17 in the Yale
Oriental Series: Babylonian Texts (
University, 1980) 17:63.
22 "The Sargon Chronicle," ANET, 266.
23 Stepheri D. Simmons, Early Old Babylonian Documents, vol. 14 in the Yale
Oriental Series: Babylonian Texts (New Haven: Yale University, 1978) 73.
12 in the Yale Oriental Series: Babylonian Texts (
25 Ibid., 50.
PRICE: ROSH: AN ANCIENT LAND KNOWN TO EZEKIEL 71
of Sulalum,26 Resh-Irra,27 Resh-Marduk son of Ipqu-Amurru,28 Resh-
Nabium,29 Resh-Shamash,30 Resh-Shubula son of Ibn-Adad,31 Resh-
Sin,32 and Resh-Zababa.33
Rosh Mentioned Twenty Times as a Place Name
The place name Rosh (or its phonetic equivalents in the respec-
tive languages) was found three times in the LXX, ten times in
Sargon's inscriptions, once on Assurbanipal's cylinder, once in Sen-
nacherib's annals, and five times on Ugaritic tablets--a total of twenty
references in five different sources. The following sections list the
Rosh in the LXX. The LXX translates Ezek 38:2, 3; 39:1 as
a@rxonta Rwj, Mosox kai> qobel. The Greek obviously transliterated
the Hebrew pronunciation.
Rosh in Sargon's Inscriptions. Various inscriptions of Sargon
(1) The Annals of Sargon (year 12, 11. 228-316):
Til-Hamba, Dunni-Shamshu, Bube, Hamanu, strong cities in the
(array) and entered Bit-Imbi.34
(2) Sargon's Display Inscription:
In the might and power of the great gods, my lords, . . . I cut
down all my foes. . . the lands of Ellipi and Rashi which are on the
Elamite border on the banks of the Tigris.35
(3) Sargon's Display Inscription of Salon XIV:
In the might of Assur, Nabu and Mardu, the great gods, my lords,
who sent forth my weapons, I cut down all my enemies . . . the lands of
26 Simmons, Early Old Babylonian Documents, 73.
27 Ibid., 73.
28 Feigin, Legal and Administrative Texts, 50.
29 Simmons, Early Old Babylonian Documents, 73.
30 A popular name, listed 3 times by Feigin, Legal and Administrative Texts, 50,
and 3 times by Simmons, Early Old Babylonian Documents, 73.
31 Feigin, Legal and Administrative Texts, 50.
32 Simmons, Early Old Babylonian Documents, 73.
Historical Records of
35 Ibid., 26.
72 GRACE THEOLOGICAL JOURNAL
Rashi and Ellipi which are on the Elamite frontier, the Arameans who
dwell on the banks of the
Rashu is also mentioned as a place name in the following additional
inscriptions of Sargon: (a) Sargon's Bull Inscription,37 (b) Sargon's
Pavement Inscription (mentioned 5 times),38 and (c) Sargon's Cylinder
Rosh in Assurbanipal's Texts. The
in Assurbanipal's Texts on the Rassam cylinder, the eighth campaign
In my eighth campaign, at the command of Assur and Ishtar, I
mustered my troops, (and) made straight for Ummanaldasi, king of
time I captured (together
Manamu with its (surrounding) district.40
in the annals of Sennacherib:
First year of Nergalushezib: . . . One year and 6 months was
day, his people made a rebellion against Hallashu, king of Elan, . . .
and killed him. . . Afterward
Sennacherib marched down to
and destroyed. . . (the country) from the
(Rishi is the equivalent of Reshu.)
Rosh in Ugaritic Literature. The Ugaritic literature mentions
people of the
(1) mit.tlt.mbhrm (1) One-hundred (and) three deficit
(2) ‘l nsk. kttglm (2) against the metal smith of Kttglm.
(3) arb’ m.tlt mhsrm (3) Forty-three deficit
(4) mtb’l.risy (4) (against) Motbaal the Reshite
36 Ibid., 41.
37 Ibid., 45-47; the Akkadian text spells the name ra-a-si. See D. G. Lyon, Keil-
schrifttexte Sargon s Konig von Assyirien (reprint;
Deutschen Demokratischen Republik, 1977) 14, 42 ,93.
38 Luckenbill, Historical Records of
39 Ibid., 60-62; Lyon, Keilschrifttexte, 2.
Historical Records of
41 ANET, 302.
42 Gordon, Ugaritic Textbook, 240.
PRICE: ROSH: AN ANCIENT LAND KNOWN TO EZEKIEL 73
(9) bms.mnt.tlt (9) five minas. Three
(10) ‘l mtn.risy (10) against Motan the Reshite.43
(Text 2078) 44
(1) rism.qnum (1) The Reshites: Qanum
(2) bn ilrs (2) the son of Ilrash
(3) etc. (3) etc.
(Text 2027) also a list of Reshites.45
(Text 2079) also a list of Reshites.46
(1) li.mat.lim.kbdsmn (1) Six hundred sixty kubdas of oil
(2) l.abrm.altyy (2) for Abram the Cypriote.
(3) mit.tltm.kbd.smn (3) One hundred thirty kubdas of oil
(4) l.abrm msrm (4) for Abram of Egypt.
(5) mitm.arb’m.tmn.kbd (5) Two hundred forty-eight kubdas
(6) for the men
(7) mit.l.bn.’zmt.risy (7) One hundred for Ben Azmot the
(8) etc. (8) etc.
These references to Rosh (Rashu / Reshu) demonstrate that it was
a well-known land in antiquity on the banks of
George C. Cameron, the noted historian of early
the land as "the Rashi tribe of Arameans, well known to the Assy-
rians from Sargon onward and located in the mountains east of Der,
where was its capital, Bit Imbi.”48 Other of its prominent cities were
Hamanu, Bube, Bit Bunakki, and Bit Arrabi.49
The cumulative effect of the preceding is that Rosh was a well
known place. The next section demonstrates that the word wxro, is
most probably not an adjective in Ezek 38:2, 3 and 39:1.
43 Translations of the Ugaritic materials are my own.
44 Gordon. Ugartic Textbook, 22*.
45 Ibid., 10*.
46 Ibid., 23*.
47 Ibid., 25*.
48 George C. Cameron. History of Early
49 Ibid., 200.
74 GRACE THEOLOGICAL JOURNAL
Contrary to the objection of Simons, the absence of a conjunc-
tion between wxr, and j`x,m, does not make it impossible for wxro, to be "
a place name. On the other hand, the fact that the word xyWin; (prince)
is a construct noun does make it extremely doubtful that wxro is an
Missing Conjunction is Inconsequential
Although it is customary for Hebrew to use conjunctions between
all the words in a series, it is not mandatory. Many exceptions to the
rule are found. After giving the previously noted rule of polysyndeton
in Hebrew, Gesenius cited the exception, "Sometimes, however, only
the last two words are joined."50 Examples are found in Gen 5:32
(tp,yA-tx,v; MHA-tx, Mwe-tx, Hano dl,Oy.va / 'And Noah begot Shem, Ham, and
Japheth'), Gen (NrAhA tx,v; rOHnA-tx, MrAb;xa-tx, dl,Oy.va / 'And he begot
livestock, in silver, and in gold '). This exception corresponds exactly
to the syntax of Ezek 38:2, 3; 39: 1; consequently wxro, can be a noun
in a series without violating normal conventions of Hebrew grammar.
Hebrew Syntax Expects wxro to Be a Name
If wxro is regarded as a name, then the syntax of the passage is in
keeping with the normal conventions of Hebrew grammar. In this
case, the construct noun xyWin; ('prince of') is followed by a compound
nomen rectum consisting of a series of three names (Rosh, Meshech,
and Tubal). Although Hebrew avoids lengthy series of coordinate
genitives depending on one nomen regens, numerous examples are
found in the Bible of short series of closely related words.51 Examples
are found in Gen 14:19 ("Possessor of heaven and earth"), Gen 28:5
("the mother of Jacob and Esau"), Exod ("the God of Abraham,
of Isaac, and of Jacob"),52 Num 20:5 ("a place of grain or figs or
vines or pomegranates"), 1 Sam 23:7 ("a town of gates and bars"), Ps
8:2 ("the mouth of babes and infants"), and Isa 22:5 ("a day of
trouble and treading down and perplexity").
These examples demonstrate that regarding wxro as a name con-
forms with known conventions of biblical Hebrew. However, the next
section demonstrates that regarding wxro as an adjective does not so
50 GKC. 154a; note other examples at Gen 10:1, 14:1, 30:39; Jer ; Ps 45:9.
51 GKC. 128a.
52 Note the absence of the conjunction between" Abraham" and "Isaac."
PRICE: ROSH: AN ANCIENT LAND KNOWN TO EZEKIEL 75
Syntax Rejects wxro as an Adjective
If wxro is regarded as an adjective, a syntactical anomaly re-
sults. One of the fundamental principles of Hebrew grammar is not
observed--a word normally does not intervene between a construct
noun and its nomen rectum. Joshua Blau stated the basic principle of
this convention of nonintervention, "Nothing must intervene between
the construct and the nomen rectum. Accordingly, even an adjective
attribute of the construct has to come after the nomen rectum.”53
As this convention applies to the words "lbAtuv; j`w,m, wxro xyWin; of
Ezek 38:2, 3; 39:1, it indicates that it is quite unlikely for the adjective
attribute wxro (chief) of the construct noun xyWin; (prince) to intervene
between the construct and the nomen rectum lbAtuv; j`w,m, (Meshech and
Tubal). Therefore, unless the principle of nonintervention permits
exceptions of this type, it is extremely improbable that wxro is an
adjective. Rather, it is extremely probable that it is a name in accord
with normal syntax. The following sections demonstrate that there
are no undisputed exceptions to the principle of nonintervention.
Hebrew Syntax Uses Other Constructions for Adjectives
When Hebrew expresses an adjective attribute for a construct
noun, it regularly uses other syntactic constructions. There are four
possible syntactic structures which could be used to express the
thought "chief prince of Meshech and Tubal."
(1) The absolute adjective may follow the nomen rectum, as
Blau's statement suggested. This construction is used most often.
Examples are found in 2 Sam (tlUtB;ha j`l,m.ha-tOnB; / 'the king's
virgin daughters') and Isa 55:3 (MynimAx<n.,ha dvidA ydes;ha / 'the sure mercies
of David'). When the statement becomes ambiguous or too complex,
alternate constructions are used. The use of this construction in Ezek
38:2 would produce the ambiguous phrase wxrohA lbAtuv; j`w,m, xyWin; where
wxrohA may modify lbATu or xyWin;. Therefore, the construction would be
(2) The construct adjective may be placed before the noun phrase
it modifies. In this case the Hebrew would read lbatuv; j`w,m, xyWin; wxro.
examples of this are MOy.ci
tBa tlaUtB; / 'virgin daughter of
(Isa 37:22), MyxiyWin.;ha ywexro / 'chief princes' (1 Chr ), MymiWAB; ywexro /
'chief spices' (Cant ), and tObxAhA ywexro / 'chief fathers' (1 Chr
53 Joshua Blau, A Grammar of Biblical Hebrew (Wiesbaden: Harrassowitz, 1976)
see also August Muller, Hebrew Syntax
76 GRACE THEOLOGICAL JOURNAL
(3) The absolute adjective may precede the noun phrase it modi-
fies. In this case the Hebrew would read lbAtuv; j`w,m, xyWin; wxro (hA).
Examples of this combination are found in Ezek (fwArA llAHA hTAxa
xyWin; ‘You, O
profane wicked prince of
(NOdyci tBa tlaUtB; hqAwAfum.;ha / 'You oppressed virgin daughter of Zidon'),
and Isa 52:2 (NOy.ci tBa hy.Abiw; / 'captive daughter of
(4) When a complex nomen regens prevents the attachment of a
genitive nomen rectum by means of a construct form, the genitive
may be attached by means of the preposition l.54 This occurs when
the nomen regens is a proper name, or has unmoveable modifiers.
Judg (bxAOml; NDer;y.aha tOrB;f;ma
Hag 1:1 (wv,yAr;dal; MyiTaw; tnaw;Bi / 'in the second year of Darius') have
examples of this construction. Although no example was found using
an attributive adjective, it seems probable that the construction xyWinA.ha
lbAtuv; j`w,m,l; wxrohA would accurately express "chief prince of Meshech
These examples demonstrate that Hebrew has regular syntactic
conventions for accommodating an adjective attribute of a construct
noun without violating the principle of nonintervention. Ezekiel used
these conventions in statements similar to 38:2, 3; 39:1 (see, e.g., Ezek
). It is highly unlikely that Ezekiel would violate such a widely
used principle of Hebrew grammar. The next section demonstrates
that alleged broken construct chains do not correspond to the syntax
at Ezek 38:2, 3; 39: I and do not justify regarding wxro as an adjective.
Ajective Modifying a Construct Does Not Apply
Some argue that, although it is unusual, there are certain cases
where attributive adjectives follow construct nouns, such as wxrohA NyeKo--
"chief priest” (2 Kgs 25: 18, Jer 52:24, etc.). However, the Ezekiel prob-
lem involves the possibility of an adjective intervening between a
construct noun and its nomen rectum, not merely following the
construct. Consequently such cases have no bearing on the Ezekiel
Broken Construct Chains Do Not Apply
Based on the evidences given by Gesenius,55 and supplemented
by M. Dahood56 and D. N. Freedman,57 some have concluded that
54 GKC, 129.
55 GKC, 130a-f.
56 M. Dahood, Psalms (AB; Garden City: Doubleday, 1970) 3:81-83.
57 David Noel Freedman, "The Broken Construct Chain," Bib 53 (1972) 534-36.
PRICE: ROSH: AN ANCIENT LAND KNOWN TO EZEKIEL 77
Hebrew grammar admits exceptions to the principle of noninterven-
tion called broken construct chains. Evidence was given by these
scholars citing several examples from the Hebrew Bible where a con-
struct noun is not followed immediately by a genitive nomen rectum.
These alleged broken construct chains are considered by some as
justification for regarding wxro as an adjective that legitimately breaks
the construct chain in Ezek 38:2.
Constructs are Created by Rhythm. The existence of alleged
broken construct chains in biblical Hebrew should not be accepted
hastily as justification for a broken chain in Ezek 38:2, 3, and 39:1.
Most syntactic constructions classified by Dahood and Freedman as
broken construct chains were previously noted by Gesenius, but were
not regarded by him as broken chains. The problem is that not every
construct form is a nomen regens that anticipates a genitive nomen
rectum. A construct form comes about when the language places two
words in such close rhythmical relationship that they receive only one
major accent. The first word of the pair loses its accent and its form
becomes a construct; the second word receives the major accent and
retains its standard (absolute) form. Gesenius said,
It is sufficiently evident. . . that the construct state is not strictly
to be regarded as a syntactical and logical phenomenon, but rather as
simply phonetic and rhythmical, depending on the circumstances of the
The genitive relationship between nouns regularly produces this
condition; the nomen regens has the construct form and the nomen
rectum has the absolute form. This construction is commonly known
as a construct chain. Because it is so common in Hebrew, it may
mistakenly be regarded as the only use of the construct form. Actually,
since the construct state is phonetic and rhythmical, not strictly syn-
tactical, Hebrew frequently exhibits other cases of the construct state
not associated with the genitive relationship between nouns. Gesenius
noted several such constructions: "The construct state . . . is frequently
employed in rapid narrative as a connective form, even apart from
the genitive relation.”59 The following constructions were listed by
Gesenius. (I) The construct state frequently governs prepositional
phrases, particularly in prophecy and poetry, especially when the con-
struct word is a participle. (2) The construct state frequently governs
a relative pronoun clause. (3) The construct state sometimes governs
an independent clause. This construction may be understood as a case
59 GKC, 130a-c.
78 GRACE THEOLOGICAL JOURNAL
where the relative pronoun is elided. (4) The construct state is some-
times followed by waw conjunctive where the connection is strong.
Miiller60 noted a few cases where a construct participle precedes an
accusative. An example is ytixo ytey;wAm; / 'those who serve me' (Jer
33:22). None of the above should be mistaken for a construct chain:
no genitive relationships were indicated. They represent the purely
phonetic and rhythmical cases.
The same phonetic and rhythmical conditions occasionally pro-
duce construct-like forms in other parts of speech, such as particles,
adverbs, prepositions, pronouns and verbs. When such words are
closely related to the following words by maqqeph, the accent is
drawn away from the word, resulting in a construct-like reduction of
the vowel. Examples of particles are wye versus -wy, versus -tx,, and
Nhe versus -Nh,. An example of an adverb is NKe versus xUh-NK, which is
found in Gen 44:10; Josh ; and 1 Sam 25:25. Examples of pre-
positions are tx, versus -tx, (with) and dfaBa (Cant 4: 1) versus NOl.Haha-dfaB
(Josh ). An example of a pronoun is hmA versus -hma (what?). Several
examples of verbs are found: drePAhi versus xnA-dr,PAhi ('please separate
yourself'; Gen 13:9), FBeha versus xnA-FB,ha ('please look'; Gen 15:5), bwe
versus MwA-bw, ('dwell there'; Gen 35:1), j`le versus xnA-j`l, ('please go';
Gen 37:14), bTok;yi versus MwA-bTAk;yi ('he wrote there'; Josh ), lwom;
versus UnBA-lwAm; ('rule over us'; Judg ), lwom;yi versus MBA-lwAm;yi ('he
will rule over them'; Isa 19:4, Joel ), rbofEya versus MfzA-rbAfEya ('the
indignation is past'; Isa 26:20), wDeqat;hii versus gHA-wD,qat;hii ('a festival is
kept'; Isa 30:29), j`lom;yi versus j`l,m,-j`lAm;yi ('a king will reign'; Isa 32:1),
smor;yi versus FyFi-smAr;yi ('he treads clay'; Isa 41:25), j`mot;x, versus OB-j`mAt;x,
('I uphold him'; Isa 42: 1), qc.ox, versus Myima-qc.Ax, ('I will pour water'; Isa
44:3), rBeDa versus qw,fo-rB,Da ('to speak oppression'; Isa 59:13), and rmow;yi
versus UnlA-rmAw;yi ('he keeps for us'; Jer 5:24).
All these examples demonstrate the role that rhythm plays in
creating construct and construct-like forms. But none of these are equivalent
to true construct chains governed by the principle of non-intervention.
True Construct Chains Involve a Genitive
The true construct chain, particularly as it relates to the problem
in Ezekiel 38, is limited to the genitive relationship between nouns. It
is in this particular case that the principle of noninterventiom applies,
and it is this particular case that must be tested for exceptions, not
whether a construct form may be succeeded by something other than
an absolute nomen rectum.
60 Muller, Hebrew Syntax, 53.
PRICE: ROSH: AN ANCIENT LAND KNOWN TO EZEKIEL 79
There are several commonly known modifications of the principle
of nonintervention that must not be regarded as violations.
(1) A construct may follow another construct when there is a
series of genitives.61 This forms a construct chain of more than two
links, the last of which is an absolute. The principle of noninterven-
tion then applies to the entire series.
(2) The definite article may precede the absolute nomen rectum.
Since it is a prepositive, it is regarded as part of the nomen rectum.
(3) The locative He may follow the first construct as a post-
positive case marker.62 It is regarded as part of the nomen regens.
(4) Although it is not common, the construct may receive a pro-
noun suffix.63 Usually the pronoun is attached to the nomen rectum
even though it modifies the construct;64 but where sense or style
requires, the construct may receive the suffix. Since it is a suffix, it
must be regarded as part of the nomen regens not as an intervening
(5) Although it is not common, a negative may precede the
nomen rectum. The negative is usually connected to the nomen rectum
by a maqqeph, making it the equivalent of another construct, or a
part of the nomen rectum. The negative must precede the word it
negates and, like another construct, it is a legitimate modification to
the principle of nonintervention. Examples of this construction are
found in Isa 31:8 (wyxi-xOl br,H, / 'a sword not of man'), Isa 31:8
(MdAxA-xOl br,H, / 'a sword not of mankind'), 2 Sam 23:4 (tObfA-xlo rq,Bo
'a morning of no clouds'), and Isa 14:6 (hrAsA yTil;Bi tKama / 'a stroke of
None of these modifications of the principle of nonintervention
corresponds to the grammar of Ezek 38:2, 3; 39:1. None accounts for
an attributive adjective intervening between a construct noun and its
genitive nomen rectum.
True Construct Chains are Seldom if Ever Broken
Now the question to be answered is this: have any clear examples
been found of a departure from the principle of nonintervention? If
so, are the exceptions sufficient to justify considering j`w,m, wxro xyWin;
lBAtuv; to be a broken construct chain? The Ezekiel case would consist
of an adjective attribute of the nomen regens interposed between the
nomen regens and the nomen rectum.
61 GKC. 128a.
63 GKC. 128d.
64 Muller. Hebrew Syntax, 54.
80 GRACE THEOLOGICAL JOURNAL
In general the syntax of biblical Hebrew is structurally consis-
tent; there are relatively few structural discontinuities. Where the
syntax exhibits discontinuity, it is for emphasis, clarification, the
avoidance of ambiguity, or due to an author's style. Occasionally a
discontinuity is created by ellipsis. But legitimate discontinuities are
purposeful and meaningful, not accidental and enigmatic.
If there is proof that the principle of nonintervention may not
always be followed (as some believe to be true in Ezek 38:2), the
proof must consist of clear, unambiguous examples from biblical
Hebrew. The examples cannot be created by speculative emendation;
they must have interventions similar to Ezek 38:2; and they cannot be
examples of the admissible modifications of the principle previously
mentioned. It should be expected that an example would exhibit a
case where the discontinuity provides clarification, emphasis, the
avoidance of ambiguity, or evidence of stylistic purpose. It is not
expected that the discontinuity should be explained as a grammatical
blunder that contributes to confusion.
Numerous examples of possible broken construct chains have
been listed by Gesenius, Dahood, and Freedman. Yet none of them
qualify as an unambiguous precedent that proves that true construct
chains may be broken.
Gesenius' Broken Construct Chains
Because Gesenius felt so strongly about the principle of non-
intervention, he was very reluctant to recognize any possible excep-
tion. He said:
As the fundamental rules are the necessary consequence not merely
of logical but more especially of rhythmical relations, . . . we must feel
the more hesitation in admitting examples in which genitives are sup-
posed to be loosely attached to forms other than the construct state.
Others have been more willing to accept broken construct chains,
but Gesenius' reluctance should serve as a warning against hastily
discovering supposed discontinuities in Hebrew syntax. Although he
did not regard these passages in Ezekiel as broken construct chains,
he did discuss certain problems related to the principle of noninter-
Intervening Pronoun Suffix. Gesenius listed several examples
of a pronoun suffix intervening between a construct and its nomen
65 GKC. 128b.
66 GKC. 130a-f.
PRICE: ROSH: AN ANCIENT LAND KNOWN TO EZEKIEL 81
rectum. He tried to explain away the noted cases as textual corrup-
tions or by emendations. Actually, according to previous discussion,
such pronoun suffixes are to be regarded as part of the nomen regens
and not a violation of the principle of nonintervention. The use of a
pronoun suffix with the nomen regens is uncommon, but required at
times to avoid ambiguity.
Special Case for the Construct of lKo. Gesenius67 recorded a
special problem with the word lKo (also noted in BDB). Three times its
construct seems to have a word interposed between it and its genitive,
a structure which BDB marks as very anomalous:
2 Sam 1:9 yBi ywip;na dOf-lKa yKi
Job 27:3 yBi ytimAw;ni dOf-lKA yKi
Hos 14:3 NOfA xWA.Ti-lKA
Gesenius suggests that lKo must be regarded as adverbial in these cases
in the sense of "wholly." That is a good suggestion; however, the
possibility remains that they may be broken construct chains. In any
case, lKo is a quantifier, not a noun, and as such it has unique rules of
syntax that vary somewhat from those of nouns. It is doubtful that
this special case can be used to justify an intervening adjective in
Intervening Adjectives. Gesenius listed several other examples
of possible broken construct chains.68 He listed Isa 28:1 as a possible
case of an intervening adjective. The text reads Nyiya ymeUlHE MynimAw; xyGe
'the rich valley of those overcome with wine' (RSV). It is understood
by some that MynimAw; is an adjective attribute of the nomen regens xyGe
intervening between it and ymeUlHE. This seems to be the way it was
understood by KJV, ASV, RSV, and NASB. However, MynimAw; is not
an adjective but a noun;69 and it is not in grammatical concord with
xyGe as expected for an adjective. The form may be the abstract plural
with the meaning "fatness," "richness," in which case "the valley of
richness" is a proper way of expressing "rich valley." Thus, it is
proper to understand "the valley of the richness of those overcome
with wine" as the equivalent of "the rich valley of those. . . ." How-
ever, to express this equivalent construction would require MynimAw; to be
in the construct state (and thus not violate the principle of noninter-
67 GKC. 128e.
68 GKC. 128c.
69 BDB. 1032.
82 GRACE THEOLOGICAL JOURNAL
Because MynimAw; is not construct, and because of the disjunctive
accent separating the two halves of the expression, many translators
regard the halves as not syntactically related (NIV, NAB, TEV,
NKJV). This seems to be the better choice since it follows conven-
tional grammar. Although it is possible to regard the example as a
broken chain, the grammar and accents are against it. Thus, it cannot
be used as an unambiguous precedent.
Gesenius also listed Isa 32:13 as a possible case of an intervening
adjective. The text reads hz,yli.fa hyAr;qi WOWmA yTeBA lKA / 'all the joyous
houses of the jubilant city.' It is possible to regard WOWmA as an adjec-
tive attribute of yTeBA interposed between it and its nomen rectum hyAr;qi.
But, as in the previous example, WOWmA is a noun meaning "exulta-
tion",70 and a disjunctive accent separates the halves of the expres-
sion. Nearly all translators understand the halves to be syntactically
unrelated, and to have an elided words between them (KJV, RSV,
ASV, NASB, NIV, TEV, NKJV), or to be appositives (NAB). It
seems to be wholly rejected as a broken construct chain.
Gesenius also listed Isa 28:16 as a possible case of an intervening
adjective. The text reads dsAUm traq;yi tn.aPi / 'a costly cornerstone of a
foundation.' It is possible to regard traq;yi as an adjective attribute of
tn.Pi interposed between it and its nomen rectum dsAUm. Although
Gesenius asserted that traq;pyi is a construct noun not an adjective,71 it is
classified as an adjective in his lexicon, in BDB and others. Yet it is
unusual for an attributive adjective to be in the construct state. The
text is problematical and cannot serve as an unambiguous precedent.
Gesenius also listed Ezek 6:11 as a possible case of an intervening
adjective. The text reads lxrAWyi tyBe tOfrA tObfEOT lKA / 'all the evil
abominations of the house of
an adjective attribute of tObfEOT interposed between it and its nomen
rectum lxErAW;y tyBe. Although' Gesenius asserted that tOfrA must be a
construct noun (evils) not an adjective, the form could be either an
adjective or a construct noun. However, since the construct noun
follows normal grammar and makes good sense, Gesenius should be
given the benefit of the doubt. Since the key word tOfrA is ambiguous,
this example cannot serve as an unambiguous precedent.
In summary, Gesenius' examples are problematical and ambigu-
ous. None can serve as proof that true construct chains may be broken
in biblical Hebrew.
70 BDB. 965.
71 GKC. 130f. n. 4.
PRICE: ROSH: AN ANCIENT LAND KNOWN TO EZEKIEL 83
Dahood's Broken Construct Chains
Dahood listed several possible examples of broken construct
chains in addition to those listed by Gesenius.72
Intervening prepositions. Dahood listed several examples of a
construct chain broken by a preposition. He has mistakenly identified
a construct governing a prepositional phrase as a construct chain. In
each case the relationship of the construct with the absolute is defined
by the preposition, not by the genitive. The meaning would be incom-
plete without the preposition. The reason for the construct form is
phonetic and rhythmical, not syntactical and logical.
In addition, 3 of the 5 examples are ambiguous--the forms are
not clearly constructs; they may properly be absolutes (Pss ; 10:1;
92:13). In the remaining two examples, Dahood revocalized the Maso-
retic text to create the example (Pss 74:12; 84:7). The Masoretic text
of Ps 84:7 does have a construct before a preposition, but it comes
under the above comment.
Intervening pronoun suffix. Dahood listed 17 examples of a
construct chain broken by a pronoun suffix. In six of the 17 examples,
Dahood revocalized the Masoretic text to create the case (Pss 16:8;
; 35:16; 88:16; 102:24-25; 140:10). In Ps 102:24-25 he made the
chain bridge the end of a verse, and in Ps 140:10 he made it bridge an
athnach--obvious departures from the Masoretic punctuation.
In three other cases his examples are construct participles govern-
ing an accusative pronoun suffix and an adverb: Ps 35:19 ("those who
are my enemies wrongfully"), Ps 35:19 ("those who hate me without
cause"), and Ps 38:20 ("those who hate me wrongfully"). These are
not examples of a construct governing a genitive nomen rectum. The
construct forms originated from rhythm and phoentics, not neces-
sarily because of grammar.
In six other cases the construct has a genitive pronoun suffix,
and the second word of the phrase is properly identified as an adverb
not an absolute noun. Construct nouns do not govern adverbs. The
noun takes the construct form because of the pronoun suffix. The
examples are not broken construct chains:
Ps 38:20 "My enemies are lively"
Ps 48:15 "This God is our God forever and ever'
Ps 61:5 "I will abide in your tabernacle forever"
Ps 66:7 "He rules by his power forever"
Ps 71:6 "My praise shall be continually of you"
Ps 105:4 "Seek his face forever"
72 Dahood, Psalms, 3:381-83.
84 GRACE THEOLOGICAL JOURNAL
In one other case his example is actually a proper name, "Mel-
chizedek" (Ps 110:4).
In all the above cases AV, NASB, NKJV, and NIV do not agree
with Dahood, but view them according to more conventional gram-
matical theory. The NIV regards the second word of Ps 38:20 as an
In only one case is there a possible broken construct chain: hTAxa
zfo-ysiHEma / 'You are my strong refuge' (Ps 71:7). This example comes
under the permissible variations of the principle of nonintervention,
which is not properly a violation. The pronoun cannot be attached to
the nomen rectum without changing the sense. The construct state of
the nomen regens has been carefully preserved in the Masoretic text
by the absence of a principle accent, and by the maqqeph, even
though there is a pronoun suffix.
Intervening emphatic yKi. Dahood listed six examples of con-
struct chains allegedly broken by an emphatic yKi. None of the exam-
ples were recognized by the Masoretes as the emphatic yKi. In each
case Dahood emended the Masoretic presentation of the text to create
the example--always by adding a space between consonants and, in
some cases, by changing the vowels. All of the alleged examples are
properly identified as pronoun suffixes of direct address followed by
a vocative, not a genitive. None are unambiguous examples of broken
Intervening enclitic mem. Dahood listed 23 examples of con-
struct chains with intervening enclitic mem. However all of these
examples involved revocalizing the Masoretic text to create the ex-
amples. Such revocalization is not strong evidence to demonstrate
that an enclitic mem actually breaks the construct chain in biblical
Intervening vocative. Dahood listed one example of a construct
chain with an intervening vocative, Ps 145:7, which reads j~b;UF-bra rk,ze /
'the record, O Master, of your goodness'. bra is regarded as the
intervening vocative. However, Dahood emended the Masoretic mark-
ing by omitting the maqqeph between bra and j~b;UF, and by ignoring
the lack of an accent on br, both of which identify bra as a construct
form. As a construct noun, bra is a member of an unbroken construct
chain that is grammatically and semantically correct. The phrase is
literally translated "the memory of the greatness of Your goodness,"
or "the memory of Your great goodness" (NKJV). Dahood's revo-
calization does not convincingly demonstrate that vocatives actually
break construct chains in Biblical Hebrew.
In summary, Dahood did not list one example of an unambigu-
ous broken construct chain; all his examples involved revocalizations,
PRICE: ROSH: AN ANCIENT LAND KNOWN TO EZEKIEL 85
ambiguous forms, or construct forms originating because of phonetics
and rhythm rather than from a grammatical genitive relationship.
Not one involves an intervening adjective and not one qualifies as a
precedent for regarding wxro as an adjective in Ezek 38:2, 3; 39:1.
Furthermore, Dahood did not list these Ezekiel passages as examples
of broken construct chains.
Freedman's Broken Construct Chains
David Noel Freedman attempted to add more examples of broken
construct chains to those listed by Dahood and Gesenius.73
Intervening enclitic mem. Freedman listed several additional
examples of an intervening enclitic mem. All involved revocalizing
the Masoretic text; none convincingly demonstrates that an enclitic
mem actually breaks a construct chain in Biblical Hebrew.
Intervening clause. Freedman proposed that there is a clause
breaking a construct chain in Isa 10:5 which reads MDAyAB; xUh-hF.,maU
ymif;za. In this rather difficult construction, he proposed that the clause
MdAyAB; xUh / 'he is in their hand' breaks the construct chain. hF.,ma
ymif;za / 'the staff of . . . my fury'. In doing so he emended the absolute
noun hF.,ma to its construct form hFema and emended the word MdAyAB; to
MydiyAB; ('in my hand') with an enclitic mem. His translation is "the staff
of my fury is he in my hand." His emendations created the broken
construct chain. The Hebrew is difficult, but it can be understood
without emending the Masoretic text. the KJV has "and the staff in
their hand is my indignation," the NKJV has "and the staff in whose
hand is My indignation," and the NASB has "and the staff in whose
hands is My indignation." All these make tolerable sense following
the Masoretic vocalization. Freedman's speculative revocalization does
not provide strong evidence to demonstrate that Isa 10:5 is an instance
where a clause really breaks a construct chain.
Intervening pronoun suffix. Freedman listed Hab 3:8 as an
example of an intervening pronoun suffix: hfAUwy; j~yt,boK;r;ma / 'your
chariots of salvation.' This is a case that comes under the permissible
variations of the principle of nonintervention previously mentioned.
In this case the pronoun cannot be attached to "salvation" without
Intervening sign of the direct object. Freedman listed Hab 3:13b
as an example of the sign of the direct object breaking a construct
chain. In the phrase j~H,ywim; tx, fwayel; / 'for salvation with Your
73 Freedman, "The Broken Construct Chain," 534-36.
86 GRACE THEOLOGICAL JOURNAL
Anointed,' Freedman proposed that the construct chain is . . . fwaye
j~H,ywim;. ‘the salvation of . . . your anointed' in parallel with the pre-
ceding line "for the salvation of your people." The tx, would then
break the construct chain. However the form of the word fwaye is
ambiguous, either absolute or construct, and the word tx, may be
either the sign of the direct object or the preposition "with." The
translation, following a more conventional grammar, would be "for
salvation with Thy Anointed" (NKJV). The absolute noun governing
a prepositional phrase makes sense. Though the line lacks poetic
parallelism, such progressive structure is not uncommon. Freedman's
ambiguous speculation does not convincingly demonstrate that the
sign of the direct object really breaks a construct chain in Biblical Hebrew.
Intervening adverb. Freedman listed Hab 3:13c as an example
of an adverb breaking a construct chain. In the
fwArA / 'You struck the head from the house of the wicked,' Freedman
proposed that the construct chain is fwArA . . . wxro / 'the head of the
. . . wicked one' and that the word tyBemi should be emended to tyiBami
(inward), so that the clause is translated "You crushed the head of the
wicked one inwards." But wxro is an ambiguous form, either absolute
or construct, and the revocalization is speculation based on poetic
parallelism. The Masoretic pointing of the text makes sense. This
revocalization does not convincingly demonstrate that an adverbial
phrase really breaks construct chains in Biblical Hebrew.
He also listed Ezek 39:11 as an example of an intervening adverb.
The text reads lxerAW;yiB; rb,q, Mw-MOqm; / 'a place there of graves in
place of . . . burial' with the adverb MwA ('there') intervening. The
translation would be "a place of burial there
has the concrete meaning "grave, sepulchre, burial place”74 rather
than the abstract sense of "burial" which is rendered by the Hebrew
hrAUbq;. Regarding rb,q, as a genitive results in an awkward, unnatural
sense ("a place of a grave there"). The last two words, lx;rAW;yiB; rb,q,,
function more naturally as an appositive ("a place there, a burial
tics created by the close rhythmical relationship between it and the
following adverb MwA. Nevertheless, the example remains a possible
broken construct chain; but, because it makes sense in the more
conventional view (i.e., as an appositive), it remains ambiguous and
does not provide a precedent for demonstrating that an adverb really
breaks a construct chain in Biblical Hebrew.
74 BDB, 868.
PRICE: ROSH: AN ANCIENT LAND KNOWN TO EZEKIEL 87
Intervening verb. Freedman listed Hos 14:3 as an example of a
verb breaking a construct chain. This example was previously noted
by Gesenius,75 and was discussed in a previous section.
He also listed Hos 6:9 as an example which reads hmAk;w, UHc.;rav; j`r,D, /
'they murder on the way to Shechem'. He proposed that the con-
struct chain is HmAk;w, . . . j`r,D,, and that the verb intervenes. There are
two reasons why this is ambiguous: (1) the form of j`r,D, is ambiguous,
being either absolute or construct; and (2) the word hmAk;w, has the
locative he and is the equivalent of Mk,w,l;. Thus the translation is "the
way to Shechem," and is not to be confused with "the way of
Shechem." Because the example is ambiguous, it does not provide
He also listed Hos 8:2 as an example which reads yhalox< UqfEz;yi yli
lj~UnfEday; / '
proposed that the word yhalox< be revocalized to yhelox< to produce the
broken chain lx;rAW;yi
. . . yhlox< / 'the God of . . .
intervening. The translation would be "O God of Israel, we know
you." This again involves revocalization of the Masoretic text. Freed-
man seems to exaggerate the change in number (from "my" to "we"),
a common phenomenon in poetry. This instance does not provide
strong evidence for demonstrating that a verb really breaks a con-
struct chain in biblical Hebrew.
In summary, Freedman did not list one example of an unam-
biguous broken construct chain; all his examples involved unneces-
sary revocalization, ambiguous forms, or construct forms originating
because of phonetics and rhythm rather than from a strictly gram-
matical genitive relationship. None involved an intervening adjective,
nor do any qualify as a precedent for regarding wxro, as an adjective in
Ezek 38:2, 3 and 39:1. Furthermore, Freedman did not list these
passages in Ezekiel as examples of broken construct chains.
No Proof Found for Broken Construct Chains
None of the examples furnished by Gesenius, Dahood, or Freed-
man are unambiguous broken construct chains; all the examples
involve unnecessary revocalization, ambiguous forms, or construct
forms originating because of phonetics and rhythm rather than from
a grammatical genitive relationship. All the possible cases of interven-
ing attributive adjectives are problematical. Not one example qualifies
as an unambiguous precedent for regarding wxro, as an adjective in
Ezek 38:2, 3; 39:1. Furthermore, none of the three scholars listed
75 GKC. 128c.
88 GRACE THEOLOGICAL JOURNAL
these Ezekiel passages as examples of a broken construct chain. Con-
sequently, it must be concluded that the existence of broken construct
chains is speculative apart from the previously mentioned normal
modifications of the principle of nonintervention. Alleged broken
construct chains provide no support for breaking the principle of
nonintervention in Ezekiel.
The origin of the translation "chief prince of Meshech and Tubal"
is traced to the Latin Vulgate. The early translators of the English
Bible were quite dependent on the Latin Version for help in trans-
lating difficult passages. They evidently followed Jerome in Ezek 38:2,
Some have supposed that the Aramaic Targum may have been
the source for interpreting wxro as an adjective. The Targum reads
lbAtuv; j`w,m, wyre bra, where bra is the equivalent of Hebrew xyWin; and wxre
(=wxre ) is the equivalent of Hebrew wxro . But Aramaic has the same
syntactic conventions for construct chains as Hebrew, so the same
arguments that favor wxro as a name in Hebrew favor wxre as a name
in Aramaic. Therefore, the Aramaic does not support regarding wxro
as an adjective, although those who do not take into account the
difference in pronunciation may erroneously think
Jerome may have thought).
Evidently by the second century A.D. the knowledge of the ancient
was in such common use as "head" or
to interpret wxro as an adjective, contrary to the LXX and nor-
mal grammatical conventions. Jerome followed the precedent set by
by removing the name from the Latin Bible.
By the sixteenth century A.D. ancient Rosh was completely un-
known in the West, so the early English translators of the Bible were
influenced by the Latin Vulgate to violate normal Hebrew grammar
in their translation of Ezekiel 38-39. Once the precedent was set in
English, it was perpetuated in all subsequent English Versions until
this century when some modern versions have taken exception. This
ancient erroneous precedent should not be perpetuated.
It has been demonstrated that Rosh was a well-known place in
antiquity as evidenced by numerous and varied references in the
ancient literature. It has also been demonstrated that an adjective
intervening between a construct noun and its nomen rectum is highly
PRICE: ROSH: AN ANCIENT LAND KNOWN TO EZEKIEL 89
improbable, there being no unambiguous example of such in the
Hebrew Bible. Furthermore, it has been demonstrated that regarding
wxro as a name is in harmony with normal Hebrew grammar and
syntax. It is concluded that wxro cannot be an adjective in Ezekiel
38-39, but must be a name. Therefore, the only appropriate transla-
tion of the phrase in Ezek 38:2, 3, and 39:1 is "prince of Rosh,
Meshech, and Tubal."
This material is cited with gracious permission from:
Grace Theological Seminary
Please report any errors to Ted Hildebrandt at: firstname.lastname@example.org