Skip to main content

Full text of ""Man by Man," Joshua 7, 17"

See other formats


Early Journal Content on JSTOR, Free to Anyone in the World 

This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in 
the world by JSTOR. 

Known as the Early Journal Content, this set of works include research articles, news, letters, and other 
writings published in more than 200 of the oldest leading academic journals. The works date from the 
mid-seventeenth to the early twentieth centuries. 

We encourage people to read and share the Early Journal Content openly and to tell others that this 
resource exists. People may post this content online or redistribute in any way for non-commercial 

Read more about Early Journal Content at 
journal-content . 

JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people 
discover, use, and build upon a wide range of content through a powerful research and teaching 
platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit 
organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please 


By Max L. Margous, Dropsie College 

In his Lectures on "The Polity of the Ancient He- 
brews" (JQR., New Series, III, 1 ff.) Judge Sulzberger 
has occasion to cite the "classical text" in Josh. 7, 16-18 
bearing on the subject of the Israelite army organization 
and, after giving his own version, aptly remarks : "The text 
is slightly defective, but a careful reading of it justifies this 

The text in question is indeed faulty. The two English 
Versions (Authorized, Revised) reproduce the Hebrew 
text as commonly printed. As the differences between the 
two Versions are only verbal we may transcribe here the 
Revised. "(16) So Joshua rose up early in the morning, 
and brought Israel near by their tribes; and the tribe of 
Judah was taken: (17) and he brought near the family of 
Judah ; and he took the family of the Zerahites : and he 
brought near the family of the Zerahites man by man; and 
Zabdi was taken: (18) and he brought near his house- 
hold man by man; and Achan, the son of Carmi, the son 
of Zabdi, the son of Zerah, of the tribe of Judah, was 

One need only turn to verse 14 to realize the two 
points wherein the received text is at fault. Verse 14 
reads in the Revised Version (which again differs only 
verbally from the Authorized) : "In the morning therefore 



ye shall be brought near by your tribes: and it shall be, 
that the tribe which the Lord taketh shall come near by 
families; and the family which the Lord taketh shall come 
near by households; and the household which the Lord 
shall take shall come near man by man." Accordingly, the 
divisions are: the tribe (MB>), the family (nnse>»), the 
household (JV3)> the man (laJ). We should therefore ex- 
pect verse 17 to have read: "And he brought near the tribe 
of Judah by families ; and he took the family of the Zera- 
hites : and he brought near the family of the Zerahites house- 
hold by household; and Zabdi was taken." Such, in sub- 
stance, is the version acceptetd by the learned Judge, and 
he is sustained by modern commentators. 

When we approach the Hebrew of verse 17, the sub- 
stitution of dtqS (household by household) for vraib (man 
by man) is a self-evident remedy. As for the beginning 
of the verse, all that is required is a change in the pointing 
of the first nnsK'D: in the received text it is pointed nnsE>D 
(family, in the singular), but we should point the word 
nftsE'p (families, in the plural). Accordingly, what the 
author intended was: "And he brought near the families 
of Judah" — which is indeed an abbreviated expression for 
the phraseology postulated on the basis of verse 14: "And 
he brought near the tribe of Judah by families." 

The two forms of the text, the received and the 
emended, may now be placed in juxtaposition: 

Received Text Emended Text 

bmw ris rip^i npaa yma* Dam 16 

a-ipV° nnrv Datr 12^1 visaed 

nx la^i miiT nnac'D nx nx la^i mm* nhae><? n« 

n« 3ipV* Trim nnse'o 


onaib imrn nnsE>D dtq^ "mm nnsj?D 

irra n« mpv 8 nnt nabi 
•'Bia p pj> na^i onaib 
niiiT ntaob mr p nat p 

The Emended Text is printed as above by Bennett in 
Haupt's Bible (1895). In his Notes (page 26) we read: 
"nnsEto, with some MSS. and (g (Septuagint) 3 (Vulgate) ; 
DTD^, with 3 (Vulgate) <g (Peshitta) and some MSS., 
Dillm(ann), &c." He also remarks: "(& (Septuagint) A 
(omits) ijb and m n« in 18; which Hollenberg (Progr. 
13) is inclined to follow." 

As will be noted, the authority of the Septuagint is 
adduced for the first change ( nftaK'D ) , but not for the 
second (DTD 1 ?), obviously because, with Hollenberg, he 
finds verse tyb to have been wanting in the Septuagint, 
though he is not ready to follow Hollenberg in the as- 
sumption that the second half of verse 17 is a late addition 
in our Hebrew text which should be excised. Driver in 
Kittel's Bible (second edition, 1909) at least notes that ten 
manuscripts of the Greek read "by households"; but that 
is still far from saying what the original Greek translator 
read and wrote. 

The aim of the present paper is to show that the 
verse-half in question was present in the Septuagint and 
that furthermore the translator read in his Hebrew text 

But before the argument is presented in substantiation 
of my double contention it may be well to ascertain upon 
what authority the Received Text rests or who are its wit- 


Both Baer and Ginsburg print the two obvious errors: 
nnsE'p and Dnaib. Baer has nothing to say on the former; 
with regard to the latter he remarks : "The Soncino edition 
prints faultily DTqS ." The Soncino edition of the Prophets 
(1485-1486) is the first print of that part of the Script- 
ures. According to Ginsburg, the second print of the 
Prophets in the first print of the entire Bible (Soncino 
1488) has likewise QTab. The other books (DnriK DnBD) 
which, according to Ginsburg, read nfratsto and DTI3? are 
certain manuscripts enumerated by Kennicott and De- 
Rossi. 1 De-Rossi adds that "by households" is found not 
only in the Vulgate and the Syriac Peshitta, but also in the 
Aldine edition of the Septuagint, on the authority of 
Fischer, Prolus, de versionibus Graecis 8, Leipzig 1772, 
page 156, who favors this reading, though Masius accepts 
the current reading "man by man." I have Fischer before 
me; he adds that the reading is apparently supported by 
verse 14. 

When prints and manuscripts differ, recourse must 
be had to (a) Masoretic evidence and (b) the testimony of 
the Jewish mediaeval commentators. 

According to the Masorah on Num. 3, 23/ four in- 
stances of nnstyp (singular) occur in Leviticus and Joshua 
(or Genesis, Exodus, Leviticus, and Joshua), to wit: Levit. 
25, 47 and the three in Josh. 7, 17. Hence the pointing 
nnQE'p at the head of verse 17 is recognized by the Masorah 
and the reading niriBB>0 must be pronounced contrary to the 

1 mnSt£*Q, so with vowel letter indicating the plural, Kenn. 182, 250; 548 
and 598 in the margin; 155 first hand, i. e. under the subsequent correction; 
DTD? 253, 257, 260, 524, 583 margo, 366 over an erasure, D'1227 DTD? 
— both readings together — 531; De-Rossi 305 first hand, 20, 174, 663 over an 
erasure, 716 in the margin. 

2 See Ginsburg, The Massorah, letter D, No. 846, compare also No. 847a. 


Masorah. — A list in Ginsburg, letter b, No. 14, consists of 
words beginning in b each of which is found only once in 
Scripture (Jinnim nb ini in b3\ b prwm ); among them 
figures D'riB^ of Josh. 7, 14. Hence the reading &mb in 
verse 17 is ruled out by the Masorah and D ,- a:6 there is 
substantiated as Masoretic. 

As for the Jewish commentators, Kimhi's exposition 
makes it plain that he read nnaeto and Dnaai). According to 
him, "the family of Judah" is the same as "the tribe of 
Judah"; supply ninat^ "by families." The first Dnaib 
he paraphrases by rinse-en Tia W) wvb "man by man who 
is the head of a household," while the second tDM means 
Ivan 'box the individual members of the household. 

But we have earlier evidence for the reading Bniib in 
verse 17. In the first place the Targum has it ( snaa^; 
the Targum also pointed the first nnst^D in the singular, 
rvjnr, Praetorius, Das Targum su Josua, 1899). But we 
may ascend still higher. We find it in the Septuagint re- 
cension of Origen which dates from the middle of the 
third century of our era. 

Origen's recension with its critical signs* is extant in 
the Syriac translation of Paul of Telia (616-7). I* was 
published by Lagarde (1892) from a British Museum 
manuscript and the critical signs where faultily placed or 
entirely wanting were restored from the commentary of 
Masius" who had before him another manuscript which has 
since disappeared. 

* A plus found in the Hebrew but wanting in the Greek text which 
was subjected to revision was introduced by an asterisk, *, and closed by a 
metobelus, : or V. 

* Andrew Du Maes, died 1573. His work on Joshua appeared in 1574; 
it is excerpted in the second volume of the Critici Sam. 


The Syriac text of verses 17 and 18 reads as follows 
(to obviate typographical difficulties it is given in Hebrew 
type) : mpnNi> mn Nonit: •'innKi | xiim * KDftio tn mpntn 
jam r6n «rpp mpnKi hit ■'•tnnsi * ttfaa tr\2i Onn ndhid 
Krme> jo * mn mn nan ma wan mi nay •'lnrwi | k-dj 
I trnnn. 

The words enclosed within < > are missing in the 
manuscript (and are omittted by Lagarde). Whether the 
omission occurred at some stage in the Syriac transmission 
of the text or was present in the Greek archetype upon 
which the translation was based is immaterial. It is sim- 
ply a case of aberration of the scribe's eye from the first 
mn (or its Greek equivalent) to the second (homoioteleu- 

The Greek for the Syriac as given above is extant in 
the manuscripts underlying Lagarde's edition of the Greek 
Old Testament (1883; one of those manuscripts served as 
the basis of the first printed edition of the Septuagint in 
the Complutensian Polyglot). There is a gap right at the 
start the nature of which escaped Lagarde. He prints in 
brackets mi irpoavx^n Kara fy/iovc But one more word should 
be added: umSa. The scribe's eye wandered from wvda with 
which verse 16 closes to wvda of verse 17. Hence the full 

text reads '. <jiai ■Kpoatixdrj Kara dq/iovf tovda^> mi aveSeix^V typoc ° 
£apaei Km TrpooijxOy Sij/ioc o £apaei Kara avipa mt avedetxOy fa/3<5£i kcu 
irpoarix^V o ot/cof avrov Kara avdpa Km avedeix^V ax a P m0 C X a Pl ttL vlov 
^afideL viov i^apa ttjq §vfai$ covda. 

As will be observed, the critical signs are missing. The 
two texts supply each other's deficiencies admirably. 
Origen accordingly wrote : tun -Kpoarix^n K <"' a fiyftove # iovSa ■. km 

avedeixfo} <%«>f o C,apau Km npooiJxQ'n <%/of o i^apaei Kara avdpa.{s) * Km 


aveSuxfty fa/3(i« mi npoarjxOri oim<; avTov Kara avSpa : nai avedeix^l 
axap wos x a PP ei mov Z a M el mov £ a P a * TWQ <j>v7ir/c lovSa : 

It is obvious that Origen read in his Hebrew of verse 

Ij D 1 *!^? tara awf/jaff), comp. D , "133;> mra avSpa verse l8. 

The error in the Hebrew is as old as Origen. 

We may take it for granted that all such manuscripts 
of the Septuagint as exhibit the reading mra avSpa^ in verse 
17 were influenced by Origen's recension. 

Two such manuscripts are F (Ambrosianus) and n (a 
Mount Athos manuscript hitherto uncollated of which pho- 
tographs have been secured by the Dropsie College. It is 
related to some twelve manuscripts grouping around the 
Catena Nicephori, but has readings of its own). They di- 
verge at the end, but otherwise present a text substantially 

identical : Fn mi ■Kpoarix^V Kara Srjfiovg mvda mi evedetxOl Sr/fio^ £apat 
(F reads £api) mi rrpoar/xSr/ 8q[to( o i^apai mr(a") avdpac ' mi evedux&l 
(omitted in F) oikoq £a/ij3pt Kat izpoarjx^ oikoc (F adds avrov) 
/cai-foQ avdpa(c~) • mi evedeixOq &X a P 

n vio( (afippi viov l,apa tov x a PP l mov avrov 
F vw( x a Pf el vwv fc/tfipet viov (apa 
Fn Trie fyvktjs tovSa. 

Opposite the portion which is enclosed above between 
two upper points there is found on the margin of n an 
asterisk (*). To the sign -+- over fr/ippi corresponds in the 
same manuscript on the margin x a PV-i- by the first hand. 

Both texts have Origen's additions. Both have the 
telltale mT[a) avdpas in verse 17. Both however, write 
Sapppi in the place of (afidei. In this they revert to the form 
of the name in the Greek text antedating Origen, which 
form goes back to a Hebrew variant "nor for the received 
"HST. In I Chron. 2, 6 Zimri is the first-born of Zerah, the 


son of Judah. As for the divergence at the end, F, bar- 
ring the form t,wPph coincides with Origen, while n text 
(not: margin) in common with its group members curtails 
the pedigree by one and then adds by way of rectification 
that the culprit was not really the son of Zimri, but his son 
Carmi's son. We shall meet below in one other manuscript 
belonging to a different group with the same ending. 

The Origen recension underlies furthermore the 
uncials A (Alexandrinus) and its recently discovered con- 
gener e (Washington Codex). Of the two A is intact, 
while the scribe of e (or a predecessor of his) made an 
omission by homoioteleuton exactly at the place where the 
scribe of the Syriac has erred. Restore e as follows: 

kcu ■Kpocrix8rj Kara Srifiovq Km svedeixOtf Srjfioq C,apai <^k<u irpoa^x^V fyftoc 

fapar> /car avdpag nai evsdeixSv o-X a P mo C fapfiP 1 viov C,apa with which 
Compare A : mi irpwrtixfy lcaTa <%*»»£ «ai evedeix9y fyu-OQ iaptei mi 
irpoarixOij Srifio^ t,apiu /car avdpag Km eveSeixSi] ax& (i. e. ax av J 
»)(0f fa/i/Hpi viov ^apa. 

The differences are trifling (note «.x«- v in A = pjj as in our 
Hebrew text; the Septuagint writes <*;w = i35?, comp. 

1 Chron. 2, 7) . Both have the curtailed pedigree at the end 
but without the postscript which we find in n. 

The manner in which the text of the two uncials was 
constructed is plain. It is an eclectic text. A pre-Origen 
text was followed in the ending (note the curtailed pedigree 
and the form Zambri) ; otherwise a transcript was made 
of Origen's revised text, omitting the additions marked by 
an asterisk but retaining a plus where Origen failed to 
mark it by signs (perhaps the signs were missing in the 
copy immediately before the scribe which may have been 
itself a transcript of Origen's text). 


On the basis of the present case it might be argued that 
Ae constitute the text which Origen made the basis of his 
revision. Hence the second cV»c o ^apau was unmarked 
from the start, since it was found in that text. But when 
all the evidence derived from a critical study of Ae 
throughout the book of Joshua is brought to bear upon the 
problem the inference is unavoidable that the two uncials 
have made use of Origen and not the reverse. Naturally 
Ae retained Origen's /car avfipag in verse 17. 

The text which Origen made the basis of his revision 
is none other than the famous Vaticanus (B) or a text 
closely related to it. Verses 16-18 read in B : /cat apdpiaev 

iTjaovc; Kai irpoaqyayev tov \aov Kara <jtv"kag nai evedeixOr/ tj tyvkri tovda icai 
irpoarix6rj Kara drip,ovq km evedeixBy Srjjioc; t^apaei nai irpoaqx^H kcit avdpa 
/cat evedeix^y &X a P vl0 C £ap.flpei viov C,apa. 

Origen, in dealing with this text, proceeded upon the 
assumption that Kara avdpa goes with what immediately pre- 
cedes it and covers ttnaab in verse 17. and that the trans- 
lator omitted tmn;6 urn m mp^i Ha? "lain. He accordingly 
filled up the gap just as he made good the other omissions. 
As was his wont, he did not translate the Hebrew afresh, 
but made use for the parts missing of one or all of the three 
later Greek translators (Aquila, Symmachus, Theodotion) 
dating from the middle of the second post-Christian cen- 
tury. In all such cases he transcribed them faithfully 
without regard to the disharmony thus produced by weld- 
ing together incongruous versions. The incongruity in the 
present case will become manifest as this investigation pro- 

It is, however, possible that Origen recognized that the 
gap occurred before and not after Kara avdpa. In placing the 
signs, Lagarde relied too confidently on Masius; it is not 


always easy to tell whether the latter did not handle the 
signs of his Syriac manuscript with considerable freedom 
and here and there operated with conjecture. The British 
Museum manuscript errs when it places the asterisk too 
far down, but, I believe, is altogether right in putting the 
metobelus after avrov. Accordingly, the asterisk should be 
moved up, against Masius, in front of mm avdpaffj. The 
Syriac certainly read mra avfyag in the plural ; the reading is 
substantiated by AFen. It is true, F reads the plural 
also below; but n has correctly the singular in agreement 
with B. In verse 14 all, with exception of nsz, have the 
singular. It would seem, therefore, that the singular ac- 
cords with the style of the original Septuagint, while the 
plural which is a bit more literal squares with the manner 
of the source from which Origen supplied the omission. 
If this be so, the error D'HnjS in verse 17 is shown to be as 
old as the times of Akiba in which period the three princi- 
pal Greek translators after the Septuagint flourished. 
Origen's text in verses 17 and 18 will then have read as 

follows : nai ■KpoatixBri mra <%/M>vf * mvda : mi aveSeixdv <%/"<>£ ° 
£apaei mi Kpoarixdri Srjjio^ o i^apasi # Kara avSpag mi aveSeixSv ^a/3dei mi 
Trpoatixdt) oiKog avrov : Kara avSpa mi avedeixdv a X a P vl0 ( X a Pf iel vwv 
£a/3dei viov t^apa % rye <j>vlt)( lOvSa : 

But even as corrected Origen's text rests upon a me- 
chanical procedure. In saying this, we are not finding 
fault with Origen. His aim was to square the Greek with 
the Hebrew as he found it, the "Hebrew truth." But our 
task is a different one. If B, the text underlying Origen's 
revision, is faulty, and we are in accord with the Church 
Father on that score, it becomes our business to correct it. 
We are in a position to correct it by consulting other texts 
which are its congeners and by a careful study of the 


manner of translation which was adopted by the author of 
the oldest Greek version of our book. 

We are in a position to confront B with three texts 
which normally go with it. They are all more or less im- 
pure, contaminated with matter borrowed from Origen's 
revision. But the basis is a text very much like B. Into 
it they work Origenic additions, but in a manner of their 
own. The three texts are h (55 in Holmes-Parsons), the 
Ethiopic translation (codices FH), and the Coptic version 
recently published by Thompson. 

Of the three, h is a mutilated text, the scribe having 
been guilty of two omissions due to homoioteleuton. I 

SUpply the gaps within <> . h Kai ■KpocsrixP'O Kara 8r/fiovg mi 

eveSetxSv Sr/jiog t^apai <jtai npoarix^V <%«>? £apai^> Kar oikovq ' aveSeixOv 

oiKog ^aft/ipi <jtai TrpooqxOy oiko( £a/ippi Kara avSpa Kai avedeixOy <^X a P 

viog X a PP- 1 viov t,afifipC^>. 

Deplorable as the omissions are, the significant mr oikovc is 
intact. (In front of aveSuxfy the conjunction mi may have 
also dropped out.) 

The Greek underlying the Coptic (©) and Ethiopic 

(lE) read : Kai ■Kpoarixfiv Kara Srjfuyvg tovSa Kai evedetxdy if//ioe 00 
£apaei \C,api Jj) nai irpoatjx^V <%/«>f Kaposi \$api f£) Kar oimvg Kai 
evedeixOy oiKog ^afifipi Kai npoatjxOv oikoq ^afiflpi Kara avSpa nai eveSeixOy 
a X a P \flX av ffi) T X a V-P L mov £apa \-\-ttk fyvkrig mvSa ^j). 

Both contain the reading mr owovg. 

Now at last it becomes an easy matter to restore the 
archetype of B. It read in verses 16-18 as follows: 16 mi 

updpioev irjoovg Kai izpoorfyayev rov Tiaov Kara tpvXag Kai evedeix@y V Qv^V 
tovSa 1T Kai ■KpoarixOil Kara Srjfiavg Kai eveSuxBy typos t^apaei Kai TTpoar/x^V 
<Cjiar oiKovg Kai evedeixOq oimg t^afijipi 18 Kai irpoO7/x0v^> Kara avSpa Kai 
evedeixQy axap viog £a/*flpei viov C,apa. 



The omission, it is clearly seen, was due to homoiote- 
leuton. The archetype of B which in this case as in many 
others may be identified with the original text of the 
Septuagint had the portion now missing and in the form 
given above, as shown by the related texts h (ftf*. For all 
three (in h, as we have seen, the last two words dropped 
out along with what follows) have the portion, and they 
could not have derived it from Origen, who, as was shown 

above, wrote mr avdpag for kcit oikovc and o owog avrov for oacog 

fr/i/ipi, both in conformity with the received Hebrew text. 

Internal evidence serves to corroborate my conclusions 
based upon the external evidence of correct grouping. The 
translator had before him verses 16-18 in somewhat the fol- 
lowing form (to facilitate an understanding of my remarks 
below, I reproduce in the parallel column the Greek as re- 
stored) : 

mpi npan ynni m^ i« 

nnats>D nx 3-ip'i " mm* one* 
with nnsK>o nx na^i mini 
ovo^ imtn nnse>o n« 3*1 pi 

mpi 81 nor nn na^i 
onsii) nor ira m 

rnt }3 not }3 -lay iai>'i 

16 k<m updpiaev ir/aovg Kai npooriyayev 
tov "kaov Kara <j>v/ui( Kai eveSeix^V V 
tyvkri mvSa 1T Kai npocrix&l Kara 
dr/ftovg mi evedux&t) Srip,oq £apaei 
mi wpooqxSv mr oiKove Kai 

eveSeix^V oikoq £afif)pi 18 Kai irpo- 
ar/xSri Kara avSpa Kai sveSeixBi} 
axap vioq £afif3pei viov t^apa 

The translator's fondness for condensation reveals 
itself at the start. "Ip33 D3£"l is simply mi updpiaev "rose up 
early" (Origen added to npm "in the morning"). While he 
uses the active voice {^poariyayzv) for the first 3npjl (verse 
16), he proceeds in the sequel with the passive construction 
{■KpoarixOv, "was brought near"). By doing this he succeeds 
in getting rid of what to him seemed unnecessary repetition 


though such is quite in accordance with the style of He- 
brew writing. He reproduces the subject as in the Hebrew 
with each new sub-division introduced by the verb "was 
taken"; he leaves it to be supplied from the context when 
the sub-division "taken" is "brought near" for another sub- 
division. Hence beginning with verse 16b he writes: "and 
the tribe of Judah was taken; and it was brought near 
deme by deme; and the deme of the Zarahites was taken: 
and it was brought near household by household; and the 
household Zamri was taken: and it was brought near man 
by man; and Achar, ec, was eaken." It was left for sub- 
sequent texts (h (giE) to introduce the explicit subject. 
That this is not to be laid at the door of the translator is 
clear from the nature of the omission in B. For with h 
(£1; as a basis, the textual form after the omission (aberra- 
tion of the scribe's eye from the second irpoovxto to the third) 

WOUld have resulted in nai n-poavx^V ockoc Safifipi Kara avdpa. The 

omission of omog Sapppi in front of Kara avdpa in B proves my 
contention as to the translator's method of condensation. 

The translator pointed the first rinsCD in verse 17 as a" 
plural (nhsE'O ) . In accordance with his method of con- 
densation, the clause "and he brought near the demes of 
Judah" became "and it (sc. the tribe of Judah) was 
brought near by demes." 

The other variations between the Hebrew underlying 
the Septuagint and our received Hebrew text need not de- 
tain us. Of utmost importance is the reading DTO7 in verse 
17. As for the omission of Carmi at the end, the curtail- 
ment may and may not have been found in the Hebrew. 
Achan's father was omitted because he was of no moment, 
each warrior, N. N. son of N. N., ranging himself imme- 
diately under the household ( tvi ) which comprised a num- 


ber of families in the modern sense of the word. Mez (Die 
Bibel des Josephus, 1895, 5 f.) is wrong in arguing that 
Carmi is an interloper ; the Septuagint certainly read wa p 
in verse 1. Josephus with his axapo; . . . frfieduov kcuq (Ant. 
V 33) merely condensed the pedigree. In § 43 where he 
gives the execution of the divine order we meet with the 

fourfold order :<jn>An (t33£}> or nDD), (pparpm (nnaCD), avyyevua 

(JV3) , avnp (13J) . Though he may have written with a view 
to verse 14, we shall not go amiss in saying that Josephus, 
if he had the Septuagint open before him, read mr owovs, 
and, if he worked with the Hebrew text before him, he 
found in verse 17 DTub. 

The error in the Hebrew accordingly crept in between 
the times of Josephus and the times of Akiba. 

I will now adduce further proof for the correctness of 
my restoration of the archetype of B from still another 
quarter. A recension which is not Origen's but which a 
recent writer (Hautsch, Der Lukiantext des Oktateuch, 
1910) would stamp as that of the martyr Lucian (died 311 
or 312) is found in a group of manuscripts enumerated by 
me in the article "The K Text of Joshua" (AJSL., 
XXVIII (1911), 1 ff.). In the present state of our knowl- 
edge it is perhaps best to forego identifying it with any one 
of the three recensions signalized by Jerome and to speak 
of it as a nameless recension (see the lucid discussion by 
Professor G. F. Moore, in AJSL., XXIX (1912), 37 ff.). 
Be that as it may, the manuscripts constituting that recen- 
sion divide themselves in two sub-groups, a larger and a 
smaller. The former which is not a pure text, contamin- 
ated as it is with Origenic matter, is nevertheless useful as 
a means for correcting the errors of the smaller group. 
The larger sub-group may itself be subdivided into a larger 


group of four (u = 84, 1 = 134; p = 76; t = 74; u = 
ulpt) and a smaller group of three (f = 106; i = Cod. 
Gr. 609 of the Paris National Library; z = 44; F = fiz). 
z departs in verses 16-18 considerably from its congeners 
(fi) and must be studied separately. It reads: mi apdpioev 

a/aovc to irpoi Kai npoorjyaye tov %aov Kara fv^ag kcu evedeixOy V tyvfo) 
lovSa mi ■upooyx^V Kara dq/iovc. kcu evedeixOy o £apai dr/fioc. mi TrpocrjxOy 
8rinoQ £apai mr avdpag kcu evedeixdr/ ax a P vl0 C ^a/j.j3pi viov t^apa o 
tov Xtyf 11 vlov avrov. 

The text is mutilated ; but the omission is not to be put to 
the account of the scribe who appears to have transcribed 
a mutilated text. For, barring the end (it shares the post- 
script of n, see above), he gives a text in verses 17 and 
18 substantially agreeing with Ae. z accordingly steps out 
for our present purposes as a representative of its narrow- 
er or wider group. If we compare the text of i — it reads 

kcu ap6piaev ir/oove. to irfiui mi irpoar/yays tov %aov avrov Kara ftvXac. 
avrov Kai avtSux&V 1 tyvkri wvda kcu irpoar/x^W av Kara dr/fiovc lovSa kcu 
avedeixOq Stjuog o t^apa mi ixpoarryayov tov Srjjmv tov t^apa kot oikovc. 
kcu avedeixSy oiKoq £a/j./3pi viov (,apa Tqc_ fv?ir/g wvda — with that Of f 
kcu ap6piaev lyaovQ to irpm mi wpocrr/yaye tov Xaov Kara tyvTiac. avrov mi 
avsdeixOv V tyvhr] lovSa Kai TtpocrixOijoav mTa Sr/fiovg lovda mi avedeixOy 
Stj/iog o £apa Kai npoar/yayov tov Srjjmv £apa kut oikovc Kai avedeixOr/ 
oikoc ^ajifipi Kai npoarryayov tov oikov £afi/3pi KaT avSpa Kai avsdeixdy axap 
vioc Xo-Pt 11 mov Zcijijipi viov $apa rr/g yvXyg wvda 

we discover an omission due to homoioteleuton in i, the 
scribe having wandered from fa/i/3p« first to {apfipi third, f 
agrees in every respect (barring, of course, trifling vari- 
ants) with u, their common text therefore representing one 
archetype. The latter ranges itself with its mr oikovc on the 
side of h (&*&; it shares iovda (verse 16) with (J|» and rye </wfaic 
tovSa (end of verse 18) with f», both Origenic additions; 


but it has a number of traits of its own ; the plural wpocyxtkcav, 
£apa bis (in the genitive), mi npoanyayov (active construction) 
bis. With F it has the full pedigree at the end. 

As for the parallel group (r; the Old latin of the 
Codex Lugdunensis (?C) is an additional witness of this 
group), ro are practically identical. The text of verses 17 

and l8 reads as follows: mi wpoar/vexdr/aav Kara dr/fiovg irarpiuv 
mi avedeixdr/ dr//j.og o %apa Kai Trpoar/yayov tov Srjftov tov t^apa Kara avdpag 
Kai evedeix&i} {avedeixOr/ o) o (omitted in o) oimg ^aftfipi mi aveSeixOy 
axap viog XVI 11 mov ?<*/><* mov (aft/ipi mr oimvg rr/g <j>v/ir/g lovSa. S has 

suffered an omission through homoioteleuton, the scribe 
wandering from aveSeixfy second (as he read with o for 
evedeixSri r) to aveSux&n third. The manuscript goes its own 
way with the reading warpiag for Sr/ftovg warpiav ro and by cut- 
ting down the end so as to read Km avedeixOy a-x a P Vl °s x a ?y- 1 
viov (apa. irpoo>/xtiv<yav it shares with ufi. 
The Greek underlying f£ read : mi rrpoor/yayov mra Sr/ftovg Km 

evedsixSrj dr/ftog o £apa Km npoar/yayev tov Srjftov tov i^apa Kara avdpag Km 
evzSux&V oimg (.aftfipi mi ■KpoarjxQri o omog Kara avdpag mi evedeixOr/ 
axap viog x a Pf- 1 vlov ^a/tfipi viov t^apa. 

Disregarding unimportant variations, there remain to be 
singled out the transposition of (,apa and ^apfipi in the pedigree, 
the omission of mr omovg rrig fvXr/g wvda which it shares with s, 
and the plus mi irpoor/x6r/ oimg mr avdpag which it has in com- 
mon with n comp. also (ft$». It is safe to say that the plus 
was introduced from another recension; |J is therefore a 
mixed text. We must therefore fall back on ro as repre- 
sentatives of the sub-group s.%. 

But ro require correction. The order viov (,apa vwv (.afijjpi 
is clearly impossible. Restore on the basis of $j and ufi: 
viov Cap-Ppt viov z,apa. As neither Carmi (first in order) nor 
Zarah (last in order) was a household {oimg), it follows 

"mak by man" — margous 335 

that Km oikovq which was read neither by s nor by JI stands 
in the wrong place. It apparently entered the text from 
the margin ; and it stood there opposite mra avSpaq which is 
an old error common to r^I for mr oimvq. mm avdpaq belongs 
at the head of verse 18 and in front thereof there had drop- 
ped out by homoioteleuton mi npoor/yayov tov oikov fau/J/M, comp. 

Hence the archetype of rJJ read: mi npoarivexdvaav mm s^ovq 

warpiav mi avedeixdq <%io<; o (,apa Kat npoarjyayov tov Srifiov tov (apa <jcar 
oikov(^> mi avedeix^l o o</cof ^afifjpi <jtat itpoaTjyayov tov oikov £a/i[)pi Kara 
av6pag^> mi avedeixdr/ axap viog x a Pf Li mov < H a / i I^PO > vlov < C£ a P c C> []• 

I am somewhat in doubt as to whether rw tyvfois avda which 
ro have at the end stood in the archetype. The points of 
similarity with the archetype of ufi are unmistakable. 
Both represent a revision of B and testify to the reading 
/car otmvg in verse 17. Interesting is the expression %«wr 
warptov (ro; for which s has simply ■KaTpia.q). In the parallel 
text I Kingdoms 9, 21 Lagarde's Lucian renders nnscD by 
warpia in the place of yvlri of the other texts, just as it writes 
(pviti for D3B» in the place of o-k^tpov of the vulgar text. 

It remains to be said that mr oimvc of the Aldina goes 
back to the text of the group 15. 18. 64. 128 which is of a 
mixed character, comp. the form £>/?<?« for ^fipi. 

By way of recapitulation, I subjoin in parallel col- 
umns (a) the reconstructed original Hebrew text; (b) the 
reconstructed original Septuagint; (c) the recension under- 
lying Rutfffi; (d) the recension of Origen; (e) the received 
Hebrew text 








a-ip 1 "! " 

"km npoor/xOv 

17 K(M 'KpOG7JX&V ffav 


11 nai wpoar/xOl 

3-|p'1 " 

nhatyp nx 

Kara dr/uov( 

Kara Srjfiovg nar- 

Kara dr/fiovg 

nnetsto nx 


# tovda : 


13^1 eve8ux@V 

Kai avedeixOt 

Kai avedeixBy 


nns^D dk 




nnaB'D nx 



o £apa 

o t^apaei 



Kai irp0G7}xdri 

Kai irpoGTjyayov 

Kai npoarjx^rj 


nnsB'D n« 

tov Srjfiov 


nnse>D ns 


TOV £apa 

o t^apaei 



<0c<rr oiKovg 


% Kara avipa^ 



Kai evedeix^V 

Kai aveSecxOy 

nai avedeixdw 


not no 

omoc {.aufipi 

o oiKog t^a/ifipi 



3-llTl 18 

ls nat npoarjx- 

18 K(« irpoorjyayw 

w iiai irpooTjxQy 

anpi I8 

•not rra ns 

tov oikov $a/ippi 

o oikoc avrov ; 

wi nx 


Kara avSpa 

Kara avSpag 

Kara avSpa 



Kai evedetx8v 

Kai avedeix^l 

Kai avedetx&v 



a X ap 




vw( x a Pl li 

viog x a Pf- sl 

'ona p 

not ja 

vio( $a/ippt 

viov ^afiflpi 

viov l^ajidei 

nat p 

mt p 

viov $apa 

viov Jdpa 

viov £apa 

% T7JQ </>vXt/C 

lovda : 

mt p