Vous êtes sur la page 1sur 5

The Shibboleth Incident (Judges 12:6)

Author(s): E. A. Speiser
Source: Bulletin of the American Schools of Oriental Research, No. 85 (Feb., 1942), pp. 1013
Published by: The American Schools of Oriental Research
Stable URL: http://www.jstor.org/stable/1355052
Accessed: 24-11-2016 00:10 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

The American Schools of Oriental Research is collaborating with JSTOR to digitize, preserve and
extend access to Bulletin of the American Schools of Oriental Research

This content downloaded from on Thu, 24 Nov 2016 00:10:47 UTC
All use subject to http://about.jstor.org/terms

Number 85 February 1942



Although the biblical term .ibblkt has long been naturalized ou

Hebrew in the sense of " criterion, catchword," the exact basis o

usage is far from clear. After more than two thousand years of interp

tation the matter is still open. The postulates and queries set dow
the medieval Hebrew commentators are reflected to this day in the
recent discussions on the subject.
The substance of the incident is plain enough. Some time in or cl
the eleventh century B. C. a group of Ephraimites sought to esca
band of vengeful Gileadites by attempting to get across the Jorda
into Palestine. Halted at the fords the fugitives pretended to be n
of Gilead. They were betrayed, however, by their inability to pron
a chosen test word in the proper Gileadite manner, with consequ
that can hardly be called academic. The telltale element was the
sound of the word .*ibbdlEt, which at that time probably had the
*Subbultu. The best that the Ephraimites could do was *subbultu,
was not good enough.

The test was thus a phonetic one, involving a dialectal differ

The dialect boundary was represented by the Jordan. That m

beyond dispute. It has also been assumed universally that the dia
peculiarity in question set off the Ephraimites from other speake
Hebrew, in Palestine and Transjordan alike, thus constituting an E
mite isogloss.a Ready to hand as such an assumption may be,
scarcely bear closer investigation. What is more, it appears that
erroneous premise has been chiefly responsible for our failure to r
struct the primary details of the shibboleth incident, even thoug
catchword itself has long since become proverbial. We have yet to
how suspects were caught by the catchword. A review of the prob
therefore, may not be without interest.
The meaning of the test word is of minor importance. Elsewhere in

Old Testament it has the sense of " ear of corn "1 or, less comm
" flood, torrent." " In our passage it is taken in the former sense b

versions as the Greek Codex Vaticanus and Aquila,3 as well as

modern scholars.4 On the other hand, reference to flowing wate

assumed by the medieval Hebrew commentators and a majority o
moderns, evidently because such an allusion would be more approp

to the occasion.

The prevailing explanation of the phonetics involved operates with the

between S and
is said
to have
been pronounced
the Ephraimites
as s.s,The
so that
But it is

difficult to reconcile this view with the available facts. We have no knowl-

edge of any West Semitic language that fails to include both s and s as
a [I. e., a linguistic phenomenon characteristic of a given area.--W. F. A.]

1'Gen. 41: 5ff.; Ruth 2: 2; Zach. 4: 12.

2Psalms 69: 3, 16; Isa. 27: 12.

3 The usual Septuagint interpretation, however, is "password."

'Cf. Liebmann, ZAW 25 (1905) 161.


This content downloaded from on Thu, 24 Nov 2016 00:10:47 UTC
All use subject to http://about.jstor.org/terms

Number 85 February 1942

independent phonemes. The tw

neither is absent from any sing
of the incident under discussion
Ephraimites were not asked, " H
Instead they were challenged sp
herited $ had changed to s in th
the slightest independent indic
required sound to save their ne

to them."

Less commonly advanced is an attempted solution based on the premise that the Ephraimites employed the spirant t in place of the Gileadite . On this assumption subbult" would be the necessarily inadequate
writing for a spoken *tubbultu. This approach to the problem was fore-

shadowed by David Qimlbi (13th cent. A.D.)." It was developed by

J. Marquart and has been endorsed cautiously by Z. S. Harris." Marquart's reasoning was as follows: The Aramaic correspondents of

-ibb'ldt are 'ebaltd and tubld. This pair presupposes an original initial t
which may yet turn up in Arabic. Accordingly, Proto-Semitic possessed

a root 'bl from which is derived the word meaning " ear of corn "; further-

more, another root *tbl which yielded the Hebrew homonym of that

form meaning " flood, torrent." By the time of Jephthah the Ephraimites

had not yet lost the t-phoneme; that is why they said *tubbult".
There are three serious objections to this view. The first is etymological. The Aramaic variant with t- signifies " ear of corn " and not

"flood." ' The latter value can scarcely be separated from Arab. sbl

" rain, flow," so that a sibilant is assured in the cognate forms of Cana-

anite and Aramaic. The second objection is chronological. Goetze has

pointed out that there is no valid reason for dating the merging of t
with S in Palestine later than in Phoenicia, where the process is attested

at least as far back as the eleventh century.1o This argues for an

Ephraimite S from t in Jephthah's time. Thirdly, there is the question of

plausibility which applies here no less than it does to the alleged mispronunciation of i as s. For even if we grant that the Ephraimites had
retained the t-phoneme, there are no grounds for denying them the
possession of the normal Canaanite sibilants. In other words, they were
in a position to duplicate cheerfully the subbultu of the Gileadite sentries.
Plainly, then, the current explanations of the shibboleth incident leave
it with all the characteristics of a " tall story." And yet, the account as
we have it bears the marks of authenticity. Something must be wrong,
therefore, with the explanations of the episode rather than the recorded
statement about it.

1To assume some peculiar differences in the pronunciation of the ?-phoneme itself,
as is done, e. g., by Budde, Richter 89, is to resort to speculation unrelieved by any

semblance of fact.

6 Miqr&'6th Ged6l6th, ad loc.

7 ZAW 8 (1888) 151 ff.

s Development of the Canaanite Dialects (1939) 64.

* Cf. Gesenius-Buhl, ad loc.

10 Language 17 (1941) 168. [On the date of the oldest true Phoenician inscriptions
see now Studies in, the History of Culture (Leland Volume), 1942, pp. 34 ff.-W. F. A.

This content downloaded from on Thu, 24 Nov 2016 00:10:47 UTC
All use subject to http://about.jstor.org/terms

Number 85 February 1942

In these circumstances it will no

from a new angle. We know that t
the initial sound of the catchword
Supposing now that in the Gilead
t, which was lacking in the Canaa
the result be? Much the same, no
the English voiceless phoneme ex
Even more to the point is the exp

to reproduce the Arabic t. The b

Persian can achieve by way of p

[sa: lis]. The Ephraimites of the el

the phoneme in their own speech

rendering of *tubbultu could re

subbultu. And that is precisely wh

All of this is as yet less than a w

we have to account for the presence

We have seen that etymological d

fact, however, that the Aramaic for

S(-ebaltA, dubaltd) or t (tubla). Th

with t (tubbultu) alongside the no

a doublet arose is beside the po

analogic interference. At all event

Aramaic. Since it was present in o
well have been current in another
At this point the objection might
differs but little from Marquart's

of corn" whereas our concern i

Marquart went astray in trying t

on an etymological basis. In real

deriving the established homonym

two distinct roots. Arabic sbl ma

we get sunbulat-, sabalat- "ear o

ranges are thus easily linked seman

apparent etymological identity.
There is, therefore, no logical obs
*tubbultu " flood " at the time of

tion inconsistent with the requireme

phy. We have learned that the m

established as earlier than the 11th

process in Palestine need not have
show that the same change had d

"I In this connection it is interesting t

th, with its twofold value, is the shib

New International Dictionary).

11a [In view of the vast number of known loanwords in Semitic, it seems to me
that the original stem began with t and that Aramaic tubl4, etc., is thus genuinely

Aramaic, whereas Sib(b)altd and Subbaltd have been borrowed from Canaanite or
Accadian. Arabic suinbulah is in any event a loan from Aramaic, and so is presumably Arab. sdbalah (from 8ebalt4). Original t is strongly suggested by the
Canaanite cognate word for "tendril, vine," preserved in New Egyptian as sabir,

since original t appears regularly in Egyptian transcription of this age as s.--W. F. A.]

This content downloaded from on Thu, 24 Nov 2016 00:10:47 UTC
All use subject to http://about.jstor.org/terms

Number 85 February 1942

Gilead until a materially later d

characterized by cultural and poli
those in Palestine. The frontier st
sode is best exemplified by the ca
suggestive from our standpoint is
tory which has not given up the t
demonstrates. In the light of the
phoneme from the dialect of Gilea


One detail remains to be cleared

allowed to proceed on its course.
in the text by s. A glance at the r
that this is the only way in whic
pressed altogether. In the first pl
and is thus represented regularly
nology furnishes a complete paral

distinction between original t and

the spirant that is written invar

appear either as ? or s.12 This is t
and Amarna. In short, t had to be
as s unless such a writing was mean
of the required sound, which is ex
To sum up, the shibboleth inciden

dialect at the time of Jephthah

Palestine. That peculiarity consist

which the Hebrews of Palestine h
The resulting misfortune might h
Galilee just as painfully as it did s
feature in question stemmed from

current interpretations of the incide

on the wrong side of the river.

In conclusion, may I be permitted

provincial, which is designed to sh
this kind. Let us assume that a pa

taken place at Ebbets Field, Bro

ensued and the followers of the in

take to their heels. But they were

ing to cross the Brooklyn Bridg

pretend that they were Brooklyni

be fooled by such a ruse. " How

was [ward]. It was a fatal mistak

were treated to a description of th

that it was a peculiarity of the Ma
Brooklyn and all the rest of the c
Such a result might well be view
make it right.
12 Cf. Speiser, JAOS 59 (1939) 187 ff.; Goetze, Language 14 (1938) 136 f.

This content downloaded from on Thu, 24 Nov 2016 00:10:47 UTC
All use subject to http://about.jstor.org/terms