Está en la página 1de 27

Grammatical Aspects of Biblical

Parallelism*
ADELEBERLIN
University of Maryland, College Park
I.

INTRODUCTION

A. Word Pairs a n d Parallelism


B. T h e Corpus
C. What is Grammatical Parallelism?
II.

MORPHOLOGICAL PARALLELISM

A. Word Pairs f r o m Different Morphological Classes


1. Noun II Pronoun
2. Noun or Pronoun II Relative Clause
3. Prepositional Phrase H Adverb
4. Substantive II Verb
B. Word Pairs f r o m t h e Same Morphological Class
j. Word Pairs of Different Tense
2. Word Pairs of Different Conjugation
3. Word Pairs of Different Gender
4. Word Pairs of Different Number
III.

SYNTACTIC PARALLELISM

A. Positive-Negative Parallelism
B. Parallelism Involving Change in Grammatical Mood
C. Subject-Object Parallelism
D. Nominal-Verbal Parallelism
IV.

SUMMARY

V .

GRAMMATICAL PARALLELISM IN PSALM 9 2

VI.

CONCLUDING REMARKS

* T h e author wishes to thank Prof. D. Hillers and Prof. M. Greenberg for reading the
manuscript and offering many helpful comments. A Summer Stipend from T h e National
Endowment for the Humanities enabled me to complete the work.
1

18

ADELE BERLIN
I.

[2]

INTRODUCTION

A. Word Pairs and Parallelism


I t has become a commonplace t o assert that parallelism is the outstanding
characteristic of biblical poetry. And while it occurs primarily in poetry, it
is not totally absent f r o m prose.1 Yet it is surprising how little is actually
known about the inner workings of this popular rhetorical feature.
T h e most important component of biblical parallelism seems t o b e
parallel word pairs. This has been the object of most studies of biblical
parallelism during the last quarter century. T h e pairs have been listed, 2
a n d t h e principles by which they were selected have been examined. Most
of the literature o n the subject accepts the hypothesis that there was a
tradition of fixed word pairs in Ugaritic a n d Hebrew poetry which automatically conditioned t h e use of the second member of a pair once t h e first
member was employed. This assumption has been recently challenged in
a study by William R. Watters, in which h e shows that the majority of word
pairs a r e not traditional (i.e. they d o not recur), a n d that the recurrence of
others can b e explained by reasons other t h a n a fixed tradition.3 But
whatever t h e truth may b e in regard to a tradition of fixed pairs, Watters
stresses again the importance of word pairs in parallelism. " T h e parallelism as well as the sense of t h e line is based u p o n t h e word pair." 4 I n fact,
not only a r e word pairs important, they a r e essential. "There can b e n o
parallelism without a word pair." 5
( 1 ) I use "prose" and "poetry" in the conventional manner, without attempting to define
either. Examples of prose verses containing parallelism are I Sam. 3:1, 2, 7. From such
examples it would appear that attempts to differentiate poetry from prose solely by the
presence o r absence of parallelism are misguided.
O n the basis of parallelism, formulaic pairs, and other rhetorical features J . S. Kselman
has recendy identified some hitherto unrecognized poetic fragments ("The Recovery of
Poetic Fragments from the Pentateuchal Priestly Source ,"/BL 97 [1978] 161-173). These
verses may, indeed, be poetic fragments, o r they may simply show, as d o several examples in
the present study, that many of the rhetorical features found in poetry also occur in prose.
O n possible criteria for differentiating prose from poetry see D. N. Freedman, "Pottery,
Poetry, and Prophecy: An Essay on Biblical Poetry," JBL 96 (1977) 5-26.
For some views on the difference between prose and poetic parallelism (not necessarily
shared by this writer) see Wm. Whallon, Formula, Character, and Context, Studies in Homeric,
Old English, and Old Testament Poetry (Cambridge: Harvard Univ. Press, 1969) 197-199.
(2) T h e most comprehensive listing of parallel word pairs in Ugaritic and the Bible is M.
Dahood, with the collaboration of T . Penar, "Ugaritic-Hebrew Parallel Pairs," Ras Shamra
Parallels I (Analecta Orientalia 49, Rome: Pontifical Biblical Institute, 1972) 71-382; I I
(Analecta Orientalia 50, 1975) 1-39
(3) Formula Criticism and the Poetry of the Old Testament (Berlin and New York: Walterde
Gruyter, 1976) 60-80.
(4) Ibid. 42.

(5) Ibid

[3]

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM

But while word pairs a r e a crucial ingredient of parallelism, there is


more t o parallelism t h a n the presence of word pairs. T h e use of word
pairs, fixed o r not, is n o t sufficient to explain t h e wide variety of parallel
stichs that occur in the Bible. T h e r e are other aspects of parallelism that
need t o b e recognized a n d analyzed. These other aspects are r e f e r r e d t o
by Roman Jakobson when h e states that
Pervasive parallelism inevitably activates all t h e levels of language
the distinctive features, inherent a n d prosodie, t h e morphologic a n d
syntactic categories a n d forms, t h e lexical units a n d their semantic
classes in both their convergences a n d divergences acquire a n autonomous poetic value. 6
I n short, it is not only words ("lexical units a n d their semantic classes") that
are paired in parallelism, b u t other aspects o r levels of language as well.
This study will examine some of the grammatical aspects ("morphologic
and syntactic categories a n d forms") of biblical Hebrew which a r e activated in biblical parallelism.
B. T h e Corpus
T h e corpus u n d e r consideration f o r this study is t h e entire Hebrew Bible.
I have purposely drawn o n all parts of the Bible because the phenomena
which I have observed a r e not limited to a particular genre, time, o r
author. 7 T h e verses cited represent examples of various types of grammatical parallelism. Some of t h e verses have been noted by others; some
are fortuitous findings of my own. T h e compilation of a complete list of all
verses illustrating all types of parallelism is not a practical undertaking
without t h e aid of a computer, n o r is it necessary t o prove the existence of
these types.
T h e r e is some injustice d o n e to a verse by lifting it out of context, as I
have done. I n many cases a verse is part of a larger structure built o n
lexical o r phonetic patterns, o r woven among other levels of parallelism.
T o appreciate the full measure of intricacy a n d beauty of a specific verse
one must always go back to t h e context. But t h e purpose of this study is t o
abstract certain general features of parallelism. When these a r e noted,
along with other rhetorical features, in entire passages, the "rhetorical
criticism" of those passages should b e more complete.
(6) "Grammatical Parallelism and Its Russian Facet,"Language 42 (1966) 423.Cf. also R.
Jakobson, "Poetry of Grammar and Grammar of Poetry," Lingua 21 (1968) 604.
(7) It may be that certain phenomena recur more often in one book or period but that is
not the concern of the present study.

2 0

[41

ADELE BERLIN

C. What is Grammatical Parallelism?


While there are many stichs which repeat the grammatical structure of
those which precede them, many others alter that structure in some way.
T h e alteration of grammatical structure in parallel stichs, or, better, t h e
pairing of two different grammatical structures in parallel stichs we t e r m
grammatical parallelism. Grammatical parallelism warrants f u r t h e r analysis
because it is one of the keys t o understanding how parallel stichs are
generated, a n d how the danger of monotony, inherent in such a repetitive
type of rhetoric, is avoided. 8
Grammar has two subdivisions: morphology a n d syntax. Morphology
deals with the individual components of a sentence, o r , in parallelism, a
stich; syntax is concerned with the sentence o r stich as a whole. When we
examine t h e morphological aspects of parallelism we will be comparing
the morphology of parallel terms (word pairs). Morphological parallelism is
the pairing of parallel terms f r o m different morphological classes (parts
of speech) o r f r o m the same morphological class but containing different
morphological components. Let us illustrate by citing two verses. T h e first
contains n o morphological parallelism, b u t complete morphological repetition. T h e second has o n e instance of morphological parallelism.
Ps 103:10

T h e corresponding terms occur in the same o r d e r in both stichs a n d a r e


quite obvious: 1. T h e negative particle is paired with the same. 2. A
term composed of a preposition + n o u n 4 possessive suffix ()
parallels a different term composed of the same parts of speech ().
I n addition, both terms a r e of the same gender a n d number. 3. A qal
perfect 3rd person singular verb is paired with a different verb with the
same morphology. 4. A preposition4 1 s t person plural suffix is paired
with a different preposition with the same suffix. Lexically these stichs a r e
in parallelism; morphologically, however, they a r e repetitive.
J o b 4:17

Although the word pairs occur in a different o r d e r in the two stichs


they a r e easily recognizable. T h e only morphological difference is in
/ / the latter containing a possessive suffix which is absent (for
semantic reasons) in the former.
(8) Other techniques employed to avoid monotony in parallelism are the omission
and/or addition of terms and the change of order of the terms.

[5]

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM

21

T h e first part o f this paper will list and examine several types o f
morphological parallelism. It will be shown that in some cases the pairing
of morphologically different terms seems to have been intentional, in
order to heighten the effect o f the parallelism beyond that already produced by the use o f semantically parallel terms.
T h e second part o f the paper will deal with syntactic parallelism that is,
parallel stichs with different syntax. The two verses quoted above are
examples o f syntactic repetition, not syntactic parallelism. In both verses
stich b has the same syntactic structure as stich a. The change in word
order in Job 4:17 does not alter this fact. T h e order of the words does not
affect the syntactic analysis o f a stich, and therefore it is not considered
grammatically significant for the purpose o f this study. 9
Both morphological and syntactic parallelism augment the total effect
of the parallelism, and provide an almost infinite number o f possibilities
for constructing parallel stichs.
II.

A.

MORPHOLOGICAL PARALLELISM

Word Pairs from Different Morphological Classes

Whenever a word from one part o f speech parallels a word from a


different part of speech we have a form o f morphological parallelism. I
have noted the following combinations:
1. Noun II Pronoun
Ps. 33:2

Ps. 3 3 : 8

,

2. Noun or Pronoun II Relative Clause10

Isa. 44:1



(Cf. Isa.4 5 : 4 ) "

(9) Word order, or, more specifically, chiasm, does have a semantic function, as demonstrated by F. I. Andersen, The Sentence in Biblical Hebrew (The Hague and Paris:Mouton,
1974) 119-140. Note especially p. 123: "Chiasm is a syntactic as well as an artistic device."
What Andersen means is that in poetry two parallel stichs arranged chiastically are to be
viewed as one sentence. However, in our study we are not interested in the two stichs as a
unit, but in the contrast between the two stichs.
(10) A relative clause is not a part of speech, but one does not find a relative pronoun by
itself, and often the pronoun is omitted.
(11) T h e word pair / / does not appear in Dahood's list of word pairs (Ras Shamra
Parallels) but occurs in Isa. 41:8, 9; 42:1; 43:10; 44:1, 2; 65:9, 15, as noted by W. Watters,
Formula Criticism 174, and also in Ps. 89:4; 105:6; and perhaps Hag. 2:23.

22

[6]

ADELE BERLIN

T h e relative clause may precede the noun, as in


,

Lam. 5:1
Song 3: !

...

3 Prepositional Phrase II Adverb


,

Ps. 34:2
4 Substantive II Verb

Ps. 3 4 : 2 a ' s o illustrates that a substantive (noun, adjective, o r participle)


can parallel a verb. This is n o t uncommon in parallelism a n d will b e
discussed u n d e r syntactic parallelism.
T h e r e is really nothing unusual about such pairings, f o r the members
of each category are normally used as substitutes f o r each other in biblical
Hebrew; pronouns take t h e place of nouns, prepositional phrases a n d
adverbs o f t e n serve the same syntactic function a n d are not always distinguishable, 12 a n d a relative clause may serve as a subject o r object t h e
same syntactic slot also filled by nouns a n d pronouns. T h e evidence seems
to indicate that any parts of speech that serve the same syntactic function
can b e used as parallel terms.1 3 This provides a grammatical dimension
f o r broadening the choice of parallel terms, in addition to the choice
provided by lexical-semantic possibilities. I n the next section we will show
that efforts were made t o broaden the selection of parallel terms even
when they a r e f r o m the same morphological class.
B. Word Pairs f r o m the Same Morphological Class
As stated in t h e preceding section, a term may b e paralleled by o n e f r o m a
different morphological class. I t is also possible to repeat the same term
(technically this is repetition, not parallelism), o r to parallel it by another
term f r o m the same morphological class. Parallel terms f r o m the same
morphological class may b e morphologically identical o r morphologically
different. T h e following sections will enumerate parallel pairs f r o m the
same class which are morphologically different. Moreover, their mor(12) Cf. P. Joon, Grammaire de l'hbreu biblique (Rome: Pontifical Biblical Institute, 1923)
267.
(13) Compare P. Kiparsky, "The Role of Linguistics in a Theory of Poetry," Daedalus,
(Summer, 1973) 235: "the linguistic sames which are potentially relevant in poetry arejust those which
are potentially relevant in grammar." (Italics in the original.)

[7]

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM

23

phological difference does not affect the meaning of the parallelism. I t


represents a type of paralleling o n the grammatical level, not o n t h e
semantic level.
. Word Pairs of Different Tense
T h e shift f r o m perfect t o imperfect forms, a n d vice versa (the qtl-yqtllyqtlqtl sequence) in parallel stichs has been recognized f o r some time a n d is
amply documented. 1 4 I t occurs with verbs f r o m the same root, as in
Ps. 29:10

o r with verbs f r o m different roots, as in


Ps. 26:4

15

It is important to emphasize that the qtl-yqtl shift occurs not f o r semantic


reasons (it does not indicate a real temporal sequence), 16 b u t f o r stylistic
reasons; it provides a n added dimension t o t h e parallelism.
2. Word Pairs of Different Conjugation
Another phenomenon in parallelism is the use of the same verbal root in
two different conjugations, such as is f o u n d in
Ps. 24:7

17

This was called the active-passive sequence by U . Cassuto, 18 a n d t h e


factitive-passive sequence by M. Held. 19 However, the shift is not limited
to specific conjugations o r grammatical voice, as can be seen f r o m t h e
following verses listed by M. Dahood: Ps. 64:5 (qal-hiphil); Ps. 77:12 (Ketiv)
(14) Cf. M. Held, "The YQTL-QTL (QTL-YQTL) Sequence of Identical Verbs in
Biblical Hebrew and in Ugaritic," Studies and Essays in Honor of Abraham A. Neuman (Leiden:
Brill, 1962) 281-290; M. Dahood, Psalms III {Anchor Bible, Garden City: Doubleday, 1970)
420-423.
(15) Cf. Held's treatment of this and the following verse, op. cit. 286.
(16) Cf. D. Clines, / , He, We, and They: A Literary Approach to Isaiah 53, Journal for the
Study of the Old Testament, Supplement Series 1 (Sheffield: Univ. of Sheffield, 1976)
47-48.
(17) Cf. Ps. 24:9 and M. Held, "The Active-Result (Factitive-Passive) Sequence of Identical Verbs in Biblical Hebrew and Ugaritic,"/L 84 (1965) 276, note 8.
(18) The Goddess Anath (Jerusalem: Magnes Press, 1971; first pub. in Hebrew in 1951)
47-48 = Biblical and Oriental Studies, vol. 2 (Jerusalem: Magnes Press, 1975) 58-59 T h e
article originally appeared in Tarbiz 14 (1943) 1-10.
(19) JBL 84 (1965) 272-282.

[8]

ADE L E BERLIN

2 4

(hiphil-qal); Ps. 139:21 (piel-qal).20 I n addition t o the verses cited by the


aforementioned scholars, 21 the following verses contain the same verbal
root in different conjugations in parallel stichs:
Gen. 6:12

#?

Gen. 17:17

22

I Sam. 1:28
Isa. 1:19-20



...

5 *

I S
I ...

This is a play o n words made possible because the root occurs in


different conjugations with different meanings: the qal means "to eat"
and the pual means "to be consumed."
Isa. 33:1

T h e root occurs here in the qal and huphal; the root only in
different forms of the qal. Not only does show variation in conjugation, but repetition of the same pattern f o r both verbs is avoided by using
in passive constructions a n d in active (impersonal) constructions.

(20) Dahood also lists Ps. 29:5; 38:3; 69:15 (Psalms III, 414).
(21) T h e verses listed by Cassuto and Held are Isa. 6:11 (?);Jer. 15:19; 17:14; 20:7; 31:3,
17; Ps. 19:13-14; 24:7; 69:15; Lam. 5:21. Held hesitates to include Isa. 6:11 because many
modern commentators, following the reading in the LXX, emend to . This
emendation, notes Held, seems also to be supported by Isa. 24:12 (JBL 84, 275, note 2).
However, while it is true that Isa. 24:12 contains the same idea and several of the same terms
found in 6:11 this does not mean that all of the terms need be identical. T h e word may
have been used in 24:12 because it makes a good phonetic complement to the word at the
end of the verse. The phonetic pattern in 6:11 is entirely different. Here one might see an
ABBA pattern composed of . I n my opinion emending is
unnecessary; the verbs and exemplify both a change of tense and of conjugation, as
well as a shift from plural to singular (see below).
(22) Cf. J . S. Kselman,JBL 97 (1978) 168. There is no need to change to or to
explain the lamed of as emphatic. T h e syntax of the two parallel stichs need not be
identical. "Will (a child) be born to a centenarian; and will Sarah who is ninety years old give
birth" makes a good parallelism. Cf. Jer. 20:14 and see below, SYNTACTIC PARALLELISM.

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM

2 5

Isa. 66:13


5?

J e r . 20:14

^
^

J e r . 23:19




(cf. J e r . 30:23)

H o s . 12:13-14

" W

^
Mie. 6:14b
Job 2 2 : 3 0

90
5

Furthermore, it has been noted that terms which are used as parallel pairs
in parallel stichs may occur elsewhere, in poetry o r in prose, in juxtaposition (one after another) o r in collocation (at some distance f r o m o n e
another). 23 T h e same is t r u e of word pairs consisting of the same verbal
root in different conjugations. 24
Gen. 7:23
Gen. 25:21
Lev. 13: !9-20

$? . . 9
! . . .
$ . $

Josh. 6:1

Isa. 4 5 : 1


5
?

(23) Cf. M. Dahood, Ras Shamra Parallels I, p. 87.


(24) Many of Cassuto's and Held's examples really belong in this category. R. Gordis
correctly distinguishes "the use of the same verb in two different tenses or voices, in two
separate and parallel stichs" and "the use of the two verbs within the same stich." H e then further
subdivides the latter category into those which constitute the "plea and response" formula,
and those which constitute the "action and result" formula (The Book ofJob, Commentary, New
Translation, and Special Studies [New York: Jewish Theological Seminary, 1978] p. 511.

26

ADELE BERLIN

Isa. 5 7 : 2
Ezek. 14:6
Zeph. 2:1
Mal. 2:10-11
_1

Ps. 92:13 4

[IO]

^


(cf. SS 1 1 1 ) 2 5

. . . *
?. . .

Both a shift in tense a n d a shift in conjugation a r e f o u n d in


Hos. 5 : 5

M. Held's explanation f o r t h e shift in conjugation is that "the device is


stylistic a n d would seem to aim at stressing a n d emphasizing the effect o r
result of t h e action referred t o in the first stichos." 26 H e also suggests that
there were fewer available parallel pairs f o r verbs than f o r nouns, a n d so
rather than repeat t h e same verb in the same f o r m , it was modified
slightly. 2 7 His observation o n the effect of t h e device appears to b e
correct, b u t I doubt that the biblical author was ever at a loss t o find a
parallel verb if h e so chose. I t would seem that using the same root in a
different conjugation is, at times, more effective than using a totally
different verb, because it produces the assonance a n d the play o n words
which is so much a part of biblical rhetoric. I n this respect verses containing the same verbal root in different conjugations a r e only one subset of
the set of verses in which two different forms of the same root occur in
parallel stichs, n o matter in what part of speech these roots may b e f o u n d .
Examples are:
J u d . 5:28

Mie. 1:16

Ps. 80:6

(25) These verses were called to my attention by Prof. Moshe Greenberg.


(26) JBL 84 (1965) 874.
(87) Ibid. 875.

[11]

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM

J o b 11:18
28

Ruth 2:12

27



3 Words Pairs of Different Gender

T h e paralleling of verbs in t h e qtl-yqtl sequence a n d verbs of different


conjugations should not b e viewed as isolated rhetorical peculiarities.
They a r e part of a broader picture of morphological parallelism. I n
addition t o alteration of tense a n d verbal conjugation, there appears t o b e
some evidence of morphological parallelism involving a shift in gender
a n d in number.
I t is obvious f r o m such verses as
Isa. 1:2

that two terms need not b e of t h e same gender o r n u m b e r to b e parallel.


Even in a n incomplete parallelism the term omitted in the second stich
(which is to be understood) may b e different in gender and/or n u m b e r
f r o m its expressed counterpart in the first stich. 29 Examples of this a r e
Ps. 30:6


)(

Ps. 19:5


)(

These examples merely prove that parallel terms, expressed o r u n d e r stood, need not be in t h e same gender o r number. What remains t o b e
investigated is to what extent a similarity o r difference in gender o r
n u m b e r influences the choice of parallel terms. We examine the case of
gender first.
Umberto Cassuto pointed o u t that there a r e Ugaritic a n d biblical
examples which show that o f t e n a masculine word is used in reference t o a
male o r masculine term, a n d a feminine synonym is applied to a female o r
feminine term. 3 0 H e cited
(28) Cf. R. Gordis, The Book ofJob, 511-513. This verse is not included in the lists, but its
assonance is noted in the commentary.
(29) Cf. R. Sappan, The Typical Features of the Syntax of Biblical Poetry in its Classical Period
(Unpub. dissertation, Hebrew University, Jerusalem, 1974) 55, note 45. Sappan discusses
only the difference in gender.
(30) The Goddess Anath 44-46 = Biblical and Oriental Studies, vol. 2, 66-68; originally
published in Leshonenu 15 (1947) 97-102.

[12]

ADELE BERLIN

2 8

J e r . 48:46





(cf. Deut. 21:10-11)

Nah. 2:13b

Isa. 3: !

I n Isa. 3-1 Cassuto explains, the word harks back to ( here


grammatically masculine, cf. Isa. 3:8), a n d to t h e feminine .
T h e pattern is chiastic.
These three verses contain three sets of nearly identical word pairs:
/ / , / / , a n d / / . Cassuto suggested that these sets,
o r a t least the second term in each, were chosen in order to match the
gender of another word in their respective stichs. T h a t is, the choice of
these words was based o n morphologic considerations.
But there is a slightly different way to view the phenomenon in these
three verses. Surely there were other word pairs which were of t h e
required gender. 3 1 O n e must ask why such similar terms were chosen in
these verses. T h e use of such closely related parallel terms is so striking as
to indicate a n intent t o emphasize their morphology. What these pairs
suggest to m e is not only that they were selected t o match the gender of
other words in their respective stichs, but also that they were intended to
parallel each other o n a morphological level, much like the pairs composed of the same root in different conjugations. Other sets of nearly
identical terms, o n e in the masculine a n d o n e in the feminine, a r e f o u n d
in
,
(cf. 30:23)

J e r . 23:19
Isa. 52:2
32


( ketiv:)

Other verses which employ both masculine a n d feminine forms of the


same root, although they may not be exactly synonymous o r parallel, a r e

(31) For example, compare Jer. 48:46 with Num. 21:29.


(32) Understanding both and to mean "captive" thus RSV.

[13]

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM

Ezek. 25:13
Ezek. 25:15

...
...

2 9




...

(This verse is part of a larger play o n the root .)


T h e use of the same adjective in different genders, although this is
determined by the modified nouns, may also b e considered morphological parallelism. (It also serves t o p u t emphasis o n t h e adjective.)
Gen. 11:6
Isa. 66:8

33

Especially interesting is
Ps. 51:19

T h e word may b e either gender, and, indeed, t h e pair a n d


appear in 51:12 where both a r e masculine. Perhaps has been used in
verse 19 as a feminine in o r d e r to produce t h e masculine-feminine alternation which stands o u t in / / .
T h e use of the same substantive in different genders in parallel stichs
creates t h e same type of assonance that is produced by the use of t h e same
verb in different tenses o r conjugations.
From all of these verses we see that one way of forming a parallel word
pair is by alternating t h e gender of t h e same word. T h e corollary to this is
that o n e way of forming parallel stichs is by alternating the gender of
those stichs. This is never t h e sole way in which a parallelism is constructed, b u t it may b e o n e of t h e levels of parallelism operating in certain
verses.
T h e r e a r e many verses besides those already noted in which there is
agreement in gender within a stich a n d alternation of gender f r o m o n e
parallel stich to the next. Examples are
Ps. 144:12

. . .

Ps. 126:2

Prov. 1:8

(33) This echoes Gen. 11:1.

ADELE BERLIN

SO

[141

T h e arrangement of the genders in these verses may b e accidental, b u t t h e


effect is a parallelism o n the morphological level. T h e question that
remains is: is this juxtaposition of genders intentional? I cannot prove that
such is the case, f o r the following reason: Most parallel terms a r e chosen
f o r lexical reasons, not f o r morphological reasons. Since all nouns in
Hebrew a r e either masculine o r feminine, there are inevitably many
combinations containing o n e term in each gender, e. g. , / /
/ / . T h e use of such combinations does not prove a n intent t o
alternate genders. Even the use of pairs which were originally based o n a
morphological principle, e.g. / / , / / , does not necessarily
prove that t h e author constructed his parallelism o n a morphological
level. Pairs such as / / a r e probably chosen o n t h e lexical level i. e.
they a r e traditional word pairs and/or are perceived as natural counterparts.
O n e verse which may b e suggestive of a n intentional alternation of
genders in parallel stichs is
Isa. 3:8a

T h e word may b e grammatically feminine, as in Ps. 114:2; Lam. 1:3,


o r masculine, as in Hos. 5:5. By choosing to construe it as masculine h e r e
(and also in 3:1), Isaiah has created a morphological parallelism. I n 3:8
the two genders appear t o balance each other, a n d may even create a
merismus. 3 4 T h e presence of this alternation in gender heightens t h e
effect of t h e parallelism.
T o summarize: Isa. 3:8 suggests that there may have been a n intentional switching of genders in parallel stichs. Verses such as Ps. 144:12;
126:2; Prov. 1:8 probably contain a coincidental shift in gender. I n the
case of word pairs, however, it seems m o r e certain that pairs of nearly
identical terms in different genders constitute a type of morphological
parallelism.
4. Word Pairs of Different Number
We t u r n now to a consideration of morphological parallelism involving
number. Here, too, o n e must take into account that some word pairs will
(34) T h e pair Jerusalem // Judah also creates a merismus by employing a part and its
whole. T h e effect of totality is emphasized by the chiastic word order. The verbs also
constitute a totality since both verbs apply to both subjects: Jerusalem and Judah have
stumbled and fallen. Thus the grammar, the choice of parallel terms, and the word order all
work toward the same end.

[1 5 ]

GRAMMATICAL ASPECTS O F BIBLICAL PARALLELISM

3 1

contain o n e singular a n d o n e plural term f o r lexical reasons. Most of these


fall into o n e of the following categories:
a. Some words, such as , , , a r e grammatically plural although they have a singular meaning. They will most likely be paired with
a singular term.
b. Some words usually occur in the dual o r plural, e.g.,, b u t
often have a singular parallel mate.
c. Some words, although grammatically singular, have a collective
meaning, e.g. , , a n d will often parallel a plural.
d . O n e of the principles by which word pairs a r e formed is the paralleling of a whole a n d its part. This seems to b e the explanation f o r such pairs
as / / in Ps. 144:10; / / in J e r . 7:17 a n d passim;
/ / in Ps. 48:12; 97:8.
e. T h e r e are traditional, logical, o r natural pairs, such as / / in
J o b 25:5; / / in Prov. 31:28; / / in Deut. 32:;35 / / in
Deut. 32:336
T h u s , there are many verses in which, f o r lexical-semantic reasons, a
singular term will parallel a plural one. But, in addition to these, there a r e
numerous verses which contain this type of parallelism f o r non-lexicalsemantic reasons. These verses contain a singular term paralleled by a
plural (or a compound, which generates a plural predicate) f o r n o apparent reason other than t o create a parallelism o n t h e morphological level.37
This is most apparent when t h e same word appears in both stichs in
different numbers, as in
J u d . 5:28
Hos. 5 : 5 b

38

(35) may mean elders of the generation preceding the father cf. Prov. 17:6.
Perhaps it might even be translated "grandfathers." T h e shift from singular to plural is
natural, since a person has only one father but more than one elder.
(36) This also follows the rule for paralleling numbers: // x+1.
(37) W. Watters has also observed the paralleling of a singular by a plural, but explains
the phenomenon as being necessary for metric reasons. Formula Criticism 105: "The poetry is
literally loaded with cases where one half of a word pair is plural and the other half is
singular. Yet in most all instances, the use of singular o r plural has no impact upon the
understanding of the line. By so varying the singular-plural aspect of the words in pair, the
lines are balanced in more uniform lengths."
(38) Prof. Greenberg pointed out to me that the plural of may have been
conditioned by the plural of ;cf. , Gen. 42:35

3 2

ADELE BERLIN

Prov. 14:12=16:25
J o b 6:15

[16]

o r when demonstrative pronouns are paralleled, as in


Isa. 66:8
J e r . 5:9
J o b 10:13

J o b 12:9

J o b 18:21

Lam. 5:17

I t is also striking t o find two parallel verbs in different numbers, as in


Deut. 32:7

Although many scholars p r e f e r to read o r instead of there is n o


evidence t o support such a reading. T h e parallelism is clearly singular //
plural. This pattern is repeated in the following stichs:
. T h e pair / / has been discussed above. While it was n o t
f o r m e d primarily o n a morphological level, it reinforces the morphological pattern of this verse.
T h e four-stich parallelism in Deut. 32:7 is only o n e of several four-stich
stanzas that have a morphological pattern based o n number. O t h e r
examples a r e
Isa. 40:4




39

(39) I have written in Hebrew Annual Review (1979) on my interpretation of this verse.

[171

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM

33

where t h e pattern is singular // plural; singular // plural,


Ps. 92:13-14

where t h e pattern is singular // singular; plural // plural, a n d


Ps. 126:5-6

where t h e pattern is plural // plural; singular // singular.


As in t h e case of gender, t h e use of the same o r similar adjectives in
different n u m b e r emphasizes t h e adjective a n d yields a morphological
parallelism.
Gen. 11:1
Isa. 54 : 7

40


41

As we can see f r o m Ps. 126:5-6 either a singular o r a plural can b e used


generically in Hebrew. Moreover, we often find that a verse o r passage
uses both thereby producing a n alternation in number. For example.
Prov. 14:33

Prov. 18:15

Prov. 29:27

Compare also Ps. 1:1-3, which speaks of the righteous in the singular, a n d
the parallel section in vv. 4-5, which describes t h e wicked in t h e plural.
Singular-plural alternation apparendy varies freely. We have arbitrarily chosen o n e term, , which always occurs in the plural, a n d have
(40) For cf. Ezek. 37:17.
(41) and are not normally word pairs, but can be contrued here as parallel
terms because they each occupy the same position in their respective stichs. T h e fact that
they are phonetically similar ( and ), and that both are modified by similar adjectives
adds to the impression that they are parallel terms.

34

[18]

ADELE BERLIN

noted that it is paralleled by in Ps. 64:11 ; 97:11 ; a n d probably 94:15,


and paralleled by in Ps. 32:11 a n d by in Ps. 36:11.
Finally, we must comment o n what importance the recognition of the
device of alternating n u m b e r has o n the interpretation of the biblical text.
T h e Massoretic T e x t contains several cases of singular-plural parallelism
which are not reflected in the versions.
Gen. 12:3a


42

(Versions: )

Isa. 44 : 26



(Versions: )

Ps. 114:2



(Versions:)

Deut. 26:13

43


(Versions: )

I n light of t h e foregoing discussion we should not conclude that the M T is


corrupt, o r that the Versions h a d a different text, but rather that the
Versions were simply not sensitive to this particular device.
T h e r e is ample evidence that the Bible contains morphological parallelism. Most likely there a r e other types besides those mentioned here.
T h e shift in person, a n accepted rhetorical device in the Bible(cf. Mie. 7:9;
Ps. 104:13; 145:6; Song 1:2) should probably b e considered morphological parallelism.
T o conclude this section o n parallel terms f r o m t h e same morphological class I o f f e r a n example in which every word pair shows morphological
parallelism.
J e r . 9:10

I n this verse none of t h e parallel nouns match in respect to number:


(42) U. Cassuto comments on this verse: "The difference between the plural those who
bless you and the singular him who curses you was introduced, it seems, for the sake of
diversification and variation in the parallelism, for which reason a change was also made in
the order of the words of the two clauses" (From Noah to Abraham (Jerusalem:Magnes Press,
1

964] 315)

(43) This verse is prose and does not, strictly speaking, contain a parallelism, but reflects
the same rhetorical usage of a shift in number in closely linked phrases.

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM

35

/ / , 44 / / , a n d even / / although strictly


speaking this is not a parallel word pair. 45 T h e verbs are in the qtl-yqtl
sequence. T h e use of morphological parallelism does not exclude t h e use
of other types of parallelism o r parallelistic devices. This verse also employs chiastic word arrangement a n d rhythmic parallelism. Phonetic parallelism is noticeable in the rhyme of ,, a n d ,, a n d
the alliteration ,.
III.

SYNTACTIC PARALLELISM

We have seen that individual terms can b e paralleled by a member of


the same morphological class (with o r without a morphological change of
the components), o r by a member of a different morphological class. A
similar principle operates in t h e paralleling of a stich as a whole: a stich
may b e paralleled by o n e with the same syntax (syntactic repetition o r
identity) o r by one with different syntax (syntactic parallelism). T h e
following a r e some ways by which syntactic parallelism is achieved: 46
A.

Positive-Negative Parallelism

This is a well known type of parallelism in which t h e same thought is


expressed twice, once positively a n d once negatively. (This is a f o r m of
synonymous parallelism a n d is not to be confused with antithetic parallelism.) Examples are
Prov. 1:8

Prov. 6:2o

Hab. 3:17

(44) On this pair see above, note 34.


(45) T h e phrases and are not lexically o r grammatically parallel. They
are phonetically similar, both beginning with mem, and metrically similar, both having the
same number of syllables and accent pattern. They also occupy the same position in parallel
stichs. Cf. above, note 41.
(46) On what basis did the poet choose an alternate syntax? P. Kiparsky, Daedalus
(Summer, 1973) 236 hypothesizes that "those syntactic elements which are counted as
parallel for the purpose of verse are, at some point in the derivation, counted as sames
according to transformational grammar." That is, the two different syntactic structures used
in parallel stichs are transformations of the same underlying sentence. While this has not
been demonstrated for biblical parallelism,the examples which follow suggest that such is
the case.

3 6

ADELE BERLIN

[20]



Hab. 3 : 1 7 contains f o u r stichs in a positive-negative-positive-negative
pattern, all expressing lacking o r ceasing.
This device is often f o u n d in prose, as in
Gen. 37:24b

Deut. 9:7

I Sam. 3: i b

I Sam. 3:2b

I Kings 3:18b

T h e negative may precede the positive, as in


Prov. 3:1
B.

Parallelism Involving Change in Grammatical Mood

T h e r e are a number of parallelisms in which two different grammatical


moods are paired. This may involve indicative // interrogative
Ps. 6:6

interrogative // indicative
Ps. 7 3 : 2 5

interrogative // imperative
Ps. 19:13

T h e r e are many cases of imperative //jussive / jussive // imperative. Some


involve second person jussive a n d imperative (i.e. also second person).
Verse of this type in Psalms have been noted by M. Dahood. 47 T h e r e are
also verses containing a n imperative and a third person jussive.
Deut. 32:1

(47) Psalms III 423-424

[21]

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM

Mie. 6:1
Prov. 3: !

37

Eccl. 5:1

These verses could be analyzed o n the morphological level as a change in


person.
C. Subject-Object Parallelism
Many parallel stichs are structured in such a way that the subject in one
becomes the object in the other.
Gen. 27:29

Ge11 37 : 33

J e r . 1:5
J e r . 20:14
Hos. 5:3

. . .
...



Ps. 2:7

Ruth 1:21

Lam. 5:4

Many of the verses listed u n d e r MORPHOLOGICAL PARALLELISM, Word Pairs


of Different Conjugation, are also examples of subject-object parallelism.
For example, J e r . 20:14 contains three types of grammatical parallelism:
subject-object, change in conjugation, a n d positive-negative, as well as
chiastic word order. Some cases of imperative // jussive also manifest
subject-object parallelism, e.g. Mie. 6:1 a n d Eccl. 5:1.

3 8

ADELE BERLIN

[22]

D. Nominal-Verbal Parallelism
T h e verses listed in this section all contain a nominal clause paralleled by a
verbal clause (or vice versa). Some verses utilize a substantive a n d a v e r b
f r o m t h e same root.
Ps. 97:9

Ps. 145:18

R u t h 2:18

T h e use o f a substantive a n d ver b f r o m t h e same root is also f o u n d i n


passages which a r e non-parallelistic.
Exod. 12:10
Lev. 13:12-13

. . . . . .
:

Lev. 13:17 (cf- 13 34)


Lev. 13:46

.. .

. . .

Some verses employ substantives a n d verbs f r o m different roots.


Mie. 6:2b
Ps. 34:2
Ps. 4 9 : 4

Com pare also Ps. 3 4 : 1 9 ! 5 : 8 ; 8 1 : 1 4 .


T h e s e categories of syntactic parallelism a r e n o t mutually exclusive. I t
is common t o find a verse displaying m o r e t h a n o n e type, e.g. J e r . 20:14,
m e n t i o n e d a b o v e ; Ps. 50:8: v e r b a l - n o m i n a l , su b j e c t - o b j e c t , a n d
negative-positive; Deut. 32:1: imperative-jussive a n d verbal-nominal
( / / ( ) also plural-singular, / / , although this is a
common word pair n o t f o r m e d o n t h e morphological level).

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM


IV.

39

SUMMARY

T h e section o n morphological parallelism has demonstrated that there


is m o r e t o paralleling terms t h a n t h e selection of semantically appropriate
word pairs. It is not only important to see which terms are used, b u t how
they a r e used. T h e f o r m of t h e word pairs is as interesting as their content.
T h e structure of the language permits the paralleling of o n e term by
another f r o m the same morphological class o r f r o m a different class.
Using a t e r m f r o m a different class automatically introduces a grammatical change in one stich a n d creates a morphological parallelism. Even
when the parallel term belongs t o the same morphological class its f o r m is
often varied in some way, by changing the tense, conjugation, gender,
number, o r person. This also produces a morphological parallelism.
We cannot know in all cases whether the biblical authors employed
morphological parallelism intentionally, instinctively, o r accidentally. I n
some cases involving a shift in gender o r n u m b e r morphological parallelism resulted f r o m lexical necessity. I n other cases it seems to have
resulted f r o m a n intentional manipulation of the grammar of the Ianguage. But whatever its origin, its presence certainly heightens the effect
of the parallelism a n d adds a dimension of interest a n d variety to t h e stichs
in which it occurs.
T h e inclination to vary t h e structure of parallel stichs is not confined to
the morphological level. O n the syntactic level we see that variety is
provided by pairing two stichs with different syntax. T h e categories of
syntactic parallelism illustrated here a r e positive-negative, change in
grammatical mood, subject-object, and nominal-verbal.
Syntactic parallelism a n d morphological parallelism together constitute grammatical parallelism. Grammatical parallelism is a device
whereby the grammar of the language is activated in order t o create
parallel stichs. T h e parallelism is achieved by structural substitutions o r
pairings in addition t o lexical-semantic pairings.
T h e purpose of this study was to call attention to the phenomenon of
grammatical parallelism by showing individual examples of many types
scattered throughout the Bible. But how frequently does grammatical
parallelism occur in relation to other types of parallelism? How does it
manifest itself within a complete passage? T h e answers to these questions
are not within the scope of the present study, but the following section
provides a sample analysis which, if carried o u t f o r many passages, would
give some indication of the prevalence of grammatical parallelism.

4 0

ADELE BERLIN
V .

[24]

GRAMMATICAL PARALLELISM IN PSALM 9 2

This brief analysis of Ps. 92, based o n the MT, shows how grammatical
parallelism may occur in a complete passage. T h e psalm was chosen at
random, b u t if it is at all typical, grammatical parallelism is a very pervasive device. Within fifteen verses (not including t h e superscription) there
are ten occurrences of grammatical parallelism. For purposes ofcomparison I have noted M. Dahood's interpretation {Psalms II, Anchor Bible) of
the pertinent verses in brackets.
v. 2 a change f r o m third person ("to the Lord") to second person ("to
your name, Elyon"). [Dahood proposes t o eliminate the "incongruity" by
parsing the lamedh preceding t h e Tetragrammaton as the vocative particle. This requires that the lamedh of the first stich be analyzed differently
f r o m t h e lamedh of the second stich, not in a n d of itself impossible, b u t this
would then leave without a n object, a r a r e occurrence (with t h e
exception of Neh. 12:24; I Chron. 25:3; I I Chron. 31:2).]
v. 3 singular // plural, / / . [Dahood translates "daybreak" a n d
"watches of the night." T h e use of here is compared to Ps. 16:7 a n d
134:1, n o n e of them convincing proof f o r Dahood's translation.]
V.5 qtl-yqtl sequence (?) a n d object-subject parallelism. T h e "me" of
the first stich becomes the "I" of the second. [Dahood understands t h e
verbs as being in the qtl-yqtl sequence a n d translates both by English past
tense. I n this case, however, it is not certain that this is a true rhetoric
qtl-yqtl; a real perfect-imperfect may have been intended: "You m a d e m e
48
happy . . . (therefore) I will sing . . . ". Dahood also notes the transition
f r o m second person subject to first person subject both here a n d in 11,
although h e does not speak in terms of subject-object parallelism.]
v. 6 interrogative // indicative. A rhetorical question, "how g r e a t . . . "
is echoed by a declarative statement. [Dahood considers t o b e a
double-duty interjection a n d translates "How g r e a t . . . How immensely
d e e p . . . ".]
v. 8 T h e construction of this verse is complex a n d difficult. T h e verse
appears t o have three stichs. 49 T h e first two a r e parallel to each other a n d
contain a nominal-verbal parallelism achieved by t h e paralleling of a n
infinitive, , with a finite verb,. T h e last stich of this verse can b e
understood as the parallel of t h e first two, that is, a + b // c, b u t may b e
(48) Cf. RSV and the new Jewish Publication Society translation.
(49) But cf. the new JPS translation:
though the wicked bloom, they are like grass;
though all evildoers blossom,
it is only that they may be destroyed forever.

[2 5 ]

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM

4 1

better taken as the parallel of v. 9. Semantically, the thought is antithetic:


"to destroy them forever" // "but You will b e exalted forever." Lexically,
the parallelism is based o n t h e word pair / / . [Dahood u n d e r stands as emphatic lamedh+finite verb, " H e completely destroyed
them. . . " This leaves o n e with t h e disturbing feeling that the " H e " of 8
is someone other than the "But you, Yahweh" of v. 9.]
v. 11 object-subject parallelism. Again t h e "me" of the first stich
becomes the "I" of the second. Cf. v. 5.
v. 12 singular // plural, "my eye" // "my ears." Cf. v. 3. [Dahood
ignores the difference in n u m b e r a n d translates both in the plural. (He
does not divide the verse in t h e usual way.)]
v. 13-14 These two verses a r e closely linked, forming a four-stich
parallelism. T h e y contain a shift f r o m the singular in v. 13 t o the plural in
v. 14, a n d a shift in conjugation, a n d ( although these terms d o
not parallel each other directly). [Dahood notes t h e shift in conjugation.
H e also notes the shift in number, a n d explains it by saying that in v. 13 t h e
word is the subject, b u t in vv. 14-15 t h e subject is a compound,
a n d . This seems t o b e taking t h e imagery a bit too literally, by picturing
o n e in v. 13 a n d two trees in vv. 14-15. T h e shift in number is better
understood as the use of either o r both numbers in a generic sense, as in
Ps. 126:5-6.]
v. 16 positive-negative parallelism
T h e feeling that o n e gets f r o m reading Dahood, a n d others, is that
there is a reluctance t o accept a wide range of grammatic "incongruities"
in biblical poetry, a n d the result is often a forced explanation of t h e m o r
a n emendation. Hopefully, this study will show that some of these grammatical shifts are stylisticthat is, they were used f o r rhetorical purposes,
and d o not require contorted interpretations of the text o r cast doubts o n
its accuracy.
VI.

CONCLUDING REMARKS

This study has shown that there is a grammatical aspect t o biblical


parallelism that parallelism activates the grammatical level of t h e language as well as the lexical-semantic level. This is really not a new idea a t
all. I t is present in Lowth's famous definition of parallelism:
When a proposition is delivered, a n d a second is subjoined t o it, o r
drawn u n d e r it, equivalent, o r contrasted with it in sense, or similar to
it in the form of grammatical construction, these I call parallel lines.50
(50) Isaiah: A New Translation (London: Wm. Tegg and Co., 1778) viii. (Italics mine.)

[26]

ADELE BERLIN

4 2

What Lowth calls "similar in t h e f o r m of grammatical construction" I take


t o m e a n either grammatically identical (grammatical repetition) o r
grammatically equivalent (grammatical parallelism). T h e existence of
grammatical parallelism has been assumed all along, but has never been
given t h e attention bestowed o n the lexical a n d semantic aspects of parallelism. I t is certainly worthy of that attention because it is used frequently
a n d is o n e of the devices which accounts f o r the wide variety of parallel
stichs in the Bible.
Having called attention t o o n e aspect of parallelism, I must emphasize
that there a r e other aspects. T h e matter of omission and/or addition of
terms a n d change of word o r d e r has been mentioned in passing. T h e r e is
also a phonetic aspect of parallelism (assonance) a n d a rhythmic o r metric
aspect. A n d , of course, t h e lexical-semantic aspect continues t o b e of
interest. Most important, it is necessary to examine t h e interaction a m o n g
all these aspects of parallelism. I would like t o o f f e r a few observations o n
this subject.
O n e conclusion f r o m t h e study of syntactic parallelism is that parallel
terms may occur in the same morphological f o r m b u t may be used t o fill
different syntactic functions in each stich. This can b e seen in some
examples of subject-object parallelism, e.g. Gen. 27:29; Lam. 5:4; Eccl.
5:1. Because of this, there is t h e possibility of creating tension between t h e
morphological a n d syntactic levels, o r between t h e lexical and t h e grammatical levels. This kind of tension (and it may involve other levels as well)
adds zest t o the parallelism. Its effect is roughly analogous t o cases in
English poetry where the units of meter o r rhyme d o not correspond t o
the syntactic units, a n d enjambment results. T w o examples of such tension, o r "play o n grammar," a r e Ps. 49:5 a n d Lam. 5:3.
Ps. 49:5

T h e morphological sequence in both stichs is t h e same: verb, p r e position+noun, noun+suffix. But if one analyzes the verse only o n t h e
morphological level, o n e emerges with the preposterous notion that
/ / a n d / / . Obviously the word pair is / / , b u t t h e
illusion is created by using as a n indirect object a n d as a direct
object. I n linguistic terms these two stichs have t h e same surface structure
b u t not t h e same d e e p structure.5 1

(51) T h e reverse is true of the parallel stichs discussed in the previous sections; they have
the same deep structures but different surface structures.

[271

GRAMMATICAL ASPECTS OF BIBLICAL PARALLELISM

Lam. 5:3

43

According to the syntax of t h e two stichs, / / , since these


are both t h e subject of their respective clauses, a n d / / , both
being attributes of the subjects. But lexically, the usual pairing would b e
/ / a n d //. This tension between t h e grammatical a n d lexical
levels is produced by using o n e part of a word pair as t h e subject in t h e first
stich a n d the other part as the predicate in the second stich.
These a r e j u s t some of the possibilities that a grammatical analysis of
parallelism yields. Parallelism is a n extremely complex device. T h e r e a r e a
multitude of potential parallel permutations f o r any given stich. T h e
biblical authors showed a mastery of parallelism which has yet t o b e
appreciated by modern critics who correct t h e text in o r d e r to make
"better," that is, m o r e simplistic, parallelisms. Parallelism is a marvelous
vehicle of literary expression a n esthetically pleasing a n d engaging
device. T h e m o r e we understand of it, t h e m o r e we will b e able t o
understand the biblical text a n d appreciate its literary qualities.

También podría gustarte