Gray: The Forms of Hebrew Poetry







                     THE FORMS


                HEBREW POETRY





                               OF THE OLD TESTAMENT







                             GEORGE BUCHANAN GRAY

                                         D.LITT., D.D.


















                                    HODDER AND STOUGHTON

                                    LONDON NEW YORK TORONTO









IT is impossible to go far at the present day in

any serious attempt to interpret the prophetical

books, or the books commonly called poetical,

or certain other parts of the Old Testament,

without being faced by questions relating to the

forms of Hebrew poetry. I was myself compelled

to consider these questions more fully than before

when I came to prepare my commentary on

Isaiah for the "International Critical Comment-

ary," and in the introduction to that commentary

I briefly indicated the manner in which, as it

seemed to me, the more important of these ques-

tions should be answered. But it was impossible

then, and there to give as full an exposition of the

subject as it requires. In the present volume I

have ampler scope. Yet I must guard against a

misunderstanding. Even here it is not my pur-

pose to add to the already existing exhaustive,

or at least voluminous, discussions of Hebrew

metre. My aim is different: it is rather to

survey the forms of Hebrew poetry, to consider

them in relation to one another, and to illustrate




their bearing on the criticism and interpretation

of the Old Testament.

            I have no new theory of Hebrew metre to set

forth ; and I cannot accept in all its details any

theory that others have elaborated. In my

judgment some understanding of the laws of

Hebrew rhythm has been gained: but much

still remains uncertain. And both of these facts

need to be constantly borne in mind in determin-

ing the text or interpreting the contents of

Hebrew poetry. Perhaps, therefore, the chief

service which I could expect of the discussion of

Hebrew metre in this volume is that it may on

the one hand open up to some the existence and

general nature of certain metrical principles in

Hebrew poetry, and that it may on the other

hand warn others that, in view of our imperfect

knowledge of the detailed working of these prin-

ciples, considerable uncertainty really underlies

the regular symmetrical forms in which certain

scholars have presented the poetical parts of the

Old Testament.

            The first six chapters of the volume are an

expansion of a course of University lectures

delivered in the spring of 1913. They were

published in the Expositor of May, June, July,

August, September, October and December of

the same year, and are now republished with

some modifications and very considerable addi-

tions. The two last chapters, though written


                        PREFACE                              vii


earlier, are' in the present volume rather of the

nature of an Appendix, being special studies in

the reconstruction of two mutilated acrostich

poems. These also originally appeared in the

Expositor, the former (Chapter VII.) in September

1898, the latter (Chapter VIII.) in September 1906.

Except for the omission of a paragraph which

would have been a needless repetition now that

the two discussions appear together, and for a

few slight or verbal alterations, and for additions

which are clearly indicated,. I have preferred to

republish these chapters as they were originally

written. They were both, and more especially

the former, written before I saw as far, or as

clearly, as ,I seem to myself at least now to do,

into the principles of Hebrew metre: but addi-

tional notes here and there suffice to point out

the bearing of these more fully appreciated prin-

ciples on the earlier discussions, which remain

for the most part, unaffected, largely, I believe,

because in the first instance I followed primarily

the leading of parallelism, and parallelism is

likely for long to remain a safer guide than metre,

though metre may at times enforce the guidance

of parallelism, or act as guide over places where

parallelism will not carry us.

            A word of explanation, if not of apology, is

required for the regularity with which I have

added translations to the Hebrew quoted in the

text. In many cases such translation was the




readiest way of making clear my meaning; in

others it is for the Hebrew student superfluous,

and parts of the book can scarcely appeal to

others than Hebrew students. But a large part

of the discussions can be followed by those who

are but little familiar or entirely unfamiliar with

Hebrew. For the sake of any such who may

read the book, and to secure the widest and

easiest use possible for it, I have regularly added

translations, except in the latter part of Chapter

IV., where they would have been not only super-

fluous, but irritating to Hebrew students, and use-

less to others.

            My last and pleasant duty is to thank the

Rev. Allan Gaunt for his kindness in reading the

proofs, and for offering various suggestions which

I have been glad to accept.


                                                G. BUCHANAN GRAY.














                                                CHAPTER I


INTRODUCTORY                                                                            3


                                                CHAPTER II


PARALLELISM : A RESTATEMENT                                            37



                                                CHAPTER III


PARALLELISM AND RHYTHM IN THE BOOK                        

            OF LAMENTATIONS                                                           87



                                                CHAPTER IV


THE ELEMENTS OF HEBREW RHYTHM                                   123



                                                CHAPTER V


VARIETIES OF RHYTHM: THE STROPHE                                  157



                                                CHAPTER VI



            ON CRITICISM AND INTERPRETATION.                       201



x                      FORMS OF HEBREW POETRY



                                    CHAPTER VII                                               Page


THE ALPHABETIC POEM IN NAHUM                            243



                                    CHAPTER VIII



            AND X.                                                                                  267




            SAME TERM IN PARALLEL LINES                                 295



                                        INDEX I


OF PASSAGES OF SCRIPTURE                                                    297



                                        INDEX II


OF MATTERS                                                                                   301













                        CHAPTER I












                   CHAPTER I




FAILURE to perceive what are the formal elements

in Hebrew poetry has, in the past, frequently led

to misinterpretation of Scripture. The existence

of formal elements is now generally recognised;

but there are still great differences of opinion as

to the exact nature of some of these, and as to

their relation to one another and large questions

or numerous important details of both the lower

and higher criticism and of the interpretation of

the Old Testament are involved in these differ-

ences. An examination of the forms of Hebrew

poetry thus becomes a valuable, if not indeed a

necessary, means to the correct appreciation of

its substance, to an understanding of the thought

expressed in it, in so far as that may still be

understood, or, where that is at present no 

longer possible, to a perception of the cause and

extent of the uncertainty and obscurity.

            More especially do the questions relating to

the two most important forms of Hebrew poetry



4                      FORMS OF HEBREW POETRY


—parallelism and metre—require to be studied in

close connexion with one another, and indeed

in closer connexion than has been customary of

late. I deliberately speak at this point of the

question of parallelism and metre; for, on the

one hand, it has been and may be contended

that parallelism, though it is a characteristic of

much, is never a form of any, Hebrew poetry,

and, on the other ,hand, it has been and still. .is

sometimes contended that metre is not a form of

Hebrew poetry, for the simple reason that in

Hebrew poetry it did not exist. Over a question

of nomenclature, whether parallelism should be

termed a form or a characteristic, no words need

be wasted; the really important question to be

considered later on is how far the phenomena

covered by the term parallelism can be classified,

and how far they conform to laws that can be

defined. A third form of some Hebrew poetry is

the strophe. This is of less, but still of considerable

importance, and will be briefly considered in its

place; but rhyme, which is not a regular feature

of Hebrew poetry, and poetical diction need not

for the purposes of the present survey be more

than quite briefly and incidentally referred to.

            The first systematic treatment of any of the

formal elements of Hebrew poetry came from

Oxford. There have been few more distinguished

occupants of the chair of Poetry in that university

than Robert Lowth, afterwards Bishop of London,

                        INTRODUCTORY                            5


and few lectures delivered from that chair have

been more influential than his De Sacra Poesi

Hebraeorum Praelectiones Academicae. These lec-

tures were published in the same year (1753) as

another famous volume, to wit, Jean Astruc's

Conjectures sur les memoires originaux dont it

paroit que Moyse s'est servi pour composer le

livre de la Genese. It is as true of Astruc as of

Lowth that "in theology he clung to the tradi-

tional orthodoxy";1 yet Astruc was the first

to apply a stylistic argument in a systematic

attempt to recover the original sources of a portion

of the Pentateuch, and Lowth, by his entire

treatment of his subject, marks the transition

from the then prevailing dogmatic treatment of

the Old Testament to that treatment of it which

rests on the recognition that, whatever else it

may be, and however sharply distinguished in

its worth or by its peculiarities from other litera-

tures, the Old Testament is primarily literature,

demanding the same critical examination and

appreciation, alike of form and substance, as

other literature. Owing to certain actual char-

acteristics of what survives of ancient Hebrew

literature, documentary analysis has necessarily

played an important part in modern criticism of

the Old Testament; and if, narrowing unduly

the conception of Old Testament criticism, we

think in connexion with it mainly or exclusively


1 T. K. Cheyne, Founders of Old Testament Criticism, p. 3.



of documentary analysis and questions of origin,

Astruc may seem a more important founder of

Modern Criticism than Lowth. But in reality

the general implications of Lowth's discussion of

Hebrew poetry, apart from certain special con-

clusions reached by him to which we shall pass

immediately, make his lectures of wider signifi-

cance than even Astruc's acute conjectures ; and

we may fairly claim that, through Lowth and

his two principal works, both of which were

translated into German, the Lectures by Michaelis,

the Isaiah by Koppe, Oxford, in the middle of

the eighteenth century, contributed to the critical

study of the Old Testament and the apprecia-

tion of Hebrew literature in a degree that was

scarcely equalled till the nineteenth century was

drawing to its close.

            It is a relatively small part of Lowth's lectures

that is devoted to those forms or formal char-

acteristics of Hebrew poetry with which we are

here concerned: of the thirty-four lectures one

only, the nineteenth, is primarily devoted to that

form with which Lowth's name will always be

associated, though the subject of parallelism was

already raised in the third lecture. The maturer

and fuller discussion of this and kindred topics

was first published in 1778 as a preliminary dis-

sertation to the translation of Isaiah. Briefly

summed up, Lowth's contribution to the subject

was twofold: he for the first time clearly

                        INTRODUCTORY                            7


analysed and expounded the parallelistic struc-

ture of Hebrew poetry, and he drew attention to

the fact that the extent of poetry in the Old

Testament was much larger than had generally

been recognised, that in particular it included

the greater part of the prophetic writings.

            The existence and general characteristics of

parallelism as claimed by Lowth have never been

questioned since, nor the importance for interpre-

tation of recognising these; nor can it be ques-

tioned, once the nature of parallelism is admitted,

that parallelism occurs in the Prophets as well

as in the Psalms, and in many passages of the

Prophets no less regularly than in many Psalms.

If, then, on the ground of parallelism, the Psalms

are judged to be poetry, the prophetic writings

(in the main) must also be regarded as poetry ;

and, if, on the ground of parallelism, a translation

of the Psalms is marked, as is the Revised Version,

by line divisions corresponding to the parallel

members of the original, a translation of the

Prophets should also be so marked; and by

failing so to mark the prophetic poetry, and

thereby introducing an unreal distinction between

the form of the Psalms and the form of the pro-

phetic writings, the Revised Version conceals

from those who use it one of the most important

and one of the surest of the conclusions which

were reached by Lowth in his discussion of

Hebrew poetry.



Whether after all parallelism is itself a true

differentia between prose and poetry in Hebrew,

may be and will be discussed; but it will be useful

before proceeding to a closer examination either

of parallelism or of other alleged differentiae

between prose and poetry, to recall the earlier

scattered and unsystematic attempts to describe

the formal elements of Hebrew poetry.

            It has always been recognised that between

mediaeval Jewish poetry and the poetry of the

Old Testament there is, so far as form goes, no

connexion ; nor, indeed, any similarity beyond

the use, especially by the earliest of these

mediaeval poets such as Jose ibn Jose and

Kaliri, of acrostic, or alphabetic schemes such as

occur in Lamentations i.-iv. and some other

poems1 in the Old Testament. The beginnings

of mediaeval Jewish poetry go back to the ninth

or tenth century A.D. at least; it arose under the

influence of Arabic culture, though it may also

have owed something to Syriac poetry; it

flourished for some centuries in the West, and

particularly in Spain. This poetry was governed

by metre and rhyme;2 and the metre was quanti-

tative. The same period was also, and again

owing to the influence of Arabic culture, an age


            1 Enumerated below, p. 244 f.

            2 The introduction of rhyme into Hebrew poetry is attributed to

Jannai; rhyme was also employed by Kaliri. Both Jannai (probably)

and Kaliri were Palestinians, and both lived in or before the ninth

century A.D.: see Graetz, Gesch. des Judenthums, v. 158, 159.

                        INTRODUCTORY                            9


of Jewish grammarians and philologists. These

recognised the difference between the old poetry

and the new, but contributed little to an under-

standing of the forms of the older poetry beyond

a tolerably general acquiescence in the negative

judgment that that older poetry was not metrical.

In any case, no living tradition of the laws of the

older Hebrew poetry, the poetry of the Old Testa-

ment, survived in the days of the poets Chasdai

(A.D. 915-970), Solomon ibn Gabirol (1021-1058,

or 1070), Judah hal-Levi (born 1085) ; of the

grammarians and philologists, of whom some

were poets also, Dunash ibn Labrat (c. 920-990),

Menahem ibn Saruk (c. 910-970),   Abu'l-Walid

(eleventh century), Ibn Ezra, and the Kimlhis

(twelfth century). The older poetry had long

been a lost art. Whatever these mediaeval

scholars say of it has, therefore, merely the value

of an antiquarian. theory; and however interest-

ing their theories may be, they need not detain

us longer now.

            But there exist a few far earlier Jewish state-

ments on the formal elements of the poetry of

the Old Testament which run back, not indeed

to the time of even the latest poems within the

Old Testament, but to a time when, as will be

pointed out in detail later on, poetry of the

ancient Hebrew type was still being written.

Statements from such a period unquestionably

have a higher degree of interest than those of the



mediaeval Jewish scholars. Whether as a matter

of fact they point to any discernment of the :real

principles of that poetry, and whether they do

not betray at once misconceptions and lack of

perception, is another question. At all events,

it is important to observe that while the authors

of these statements were Jews, the readers with

a view to whom they wrote were Greeks. So far

as I am aware, there is no discussion of metre,

or parallelism, or in general of the formal elements

of Hebrew poetry, in the Rabbinical writings, that

is to say in Jewish literature written in Hebrew

or Aramaic, until after the gradual permeation

of Jewish by Arabic scholarship from the seventh

or eighth century A.D. onwards. We owe the

earliest statements on Hebrew poetical forms to

two Jews who wrote in Greek—to Philo and to


            Philo's evidence is slight and indirect as to

the poetry of the Old Testament. In the De

vita Mosis i. 5 he asserts that Moses was taught

by the Egyptians " the whole theory of rhythm,

harmony and metre " (th<n te r[uqmikh>n kai> a[rmonikh>n

kai> metrikh>n qewri<an); but he nowhere states that

the poems attributed to Moses in the Pentateuch

are metrical. Of Jewish poetry of a later age he

speaks more definitely, if the De vita contem-

plativa is correctly attributed to him, and if the

sect therein described was a Jewish sect. It is

asserted in this tract (cc. x. xi.) that the thera-

                        INTRODUCTORY                            11


peutae sang hymns " in many metres and tunes,"

and in particular in iambic trimeters.

            The three statements of Josephus on the

subject are much more specific and definite. Of

Moses he says, in reference to Exodus xv. 2 if.,

that " he composed a song to God . . . in hexa-

meter verse" (e]n e[came<tr& to<n&);1 and again,

in reference to Deut. xxxii., that Moses read to

the Israelites "a hexametrical poem" (poi<hsin

e[ca<metron), and left it to them in the holy book.2

Of David he says that " he composed songs and

hymns in various metres ..(me<trou poiki<lou), making

some trimetrical, others pentametrical."3

            These exhaust the direct testimony of Jews,

who lived while poetry similar to that in the Old

Testament was still being written, to the metrical

character of that poetry. It is possible that we

have an indirect testimony to more specific

Jewish statements or theories in certain of the

patristic writers. It will be sufficient here to

refer to what is said by Origen and Eusebius and

Jerome;4 all these scholars belong to a period

before the new style of poetry adopted by the

mediaeval Jews had begun to be written, though

perhaps none of them belong quite to the age

when the older poetry was still practised as a

living art.


1 Ant. ii. 16. 4:   2 Ant. iv. 8. 44.  3 Ant. vii. 12. 3.

4 The passages, from these and other patristic writers have been

brought together and discussed by J. D611er (Rhythmics, Metrik and

Strophik in der bibl.-hebr. Poesie, Paderborn, 1899 ; see pp. 18-35).



Origen's reference to the subject of Hebrew

metre is to be found in a scholion on Psalm

cxviii. 1 (LXX). He agrees with Josephus that

Deuteronomy xxxii. is hexametrical, and that

some of the Psalms are trimetrical; but as an

alternative metre used in the Psalter, he gives

not the pentameter, as Josephus had done, but

the tetrameter. At the same time he clearly

recognises that Hebrew verses are different in

character (e!teroi) from Greek verses. Ley finds

two further statements in Origen's somewhat

obscure words: (1) that the metrical unit (den

vollen Vers) in Hebrew consists of two stichoi, not

of a single stichos; (2) that Hebrew metre was

measured by the number of accented syllables.

Eusebius refers to metre in Hebrew poems as

follows:  "There would also be found among them

poems in metre, like the great song of Moses and

David's 118th Psalm, composed in what the


            1 The scholion in question was published by Cardinal Pitra in Ana-

lecta Sacra, ii. 341, and reprinted thence by Preuschen in the Zeitschrift

fur die AT. Wissenschaft, 1891, pp. 316, 317; in the same Zeitschrift

for 1892 (pp. 212-217) Julius Ley translated and commented on the

scholion. The text being still none too well known or accessible, it

may be well to reproduce it here. The words commented on are

Maka<rioi oi[ a@mwmoi e]n o[d&, oi[ poreuo<menoi e]n no<m& kuri<ou, and the scholion

runs as follows:—ou!tw ge sti<xoj e]sti<n: oi[ ga>r par ]  [Ebrai<oij sti<xoi, w[j

e@lege< tij, e@mmetroi< ei]sin: e]n e[came<tr& me>n h[ e]n t&? Deuteronomi<& &dh<: e]n trime<tr&

de> kai> tetrame<tr& oi[ yalmoi<. oi[ sti<xoi ou#n, oi[ par ] [Ebrai<oij, e!teroi< ei]sin para>

tou>j par ] h[mi?n.   ]Ea>n qe<lwmen e]nqa<de thrh?sai, tou>j sti<xouj poiou?men.  “Maka<rioi

oi[ a@mwmoi e]n o[d&?, oi[ poreuo<menoi e]n no<m& kuri<ou.”  Kai> ou@twj a]rxo<meqa deute<rou

 [Ebrai<oij sti<xon e]n toiou<toij du<o (w[j [o[ ] tou?to a]nti<grafon gra<yaj oi[onei> pepoi<hke

th>n a]rxh>n tou? sti<xou met ] e]kqe<sewj):  to>n de> dokou?ntej deu<teron, mh> o@nta deu<teron,

a]lla> lei?mma tou? prote<rou met ] ai]sqh<sewj:  kai> tou?to pepoi<hken e]pi> o!lou tou?


                        INTRODUCTORY                            13


Greeks call heroic metre. At least it is said

(Octal, (pi-iv) that these are hexameters, consisting

of sixteen syllables ; also their other composi-

tions in verse are said to consist of trimeter and

tetrameter lines according to the sound of their

own language."1 The reference to Deuteronomy

xxxii. and Psalm cxviii. (cxix.) and the specific

metres mentioned are as in Origen; but whether

or not Origen suspected or asserted measurement

by accented syllables, Eusebius clearly refers to a

measurement by syllables, and thereby produces

the impression that the Hebrew hexameter was

of the same nature as the Greek: whereas Origen

distinctly asserts that Hebrew metres are as

compared with the Greek e!teroi. At the same

time, the final words in Eusebius have something

of the character of a saving clause.

            Scattered over Jerome's writings are a larger

number of specific statements, which may be

summarised as follows :

            1. Job iii. 2-xl. 6 consists of hexameters ; but

the verses are varied and irregular.2

            2. Job, Proverbs, the songs in Deuteronomy

(i.e. Deut. xxxii.) and Isaiah, "Deuteronomy et

Isaiae Cantica," are all written in hexameters or


            1 Praep. Ev. xi. 5. 5 : the translation given above is Gifforci's.

            2 "Hexametri versus sunt, dactylo spondaeoque currentes ; et

propter linguae idioma crebro recipientes et alios pedes non earumdem

syllabarum, sed eorumdem temporum. Interdum quoque rhythmus

ipse dulcis et tinnulus fertur numeris lege metri solutis," Praef. in

Job (Migne, Patr. Lat. xxviii. 1082).

14                    FORMS OF HEBREW POETRY


pentameters.1 Yet elsewhere2 "Deuteronomii

Canticum" is said to be written in iambic tetra-


            3. Psalms cx. and cxi. are iambic trimeters.2

            4. Psalms cxviii., cxliv. and Proverbs xxxi.

10-31 are iambic tetrameters.2

            5. Lamentations i. ii. are in " quasi sapphico

metro"; but Lamentations iii. in trimeters.2

            6. The prophets, though the text of them

is marked off by commas and colons, are not


            But these statements, occur in such connexions,

or are accompanied by such qualifying phrases,

as to indicate that Jerome did not intend them

to be taken too strictly, or as exactly assimilating

Hebrew poetry in respect of its measurements to

classical poetry. Thus, the hexameters in Job

are said to admit other feet in addition to dactyls

and spondees; the "sapphic metre" of Lamenta-

tions i. ii. iv. is qualified as "quasi"; and in

forestalling incredulity, such as the Emperor

Julian is said to have expressed, as to the existence

of metre in Hebrew literature, Jerome speaks of

the Hebrew poems as being "in morem, nostri

Flacci"--after the manner of Horace.

            There is one further important observation

to be made with regard to Jerome: the authori-


            1 " Quae omnia hexametris et pentametris versibus . . . apud suos

composita decurrunt," Praef. in Chron. Eusebii (Migne xxvii. 36).

            2 Ep. xxx. (ad Paulam) (Migne xxii. 442).

            3 Praef. in Isaiam (Migne xxviii. 771).

                        INTRODUCTORY                            15


ties whom he cites for his statements are not his

own Hebrew teacher, but Philo, Josephus, Origen,

and Eusebius,1 to the first two of whom Origen

in turn may refer indefinitely in his phrase

e@lege< tij.

            From this we may with some probability con-

clude (1) that Jerome's views of the nature of

Hebrew poetry do not represent those of Jewish

scholarship of his day; but (2) that they are a

reproduction of the statements of Josephus, or

deductions made by Jerome himself from or in

the spirit of Josephus' statements. On whom

Eusebius relied for the statement (fasi> gou?n)

that the Hebrew hexameter contained sixteen

syllables we cannot say, but his informants were

scarcely Jewish contemporaries of his.

            If, then, any theory or tradition of the metrical

character of the old Hebrew poetry formulated


            1 " If it seem incredible to any one that the Hebrews really have

metres, and that, whether we consider the Psalter, or the Lamentations

of Jeremiah, or almost all the songs of Scripture, they bear a resemblance

to our Flaccus, and the Greek Pindar, and Alcaeus, and Sappho, let

him read Philo, Josephus, Origen, Eusebius of Caesarea, and with the

aid of their testimony he will find that I speak the truth: Preface to

the translation of Job (Fremantle's translation, p..491): Migne xxviii.

1082. This was written about A.D. 392; but Jerome had expressed

himself to much the same effect ten years earlier in a passage, partly

cited already in the original, in his Preface to the Chronicle of Eusebius :

"What can be more musical than the Psalter? Like the writings of

our own Flaccus and the Grecian Pindar it now trips along in iambics,

now flows in sonorous alcaics, now swells into sapphics, now marches

in half-foot metre. What can be more lovely than the strains of

Deuteronomy and Isaiah? What more grave than Solomon's words?

What more finished than Job? All these, as Josephus and Origen tell

us, were composed in hexameters and pentameters, and so circulated

amongst their own people."—Fremantle, p. 484: Migne xxvii. 36.

16                    FORMS OF HEBREW POETRY


by those who actually wrote it still survives, our

primary source for it is Josephus. But does

what Josephus says depend on a previously

existing theory or tradition? In all probability

it does not. Josephus, in commending Hebrew

poetry to his Greek readers, followed his usual

practice of describing things Jewish in terms that

would make a good impression on them. And

so he calls Deuteronomy xxxii. hexametrical--a

term which some modern scholars would still

apply to it—but he gives his readers no clue to,

even if he himself had any clear idea of, the

difference between these hexameters and those

of Greek and Latin poetry. Neither he nor any

of the Christian scholars who follow him defines

the nature of the feet or other units of which six,

five, four, and three compose the hexameters,

pentameters, tetrameters, and trimeters respect-

ively of which they speak ; and, indeed, so loosely

are these terms used that Jerome describes

Deuteronomy xxxii. on one occasion as hexa-

meter, and on another as tetrameter. Some

modern scholars continue to use these same terms,

but define more or less precisely what they mean

by them; and the Hebrew hexameters of the

modern metrist have far less resemblance to a

Greek or Latin hexameter than any of the numer-

ous English hexameters with which English poets

have at intervals experimented from the age of

Elizabeth down to our own times. There is no

                        INTRODUCTORY                            17


reason for believing that Josephus, Origen, or

Jerome really detected, or' even thought that

they detected, any greater similarity; Jerome's

“quasi," Origen's e!teroi, cover, as a matter of

fact, a very high degree of ,difference.

            Early Jewish observations on Hebrew metre

are neither numerous nor valuable ; but observa-

tions on the characteristic parallelism of Hebrew

poetry seem to have been entirely non-existent

earlier than the time of the mediaeval Jewish

grammarians. Josephus was stimulated to dis-

cover or imagine metre in Hebrew poetry by his

desire to commend it to the Greeks ; he had no

such stimulus to draw attention to parallelism,

for that corresponded to n6-thing in the poetry

of Greece or Rome. And another cause worked

against the recognition by the Jewish Rabbis of

the part played by parallelism in Hebrew poetry.

But before defining this cause it will be convenient

to record the extent to which Lowth's analysis

of parallelism was anticipated by the mediaeval


            Dukes1 drew attention to the fact that D.

Kimhi (c. A.D. 1160-1235) in his comment on

Isaiah xix. 8 calls parallelism "a reduplication of

the meaning by means of synonymous terms "

(tvnw tvlmb Nybf lvpk), and that Levi ben Gershon

had called it an elegance (tvHc jrd), and also

noted the fact that the same style was customary


            1 Zur Kenntnis der neuhebr. religiosen Poesie (1842), p. 125.

18                    FORMS OF HEBREW POETRY


with the Arabs. Schmiedl, in 1861,1 drew atten-

tion to the still earlier use by Ibn Ezra (A.D.

1093-1168) of these same expressions as well as

of some others with reference to parallelism. So

far as I am aware, similar observations in writers

earlier than Ibn Ezra have never yet been dis-

covered.2 Ibn Ezra's observations mar be sum-

marised as follows: it is an elegance of style, and

in particular a characteristic of the', prophetic

style, to repeat the same thought ,by means

of synonymous words.3 Whether in regarding

parallelism as peculiarly characteristic of the

prophetic style (tvxybnh jrd) Ibn Ezra anticipated

Lowth's observation that Old Testament pro-

phetic literature is, in the main, poetical in form,

is doubtful: for the examples of parallelism

given by Ibn Ezra are drawn, not from the

prophetical books, but from the prophetic poems

in the Pentateuch attributed to Jacob, Moses,

and Balaam.

            Far more important is Ibn Ezra's insistence

that parallelism is a form of poetry, and that

when a writer repeats his thought by means of

synonymous terms he is not adding to the sub-

stance, but merely perfecting the form of what

he had to say. This represents a reaction against


            1 In Monatsschrift fUr Gesch. u. Wissenschaft des Judenthums, p.157.

            2 Cardinal Pitra was of opinion that Origen's scholion given above

(p. 12 n.) recognised parallelism, but this is doubtful:

            3 Ibn Ezra cites as examples Genesis xlix. 6 a, b, Deuteronomy

xxxii. 7 c, d, Numbers xxiii. 8.

                        INTRODUCTORY                19


a mode of exegesis that treated such repetition

as an addition to the substance. It was this

mode of exegesis, doubtless, that militated against

the discernment of the real nature of parallelism

by earlier Jewish scholars. How could inter-

preters who attributed importance to every letter

and every external peculiarity of the sacred text

admit that it was customary in a large part of

Scripture to express the same thought twice over

by means of synonymous terms? If the fact

that RCYYV in Genesis ii. 7 is written with two

yods, though it might have been written with

one, was supposed to express the thought not

only that God “formed” man, but that He

formed him with two "formations," or "inclina-

tions," to wit, the evil inclination and the good

inclination, how could two parallel lines convey

no fuller meaning than one such line standing

by itself? The influence of this exegetical prin-

ciple lingers still; at an earlier time it was far-

reaching. For example, in Lamech's song (Gen.

iv. 23), " the man" and "the young man" came

to be treated not as what in reality they are,

synonymous terms with the same reference, but

as referring to two different individuals, one old

and one young, who were, then, identified with

the ancient Cain and the youthful Tubal-Cain.1

Again, the reduplication of the same thought in


            1 See the commentary of Rashi (eleventh century A.D.) on Gen.

iv. 23.

20                    FORMS OF HEBREW POETRY


two parallel lines is not recognised in.


Therefore, the wicked shall not stand in the judgment,

   Nor sinners in the congregation of the righteous (Ps: i. 1).


Rabbi Nehemiah, a Rabbi of the second century

A.D., said "the wicked mean the generation of

the Flood, and the sinners mean the men of

Sodom."1 If no other difference of reference

could be postulated between two parallel terms

or lines or other repetitions of a statement, it

was customary to explain one of the present world

and the other of the world to come.2 "Day and

night" is a sufficiently obvious expression for

"continually"; and a poet naturally distributed

the two terms between two parallel lines without

any intention that what he speaks of in the one

line should be understood to be confined to the

day, and what he speaks of in the second line to

the night: thus, when a Psalmist says (xcii. 1),


            It is a good thing . . .

            To declare thy kindness in the morning

                 And thy faithfulness in the night,


what he means is that it is good to declare both

the kindness and the faithfulness of God at all

times. Yet even some modern commentators

still continue to squeeze substance out of form

by making Psalm xlii. 9 (8)--

            By day will Yahweh command his kindness,

                And in the night his song shall be with me--


            1 Sanhedrin x. 3.

            2 See e.g. Sanhedrin x. 3 for several examples of second-century

exegesis of this kind.

                        INTRODUCTORY                            21


mean more than that the Psalmist is the constant

recipient of God's goodness; and herein these

modern commentators follow, in misconceiving

the influence of form, the early Jewish interpreter

Resh Lakish (third century A.D.) who explained the

verse thus: "Every one who studieth in the Law

in this world which is like the night, the Holy One,

blessed be He, stretches over him the thread of

grace for the future world which is like the day."

            To sum up this part of our discussion: Jewish

Rabbis in the second century A.D. misunderstood

the parallelism that is characteristic of most of

the poetry of the Old Testament, and, with the

exception of Philo and Josephus, no Jews appear

to have given any attention to any metrical laws

that may also have governed that poetry;2 and


            1 Talmud B. Hagigah 12 b ; ed. Streane, p. 64. Another passage

where some modern commentators have failed to see how much the

real range of thought is defined by parallelism is Hos. ii. 5 a, b

                        Lest I strip her naked,

                                    And set her as on the day she was born.

These two lines are entirely synonymous. For the correct understand-

ing of the second line the most important thing is to recall Job i. 21,

" Naked came I out of my mother's womb"; the two lines mean simply

this : Lest I strip her to the skin so that she becomes as naked as a

child just drawn from the womb. Such a note as Harper's in the Inter-

national Critical Commentary (p. 227), which is partly based on Hitzig's,

is not really interpretation: the lines do not mean that Israel is to

become a nomadic people again. Strangely enough, the modern

commentaries which I have consulted do not give the really pertinent

reference to Job i. 21: and it was not until I turned to Kimhi that I

found a commentator who did. He very correctly paraphrases the

second line: I will cause her to stand naked as on the day of her birth,

and regards it as repeating the meaning of the first line by synonymous

terms (nlmu m'7n7  '71:22 1>3sn).

            2 It is possible enough that the practice of distinguishing certain

poems (viz. those in Ex. xv., Deut. xxxii., Judg. v. and 2 Sam. xxii.)

by spacing within the lines, a practice still regularly observed in printed

22                    FORMS OF HEBREW POETRY


what Josephus says on that subject is expressed

in Greek terms, was written as part of his apology

for all things Jewish, and appears at most to

imply that Josephus had some perception of

difference of rhythm in different Hebrew poems.

The account he gives wears a rather more learned

air, but is in reality as vague and insufficient as

the account given to Dr. Dalman by some of

those who supplied him with his specimens of

modern Palestinian poetry.1


editions of the Hebrew Bible even when other poems such as Psalms

and Job are not so distinguished, goes back to this period. It is

certainly vouched for by sayings in both Talmuds (j. Meg. iii. 74, col.

2, bottom; b. Meg. 716 b; cp. Shabbath, 103 b, bottom), of which the

Jerusalem Talmud is commonly considered to have been completed

c. A.D. 350, the Babylonian c. A.D. 500; and by the time that the

tractate Soferim was written (probably c. A.D. 850), according to state-

ments therein contained (Soferim, ed. Joel Muller, xiii. 1, p. xxi), it was

customary in accurately written MSS. to distinguish Psalms, Proverbs,

and Job in the same way ; and in some of the earliest existing MSS.

Psalms and Job as well as the four passages above mentioned are so

distinguished. But it is difficult, not to say impossible, to derive from

these facts any theory of the nature of parallelism, or of the rhythm

of the lines so distinguished : on the contrary, the different divisions

of these poetical passages in different MSS., the failure to distinguish

at all such obvious poems as the blessing of Jacob in Gen. xlix., the

poems attributed to Balaam in Num. xxiii., xxiv., and the blessings of

Moses in Deut. xxxiiii. (cp. Ginsburg's edition of the Hebrew Bible),

and the fact that the directions in the Talmud for writing certain

passages vrcx,yipc;,s group together''the poems in Ex. xv., Deut. xxxii.,

etc., and the lists of the kings of Canaan in Jos. xx. 9-24 and of the sons

of Haman in Esth. ix., rather suggest the absence of any clear theory

of either parallelism or rhythm.

            1 "In modern Arabic folk-poetry the purely rhythmical has begun

to drive out the quantitative principle so that a distinction may be

drawn between quantitative and rhythmical poems." . . .

            "I have never been able to discover how the composers of this folk-

poetry go to work in the composition of these poems. To the question

whether there was nothing at all in his lines that the poet numbered so

as to secure regularity (Gleichmass), I received from several different

quarters the reply, that nothing at all was numbered, that for the folk-


                        INTRODUCTORY                            23


            And yet, in the second century A.D., Hebrew

poetry of the type found in the Old Testament

had not yet become a long obsolete type, as it had

become when the new art of rhymed, metrical

poems without parallelism was brought to per-

fection in the tenth to the twelfth centuries ; con-

temporaries of Josephus were still employing

parallelism with as much regularity and skilful

variation as the best writers of the Old Testament

period ; and in all probability, in many cases at

least, rhythmical regularity of the same kind, and

as great, accompanied these parallelistic com-

positions, as is found in any of the Biblical poems.

But later than the second century A.D. only

meagre traces of parallelism of the types found

in the Old Testament, or of the same kind of

rhythms as are used there, can be found;

and certainly, when the new Hebrew poetry

was created, it dispensed with parallelism—with

parallelism, at all events, as any constant feature

of the poems.

            Without prejudging the question whether

parallelism in Hebrew necessarily constitutes or

implies poetical form, it will be convenient at

this point to take a survey of those parts of

ancient Jewish literature outside the Old Testa-

ment in which either parallelism is conspicuous,


poetry there was only one standard (Mass)—absolute caprice. No

doubt it may be supposed that the individual poet instinctively imitates

the form of some poem that is known to him."—G. H. Dalman, Paid-

stinischer Divan, pp. xxii, xxiii.



or other features are prominent which distinguish

those parts of the Old Testament commonly

regarded as poetry. Most of this literature,

especially the latest of it, survives only in trans-

lation; and, with regard to much of it, it is

disputed whether it actually runs back to a

Hebrew original at all. The exact date, again,

of much of it is uncertain, and I shall, therefore,

attempt no rigid chronological order of mention;

in general the period in question is from the third

or second century B.C. to the second century A.D.

            Of the apocryphal books it was clear even

before the discovery of the Hebrew original that

Ecclesiasticus (c. 180 B.C.) must have possessed

all the characteristics of ancient Hebrew poetry ;

and even the alphabetic structure of li. 13-30 had

been inferred.1 But Ecclesiasticus may well be

older than some of the latest poems in the Old


            The Hebrew original of the first book of

Maccabees (c. 90 B.C.) has not yet been recovered:

but, even through the translations, it is easy to

detect certain passages to which the use of

parallelism gives an entirely different character

from the simple prose narrative of the main body

of the work. Such passages are the eulogies of

Judas (iii. 3-9) and Simon (xiv. 6-15) and also

i. 25-28, 36 b-40, ii. 8-11 (13 a). Isolated distichs,


            1 By G. A. Bickell in the Zeitschrift fur katholische Theologie, 1882,

pp. 319 ff.

                        INTRODUCTORY                            25


such as occur in ii. 44 and ix. 41, may be citations

from now lost poems, as vii. 17 is from a still

extant Psalm (lxxix. 2, 3). In ix. 20, 21 reference

is made to an elegy on Judas and the opening

words are cited. It is possible to infer the Hebrew

original of these words with practical certainty,

and to detect in

            lxrWy fywvm |  rvbgh lpn jyx

            How hath the valiant man fallen,

                        He that delivered Israel,


the opening of a poem constructed after the same

form1 as elegies in the Old Testament.

            In the book of Judith, which may have been

written about 150, or as some think about 80 B.C.,

we find a long poem of praise and thanksgiving;

in part, it is a close imitation of earlier poems in

the Old Testament; but its parallelistic, as was

also presumably its rhythmical, regularity is by

no means least where it is most independent, as,

for example, in the lines (xvi. 8-10)

            She anointed her face with ointment,

                        And bound her hair in a tire;

            And she took a linen garment to deceive him,

                        Her sandal ravished his eye,

            And her beauty took his soul prisoner,

                        The scimitar passed through his neck,

            The Persians quaked at her daring,

                        And the Medes at her boldness were daunted.


            Not only the Apocrypha, but the Pseudepi-

grapha, contain much, the New Testament,


            1 See below, pp. 96 ff.

26                    FORMS OF HEBREW POETRY


perhaps, a little, that was originally written in

Hebrew and was poetical in form. Among these

specimens of late Hebrew poetry we may certainly

include the eighteen " Psalms of Solomon " (c.

50 B.C.)1 and perhaps some of the most ancient

elements of the Jewish liturgy, such as the "Eight-

een Blessings " (c. A.D. 100), and the blessings

accompanying the recitation of the Shema’; 2

possibly also the Magnificat and other New Testa-

ment Canticles.3 Several of the apocalypses also

include poems; in those which he has edited

more recently, Dr. Charles has distinguished the

poetry from the prose by printing the former in

regular lines. Without admitting that all parts

thus distinguished by him or others possessed


            1 The parallelistic structure is indicated in my translation of these

Psalms in The Apocrypha and Pseudepigrapha of the Old Testament

(ed. R. H. Charles), ii. 631-652.

             2 The Hebrew text of these and of the " Eighteen " is conveniently

brought together in W. Staerk, Altjudische liturgische Gebete (Bonn,

1910). The rhythm is indicated in the notes and German translation

in P. Fiebig, Berachoth: Der Mischnatractat Gegenspruche, pp,. 26 if.

            3 Dr. Burney has recently argued that the parable of the last Judg-

ment in Matt. xxv. 31-46 was a Hebrew poem ; and his Hebrew trans-

lation from the Greek text of the Gospel, his metrical analysis of the

poem and his English translation, as far as possible in the rhythm of

his Hebrew reconstruction, deserve careful attention. See the Journal

of Theological Studies for April 1913 (vol. xiv. 414-424).

            Parts, but parts only, of Matt. xxv. 31-46 are thrown into parallel

lines by Dr. Moffat also in The New Testament : a new translation.

That parts only are so arranged in this passage is the more noticeable

because in a considerable number of other, longer or shorter, passages

in this translation of the New Testament an arrangement in lines is

adopted. It is, however, tolerably clear that this line arrangement is

not always intended to imply poetical form. And certainly, even for

example in the parts of 1 Cor. xiii. which are so arranged, the form is

not that of Hebrew parallelism; in vv. 1-3 the formal effect is obtained

by exact repetition of the same phrase ("but if I have no love"), not

by repetition of the same thought by means of synonymous terms.

                        INTRODUCTORY                27


poetical form in the original, I think it may be

safely said that such apocalypses as the Twelve

Patriarchs, the Book of Jubilees, the Apocalypse

of Baruch and IV. Esdras do each contain some

such passages.

            Now of these books or passages which show

the same characteristics as the poetry of the Old

Testament, some at least were written by men

who were contemporary both with Josephus

and also with those who after A.D. 70 founded

that Jewish school at Jamnia of whose methods

of exegesis (in the second century A.D.) examples

have been given above. At the very time that

the Rabbis were examining scripture with eyes

blind to parallelism, other Jews were still writing

poems that made all the old use of parallelism.

This may be proved by reference to the Apocalypse

of Baruch: for with regard to this book I believe

that it may be safely asserted1 (1) that it was

written in Hebrew, (2) that it was written not

earlier than c. A.D. 50, and therefore (3) that

its author was in all probability a contempo-

rary, though perhaps an elder contemporary, of

Josephus and of the founders of the school of

Jamnia. But this book contains a long passage

(xlviii. 1-47) that is among the most regular and

sustained examples of parallelism in the whole

range of Hebrew literature ; a sufficiently large

portion of it may be cited here to prove this


            1 Cp. R. H. Charles, The Apocalypse of Baruch.



the translation is in the main that of Dr. Charles;

for the line division, which in one place (v. 14)

involves an important change of punctuation, I

am responsible).

            2 O my Lord, Thou summonest the advent of the times, and

                                                they stand before Thee;

            Thou causest the power of the ages to pass away, and

                                                they do not resist Thee:

            Thou arrangest the method of the seasons, and they

                                                            obey Thee.

            3 Thou alone knowesib the goal of the generations,

                        And Thou revealest not Thy mysteries to many.

            4 Thou makest known the multitude of the fire,

                        And Thou weighest the lightness of the wind.

            5 Thou explorest the limits of the heights,

                        And Thou scrutinisest the depths of the darkness.

            6 Thou carest for the number which pass away that they

                                                                        may be preserved,

                        And Thou preparest an abode for those that are to be.

            7 Thou rememberest the beginning which Thou hast made,

                        And the destruction that is to be Thou forgettest not.

            8 With nods of fear and indignation Thou givest command-

                                                                        ment to the flames,

                        And they change into spirits,2


            1 The translation, without line division, referred to above is that in

R. H. Charles, The Apocalypse of Baruch (1896). Since the above words

were written, Dr. Charles has published a revised translation with

division into parallel lines in The Apocrypha and Pseudepigrapha of the

Old Testament (Oxford, 1913), vol. ii. p. 504 f. In this later translation

Dr. Charles has adopted the punctuation in v. 14, given above ; its

correctness, indeed, becomes obvious so soon as the sustained parallel-

ism of the passage is recognised. Verse 2 is now divided by Dr. Charles

into six lines : the division into three, as above, shows the parallelism

more clearly.

            2 I suspect corruption in v. 8 a, b. In the original text " flames "

was probably a parallel term to " spirits " (cp. Ps. civ. 4), and not, as

in the present text of the versions, that which changes into spirits.

Moreover, the two lines are likely to have been more nearly equal to

one another in length : the inequality between them presents a striking

contrast to what is found in the rest of the poem.

                        INTRODUCTORY                            29


And with a word Thou quickenest that which was not,

            And with mighty power Thou .oldest that which has not

                                                                                    yet come.

9 Thou instructest created things in the understanding of


            And Thou makest wise the spheres so as to minister in

                                                                                    their orders.

10 Armies innumerable stand before Thee,

            And they minister in their orders quietly at Thy nod.

11 Hear Thy servant,

            And give ear to my petition.

12 For in a little time are we born,

            And in a little time do we return.

13 But with Thee, hours are as a time (?),

            And days as generations.

14 Be not therefore wroth with man; for he is nothing ;

            And take not account of our works; 15 for what are we?

For lo! by Thy gift do we come into the world,

            And we depart not of our own will.

16 For we said not to our parents, "Beget us,"

            And we sent not to Sheol, saying, "Receive us."

17 What, then, is our strength that we should bear Thy wrath,

            Or what are we that we should endure Thy judgment?

18 Protect us in Thy compassions,

            And in Thy mercy help us.


            The Apocalypse of Esdras (IV. Esdras) was

probably written shortly after A.D. 100, and

though it contains nothing quite so regular and

sustained as the passage just cited from the

Apocalypse of Baruch, a considerable number of

passages are printed both by Professor Gunkel

and Mr. Box 2 as poetry, and, some (e.g. viii.

20-30) at least, with good. reason.


            1 In E. Kautzsch, Die Apokryphen and Pseudepigraphen des AT.,

ii. 352-401 (cp. p. 349).

            2 G. H. Box, The Ezra-Apocalypse; and also in The Apocrypha and

Pseudepigrapha of the Old Testament (ed. R. H. Charles), ii. 542-624.



            Parallelism, then, certainly continued into the

second century A.D. to be a feature in Hebrew

poetry, or in Hebrew literature written in a form

differing from ordinary prose. Whether poetry

distinguished by the sustained use of parallelism

was still composed after the second century is

doubtful; but in this connexion two recently re-

covered documents may be very briefly referred to.


            1 Certainly no literary work that is at present generally admitted

to be later than the second century is marked by such sustained

parallelism as we find in parts of the Apocalypse of Baruch, or by any-

thing approaching it. But the Talmud contains a few snatches of

occasional poetry one or two of which, at least, are characterised by

parallelism and by something closely resembling rhythms found in the

Old Testament. The most pertinent example is that attributed in

Moed Katan 25 b to an elegist (xnrps) on the death of Hanin who is

described as hxyWn ybd hyntH, which is interpreted by Levy (Neuheb.

Worterbuch, ii. 83 a) as meaning that Hanin was a son-in-law of R.

Juda Nasi. The elegy alludes to the fact that Hanin died on the day

that his son was born. It runs:--

            vqbdn Nvgyv Nvww |  hnphn hgvtl hHmw

            xnynH dbx vtnynH tfb | hnxn vtHmw tfb

This may be rendered, tl;Lough the last lines are not free from ambiguity

(see Levy, loc. cit.) :

            Joy was turned into weariness,

               Gladness and sadness were united;

            When his gladness came, he sighed,

               When his favour came, he that was favoured, perished.

The parallelism is obvious; and the rhythm of the first distich is

3:3 (see below, p. 159 f.). Parallelism and rhythm are rather less con-

spicuous in another elegy cited at the same place, viz.:

            rmHk qydc lf | wxr vfynh Myrmt

            Mymyk tylyl Mywm lf | Mymyk tvlyl Mywn

The palm-trees shook their head

            Over the righteous that was as a palm-tree (cp. Ps. xcii. 13).

(So) let us turn night into day (i.e. weep unremittingly)

            Over him who turned night into day (in the study of the law).

Yet another elegy cited the same place contains the lines

            ryq ybvzx vWfy hm | tbhlw hlpn Myzrxb Mx

            If on the cedars the flame fell,

            What can the hyssops on the wall do?

                        INTRODUCTORY                31


            Dr. Charles1 finds a considerable element of

poetry in the fragments of a Zadokite work of

which the Hebrew text was first edited (with

translation and introduction) by Dr. Schechter2

in 1910. In the opinion of some this work is

considerably later than IV. Esdras; but Dr.

Charles has strong reasons for concluding that

it was written before A.D. 70. Be the date, how-

ever, what it may, except in quotations from the

Old Testament, parallelism in this work is not at

all conspicuous; whether, therefore, the passages

marked by Dr. Charles as possessing poetical

form actually do so, turns on matters which have

to be considered later. Happily, in this case the

question can be considered, not through transla-

tions merely, but with the original text before us.

            The Odes of Solomon, of which the Syriac text

was first edited by Dr. Rendel Harris3 in 1909,

were scarcely written before A.D. 70, and they

may belong to the second century A.D. ; in the


which recall, though the lines are longer, the ring of Ps. xi. 3. Two

similar distichs follow. A further example occurs in Hagigah 15 b

            vnybr jynpl dmf xl |   Htph rmw vlypx

            Even the keeper-of-the-door (of Gehenna)

                        Stood not his ground before thee, 0 our teacher.

As the sustained parallelism which is so characteristic of much of

the Old Testament and Jewish literature to the second century A.D.

appears to run back to origins in the popular poetry of the early

Hebrews, so parallelism seems to have maintained an existence for

some time in the occasional poetry of the later Jews, after it had

ceased to be employed in more formal literature.

            1 Fragments of a Zadokite work translated . . . 1912.

            2 In Documents of Jewish Sectaries, vol. i.

            3 The Odes and Psalms of Solomon published from the Syriac Version,

1909 (ed. 2, 1911).

32                    FORMS OF HEBREW POETRY


opinion of some they were written even later.

The original language of these Odes is still un-

determined. But some of them (e.g. v., vi.,

vii.) are strongly parallelistic in character, though

Dr. Harris refrained from distinguishing the

parallel members in his translation.

            It was long ago pointed out by Lowth that

parallelism can be retained almost unimpaired

in a translation; easier still, therefore, was it for

Jews to reproduce this feature in works written

in the first instance in some other language than

Hebrew ; and to some extent they did so. The

Book of Wisdom, which rests on no Hebrew

original, but was written, as it survives, in Greek,

is the best proof of this. It is possible that the

author of Wisdom attempted to imitate other

features of ancient Hebrew poetry as well as its

parallelism in his Greek work; but these are

questions that cannot be pursued now.

            There is no other considerable book originally

written in Greek which employs parallelism

throughout ; but it has been held with differing

degrees of conviction and consensus of opinion

that Tobit's prayer (Tob. xiii.), the Prayer of

Manasses, the Song of the Three Holy Children,

and the latter part of Baruch were written in

Greek, or at least, not in Hebrew; and a Hebrew

original for the Odes of Solomon was postulated

neither by their first editor, nor by many who

have followed him, though more recently Dr.

                        INTRODUCTORY                            33


Abbott1 has adduced some evidence which he

thinks points to such an original.

            The question of the original language of each

of these works might, perhaps, with advantage,

be reconsidered in connexion with the general

question of the extent to which parallelism was

adopted in Jewish writings not written in Hebrew.

We have on the one hand the clear example of

the use of parallelism in Wisdom, and on the

other the exceedingly slight use of parallelism,

for example, in the Sibylline oracles ; and we

may recall again in this connexion the avoidance

of parallelism in mediaeval Hebrew poetry. These

avoidances or absences of parallelism are certainly

worthy of attention in view of the ease with which

this feature of Hebrew poetry could have been

reproduced in Greek works, and even combined,

if necessary, with the use of Greek metres like the

hexameters of the Jewish Sibylline books. Was it

merely due to the fact that the one was writing

in Hebrew and the other in Greek, that the author

of the Apocalypse of Baruch in his loftier passages

employs the form of ancient Hebrew poetry,

whereas his contemporary, St. Paul, even in such

a passage as 1 Corinthians xiii.,2 avoids it ? Or

may we detect here the influences of different

schools or literary traditions?


            1 E. A. Abbott, Light on the Gospel from an Ancient Poet.

            2 See above, p. 26, n. 3.









                        CHAPTER II






































                                    CHAPTER II




THE literature of the Old Testament is divided

into two classes by the presence or absence of

what since Lowth has been known as paralle-

lismus membrorum, or parallelism. The occur-

rence of parallelism characterises the books of

Psalms, Job, Proverbs, Ecclesiastes (in part),

Lamentations, Canticles, the larger part of the

prophetical books, and certain songs and snatches

that are cited and a few other passages that occur

in the historical books. Absence of parallelism

characterises the remainder of the Old Testament,

i.e. the Pentateuch and the books of Joshua,

Judges, Samuel, Kings and Chronicles (with

slight exceptions in all these books as just in-

dicated), Ezra, Nehemiah, Esther, Ruth, and

part of the prophetical books, including most of

Ezekiel, the biographical parts of Jeremiah, the

book of Jonah (except the psalm in chapter ii.),

and some passages in most of the remaining

prophetical books. It had become customary to




distinguish these two divisions of Hebrew litera-

ture as poetry and prose respectively : parallelism

had come to be regarded as a mark of poetry, its

absence as a marls of prose; and by the application

of the same test the non-canonical literature of

the Jews from the second century B.C. to the

second century A.D. was likewise coming to be

distinguished into its prose and poetical elements.

The validity of parallelism as a test to dis-

tinguish between prose and poetry in Hebrew

literature might be, and has been either actually

or virtually, challenged on two grounds: (1)

that parallelism actually occurs in prose; and

(2) that parts of the Old Testament from which

parallelism is absent are metrical and, therefore,

poetical in form.

            Parallelism is not a feature peculiar to Hebrew

literature:1 it is characteristic of parts of Baby-

lonian literature, such as the Epics of Creation


            1 Nor even to Semitic literature. Many interesting illustrations

from folk-songs and English literature are given by Dr. G. A. Smith in

The Early Poetry of Israel, pp. 14-16. Yet in most of these there is

more simple repetition without variation of terms than is common in

Hebrew, and an even more conspicuous difference is the much less sus-

tained use of parallelism. In view of the great influence of the Old

Testament on English literature and the ease with which parallelism

can be used in any language (cp. p. 32 above), it is rather surprising

that parallelism, and even sustained parallelism, is not more conspicu-

ous in English. But abundant illustrations of this sustained use may

be found in the Finnish epic, The Kalevala, if Mr. Crawford's transla-

tion keeps in this respect at all close to the original, with which I have

no acquaintance. Even here there are differences, as for example in

the absence of the tendency, so marked in Hebrew, for parallelism to

produce distichs. I cite a sufficiently long passage to illustrate what is

a frequent, though not a constant, characteristic of the style of The

Kalevala :—

            PARALLELISM : A RESTATEMENT                    39


(the Enuma elis and others), the Gilgamesh epic

and the hymns to the gods.l It is as apparent

in translations from Babylonian as in the English

versions of the Psalms or the prophets ; as ex-

amples from Babylonian literature it may suffice to

cite the well-known opening lines of Enuma elis2--


            When above the heaven was not named,

                        And beneath the earth bore no name,

            And the primeval Apsu, the begetter of them,

                        And Mummu and, the mother of them all--


                        Listen, bride, to what I tell thee :

                        In thy home thou wert a jewel,

                        Wert thy father's pride and pleasure,

                        ‘Moonlight,’ did thy father call thee,

                        And thy mother called thee ‘Sunshine,’

                        ‘Sea-foam’ did thy brother call thee,

                        And thy sister called thee ‘Flower.’

                        When thou leavest home and kindred,

                        Goest to a second mother,

                        Often she will give thee censure,

                        Never treat thee as her daughter,

                        Rarely will she give thee counsel,

                        Never will she sound thy praises.

                        ‘Brush-wood,’ will the father call thee,

                        ‘Sledge of Rags,’ thy husband's brother,

                        ‘Flight of Stairs,’ thy stranger brother,

                        ‘Scare-crow,’ will the sister call thee,

                        Sister of thy blacksmith husband ;

                        Then wilt think of my good counsels,

                        Then wilt wish in tears and murmurs,

                        That as steam thou hadst ascended,

                        That as smoke thy soul had risen,

                        That as sparks thy life had vanished.

                        As a bird thou eanst not wander

                        From thy nest to circle homeward,

                        Canst not fall and die like leaflets,

                        As the sparks thou canst not perish,

                        Like the smoke thou canst not vanish."

                                    J. M. CRAWFORD, The Kalevala, i. 341, 2.


            1 A convenient collection of all of these (transliterated text and trans-

lation) will be found in R. W. Rogers, Cuneiform Parallels to the Old


            2 Cp. Rogers, pp. 3ff.



and these lines from a hymn to the god Sin1--


When Thy word in heaven is proclaimed, the Igigi prostrate


When Thy word on earth is proclaimed, the Anunaki kiss

                                                                                    the ground.

When Thy word on high travels like a storm-wind, food and

                                                                                    drink abound;

When Thy word on earth settles down, vegetation springs


Thy word makes fat stall and stable, and multiplies living


Thy word causes truth and righteousness to arise, that

                                                            men may speak the truth.


            Whether these passages are prose or poetry,

and whether, if poetry, they are such primarily

because of the presence of parallelism, turns on

the same considerations as the corresponding

questions with reference to parallelistic passages

in Hebrew: and further discussion of these must

be postponed.

            But parallelism is characteristic not only of

much in Babylonian and Hebrew literature: it

is characteristic also of much in Arabic literature,.

And the use of parallelism in Arabic literature is

such as to give some, at least apparent, justifica••

tion to the claim that parallelism is no true

differentia between prose and poetry ; for parallel--

ism in Arabic accompanies prose—prose, it is true,

of a particular kind, but at all events not poetry,

according to the general opinion of Arabian

grammarians and prosodists. Not only is paral-


            1 Cp. Rogers, pp. 144, 145.

            PARALLELISM: A RESTATEMENT         41


lelism present in much Arabic prose: it is

commonly absent from Arabic poetry, i.e. from

the rhymed and carefully regulated metrical

poetry of the Arabs. In illustration of this, two

passages may be cited from the Makamat of

Hariri. The translations here given are based

on Chenery's,l but I have modified them here

and there in order to bring out more clearly the

regularity of the parallelism in the original : for

the same reason I give the translation with line

divisions corresponding to the parallel members.

The first passage, which consists of part of the

opening address of Abu Zayd in the first Makamah,

is from the prose fabric of Hariri's work; the

second is one of the many metrical poems which

are wrought into the prose fabric. The parallel-

ism of the prose passage, as of innumerable other

passages which might equally well have served as

examples, is as regular and as sustained as that

of any passage in Hebrew or Babylonian litera-

ture, and indeed in some respects it is even more,

monotonously regular : it is complex too, for at

times there is a double parallelism—a parallelism

between the longer periods, the lines of the trans-

lation, and also between the parts of each of

these (the half lines of the translation). This

prose passage is as follows2:--


            1 T. Chenery, The Assemblies of Al Hariri, i. 109 f. and 192.

            2 In order that parallelism may be better studied I have hyphened

together word groups in English that correspond to a single word (com-

bined in some eases with inseparable particles) in Arabic. But I have



0-thou-reckless in petulance, trailing the garment of vanity!

            0-thou-headstrong in follies, turning-aside to idle-tales!

How long wilt-thou-persevere in thine error, and eat-sweetly-

                                                            of the pasture of thy wrong ?

And how far wilt-thou-be-extreme in thy pride, and not

                                                            abstain from thy wantonness ?

Thou provokest by-thy-rebellion the Master of thy forelock

            And thou goest-boldly in-the-foulness of thy behaviour

                                                            against the knower of thy secret;

And thou hidest-thyself from thy neighbour, but thou-art

                                                            in sight of thy watcher

And thou concealest-thyself from thy slave, but nothing

                                                            is-concealed from thy Ruler.

Thinkest thou that thy state will-profit-thee when thy

                                                            departure draweth--near?

Or-that thy wealth will-deliver-thee, when thy deeds


Or-that thy repentance will-suffice for thee when thy foot


Or-that thy kindred will-lean to thee in-the-day-that thy

                                                            judgment-place gathereth-thee?

How-is-it thou-hast-walked not in-the-high-road of thy

            guidance, and hastened the treatment of thy disease?

And blunted the edge of thine iniquity, and restrained

                                                            thyself—thy worst enemy.

Is-not death thy doom? What-then-is thy preparation?

And is-not-grey-hair thy warning? What-then-is thy


And is-not-in the grave's-niche thy sleeping-place? What-

                                                            then-is thy speech?

            And is-not-to God thy going? Who-then-is thy defender?

Oft the time hath-awakened-thee, but-thou-hast-set-thyself-


And admonition hath-drawn-thee, but-thou-past-strained-


And warnings have-been-manifested to thee, but-thou-hast-



generally omitted to hyphen the frequently recurring article, “of”

(before a genitive), pronouns and the copulative particle ("and")

none of these form separate words in Arabic.

            PARALLELISM : A RESTATEMENT                    43


And truth hath-been-established to thee, but-thou-hast-


And death hath-bid-thee-remember, but-thou-hast-sought-


And it-hath-been-in-thy-power to impart, and thou-

                                                            imparted'st not.


            The poem I select as an example is translated

by Chenery as follows:--


1 Say to him who riddles questions that I am the discloser

                                                            of the secret which he hides.

Know that the deceased, in whose case the law preferred

                        the brother of his spouse to the son of his father,

Was a man who, of his free consent, gave his son in marriage

                        to his own mother-in-law : nothing strange in it.

Then the son died, but she was already pregnant by him,

                                                and gave birth to a son like him :

And he was the son's son without dispute, and brother of

                        the grandfather's spouse without equivocation.

6 But the son of the true-born son is nearer to the grand-

            father, and takes precedence in the inheritance over

                                                                                    the brother;

And therefore when he died, the eighth of the inheritance

            was adjudged to the wife for her to take possession;

And the grandson, who was really her brother by her

                                                                mother, took the rest;

And the full brother was left out of the inheritance, and

                                    we say thou past only to bewail him.

This is my decision which every judge who judges will

                                                       pattern by, every lawyer.


            Nothing could be more prosaic than this last

passage : and the only approximation in it to

parallelism is line 5 ; nevertheless it is, so far as

form goes, a perfect poem in the original : the

rhymes are correct, and the well-known metrical

form called khafif is maintained throughout.



            So far, then, as Arabic literature is concerned,

it is an unquestionable fact that sustained and

regular parallelism is a frequent characteristic of

prose, while the absence of parallelism is frequently

characteristic of metrical poems. And yet this

is not of course the whole truth even in regard

to Arabic literature. Most literatures consist of

poetry and prose: and what in them is not

poetical in form is prose, and vice versa. But in

Arabic there are three forms of composition: (1)

nathr; (2) nazm, or si’r; (3) saj’. The usual

English equivalents for these three Arabic terms

are (1) prose, (2) poetry, (3) rhymed prose; but

"rhymed prose" is not, of course, a translation

of saj’: that word signifies primarily a cooing

noise such as is made by a pigeon; and its trans-

ferred use of a form of literary composition does

not, as the English equivalent suggests, represent

this form as a subdivision of prose. We should

perhaps do more justice to some Arabic discus-

sions or descriptions of saj’ by terming it in

English "unmetrical poetry";1 and in some

respects this " rhymed prose " or " unmetrical

poetry " is more sharply marked off from ordinary


            1 ”The oldest form of poetical speech was the saj'. Even after this

stage of poetical form had long been surpassed and the metrical schemes

had already been fully developed, the saj' ranked as a kind of poetical

expression. Otherwise his opponents would certainly never have called

Mohammed sa'ir (poet), for he never recited metrical poems, but only

spoke sentences of saj'. In a saying attributed to Mohammed in the

Tradition, too, it is said: ‘This poetry is saj'.’"—Goldziher, Abhand-

lungen zur arabischen Philologie, p. 59.

            PARALLELISM A RESTATEMENT                       45


prose than from the metrical poetry between

which and itself the simplest form of metrical

verse, termed rejez,l may be regarded as a transi-

tional style.

            To the Arabic saj’, as rhymed prose, Hebrew

literature has, indeed, little or nothing analogous

to show; to saj’ as unmetrical poetry possibly,

and certainly in the opinion of some writers it has

much. For example, if we disregard the rhyme,

such passages as that cited above from Hariri

have, in respect of parallelism of terms and the

structure of the corresponding clauses, much that

is similar alike in Hebrew psalms and Hebrew

prophecy. And to some of these we may return.

            At this point I raise this question with reference

to Hebrew, and a similar question might be raised

with reference to Babylonian literature : ought

we to recognise three forms of composition as in

Arabic, or two only as in most literatures ? Since

rhyme is so conspicuous in Arabic, and so incon-

spicuous in Hebrew, this may at first seem a

singularly ill-considered question : and yet it is

not ; for however prominent rhyme may be in

Arabic poetry, it is perfectly possible to think

the rhyme away without affecting the essential

form of Arabic poetry, or of the Hebrew mediaeval

poetry that was modelled on it. It would have

been as easy for an Arabic poet, had he wished


            1 " Fundamentally rejez is nothing but rhythmically disciplined

saj’." "Many Arabic prosodists do not admit that rejez possesses the

character of si’r."—Goldziher, ibid. pp. 76, 78.



it, as it was for Milton, to dispense with rhyme:

his poetry would have remained sufficiently dis-

tinguished from prose by its rigid obedience to

metrical laws. So, again, it is possible to think

away rhyme from the rhymed prose without

reducing that form of composition to plain prose;

the parallelism, and a certain balance of the

clauses, would still remain ; and as a matter of

fact much early Arabic parallelistic composition

existed from which regular rhyme was absent.1

            Had then the ancient Hebrew three forms of

composition—metrical poetry and plain prose,

and an intermediate type differing from poetry

by the absence of metre, and from prose by obedi-

ence to certain laws governing the mutual relations

between its clauses—a type for which we might

as makeshifts employ the terms unmetrical poetry

or parallelistic prose ?

            I am not going to answer that question im-

mediately, nor, perhaps, at all directly. But it

seems to me worth formulating, even if no certain

answer to it can be obtained. It may help to

keep possibilities before us : and, perhaps, also

to prevent a fruitless conflict over terms. In the

present discussion it is not of the first importance

to determine whether it is an abuse of language


            1 Goldziher (op. cit. pp. 62 ff.) argues that rhyme first began to be

employed in the formal public discourses or sermons (khutba) from t;he

third century of the Hejira onwards. " The rhetorical character of

such discourses in old time was concerned only with the parallelism of

which use was made " (p. 64).

            PARALLELISM: A RESTATEMENT                     47


to apply the term poetry to any part of Hebrew

literature that does not follow well-defined metrical

laws simply on the ground that it is marked by

parallelism; what is of importance is to deter-

mine if possible whether any parts of the Old

Testament are in the strictest sense of the term

metrical, and, alike whether that can be deter-

mined or not, to recognise the real distinction

between what is parallelistic and what is not, to

determine so far as possible the laws of this

parallelism, and to recognise all parts of the

ancient Hebrew literature that are distinguished

by parallelism as related to one another in respect

of form.

            It is because I approach the question thus that

I treat of parallelism before metre: parallelism

is unmistakable, metre in Hebrew literature is

obscure: the laws of Hebrew metre have been

and are matters of dispute, and at times the very

existence of metre in the Old Testament has been

questioned. But let us suppose that Sievers, to

whose almost overwhelming contributions1 to

this subject we owe so much, whatever our final

judgment as to some even of his main conclusions

may be, is right in detecting metre not only in

what have commonly been regarded as the

poetical parts of the Old Testament, but also

throughout such books as Samuel and Genesis;2


            1 See below, pp. 143-154.

            2 Ed. Sievers, Metrische Studien, ii. "Die hebraische Genesis," and

Metrische Studien, iii. “Samuel.”



even then the importance and value of the

question formulated above remains. It is true

that some questions may require resetting : if

Samuel and Genesis are metrical throughout, if

even the genealogies in Genesis v. and xxxvi. are,

so fare as form goes, no less certainly poems than

the very prosaic Arabic poem cited above, it will

become less a question whether the Old Testa-

ment, contains metrical poems than whether it

contains any plain prose at all. But the distinc-

tion between what is parallelism and what is not

will remain as before: we shall still have to dis-

tinguish between parallelistic prose and prose

that is not parallelistic, or, if the entire Old Testa-

ment be metrical, between parallelistic and non-

parallelistic poetry.

            The general description and the fundamental

analysis of parallelism as given by Lowth, and

adopted by innumerable subsequent writers, are

so well known that they need not be referred to

at length here: nor will it be necessary to give

illustrations of the familiar types of parallelism

known as synonymous and antithetic. But I

may recall Lowth's own general statement in the

Preliminary Dissertation (Isaiah, ed. 3, p. xiv):

"The correspondence of one verse, or line, with

another, I call parallelism. When a proposition

is delivered, and a second is subjoined to it, or

drawn under it, equivalent, or contrasted with it,

in sense; or similar to it in the form of gram-

PARALLELISM: A RESTATEMENT                     49


matical construction; these I call parallel lines,

and the words or phrases, answering one to

another in the corresponding lines, parallel terms.

Parallel lines may be reduced to three sorts:

parallels synonymous, parallels antithetic and

parallels synthetic.”

            The vulnerable point in Lowth's exposition of

parallelism as the law of Hebrew poetry lies in

what he found it necessary to comprehend under

the term synthetic parallelism : his examples

include, indeed, many couplets to which the term

parallelism can with complete propriety be ap-

plied ; in such couplets the second line repeats

by means of one or more synonymous terms part

of the sense of the first; and by means of one or

more other terms adds something fresh, to which

nothing in the first line is parallel. In virtue of

the presence of some parallel terms such lines

may be called parallel, and in virtue of the pre-

sence of some non-parallel terms they may be

called synthetic, or in full the lines may be termed

synthetic parallels, and the relation between them

synthetic parallelism; but more convenient terms

for such lines, which are of very frequent occur-

rence,1 and for the relation between them, would

be incomplete parallels and incomplete parallelism.

In any case, term them as we will, such examples

as these are in reality not distinct from, but mere

subdivisions of synonymous or antithetic parallel-


            1 Many examples are cited below: see pp. 72-82.



ism as the case may be. On the other hand there

are other examples of what Lowth called syn-

thetic parallelism in which no term in the second

line is parallel to any term in the first, but in

which the second line consists entirely of what is

fresh and additional to the first; and in some of

these examples the two lines are not even parallel

to one another by the correspondence of similar

grammatical terms. Two such lines as these

may certainly be called synthetic, but they are

parallel to one another merely in the way that

the continuation of the same straight line is

parallel to its beginning; whereas synonymous

and antithetic parallelisms, even of the incomplete

kind, do really correspond to two separate and,

strictly speaking, parallel lines. Now, if the

term parallelism, even though it be qualified by

prefixing the adjective synthetic, be applied to

lines which, though synthetically related to one

another, are connected by no parallelism of terms

or sense, as well as to lines which are connected

by parallelism of terms or sense, then this term,

(synthetic) parallelism, will really conceal an all-

important difference under a mere semblance of

similarity. And, indeed, Lowth himself seems

to have been at least half-conscious that he was

making the term synthetic parallelism cover too

much: for he admits that “the variety in the

form of this synthetic parallelism is very great, and

the degrees of resemblance almost infinite; so that

PARALLELISM: A RESTATEMENT                     51


sometimes the scheme of the parallelism is very

subtile and obscure” (Lectures, ii. 52); he very

fairly adds in illustration a really test couplet, viz.


            I also have anointed my king on Sion,

                        The mountain of my sanctity (Psa. ii. 6).1


He perceives, though he does not dwell on the

point, that this couplet marks zero among " the

degrees of resemblance almost infinite"; for

when he says, "the general form and nature of

the Psalm requires that it should be divided into

two parts or versicles; as if it were,


‘I also have anointed my king ;

      I have anointed him in Sion, the mountain of my sanctity,'”


he supplies, by repeating the words, "I have

anointed," the one and only point of resemblance

that exists between the two lines in his own

reconstruction of a couplet which, in its true

original form, is really distinguished by the entire

absence of parallelism between its lines. As in

this instance, so often, the use of the term syn-

thetic parallelism has served to conceal the fact

that couplets of lines entirely non-parallel may

occur in poems in which most of the couplets are

parallels, and in which the "general form and

nature " of the poem suggest a division of the

synthetic but non-parallel elements" into two

parts or versicles."


            1 The verse is so divided by Lowth; for reasons which will appear

Iater it should rather be divided:

            I also have anointed my king,

                        On Sion, the mountain of my holiness.




            Not only did. Lowth thus experience some

doubt whether parallelism as analysed by himself

was the one law of Hebrew poetry, but he ex-

pressly concludes his discussion of these " subtile

and obscure " examples of synthetic parallelism

with a suggestion that behind and accompanying

parallelism there may be some metrical principle,

though he judged that principle undiscovered and

probably undiscoverable.

            In spite of the general soundness of Lowth's

exposition'of parallelism, then, there is, perhaps,

sufficient reason for a restatement ; and that I

shall now attempt.

            The extreme simplicity of Hebrew narrative

has often been pointed out: the principle of

attaching clause to clause by means of the "waw

conversive" construction allows the narrative to

flow on often for long periods uninterrupted, and,

so to speak, in one continuous straight line. Now

and again, and in certain cases more often, the

line of successive events is broken to admit of

some circumstance being described; but the same

single line is quickly resumed. An excellent

example of this is found in Genesis i.: with the

exception of verse 2, which describes the condi-

tions existing at the time of the creative act

mentioned in verse 1, the narrative runs on in a

single continuous line down to verse 26; thus

1          2          3                                              26

__        ____    ____________________


PARALLELISM: A RESTATEMENT                     53


            The continuity of a single line of narrative is

in parts of Genesis ii. nearly as conspicuous: as

to other parts of Genesis ii. something will have

to be said later.1 But if we turn to certain other

descriptions of creation elsewhere in the Old

Testament, we immediately discern a difference.

Thus we read in Psalm xxxiii. 6, 7, 9:


By the word of Yahweh the heavens were made,

            And by the breath of his mouth all their host.

He gathered as into a flask the waters of the sea,

            He put into treasure-houses the deeps.

For he spake and it came to pass,

            He commanded and it stood sure;


and in Isaiah xlv. 12 the words of Yahweh run

as follows:--


I made the earth,

            And man upon it I created ;

My hands stretched out the heavens,

            And all their host I commanded.

And again in Proverbs viii. 24-2 9 creation is

            described in a series of subordinate periods :

When there were no depths . . .

            When there were no fountains abounding with water ;

Before the mountains were settled,

            Before the hills . . .

While as yet he had not made the earth, nor the fields,

            Nor the beginning of the dust of the world ;

When he established the heavens . . .

            When he set a circle upon the face of the deep ;

When he made firm the skies above,

            When the fountains of the deep became strong,

When he gave to the sea its bound,

            That the waters should not transgress his commandment,

            When he marked out the foundations of the earth.


            1 See pp. 221 f.

54                    FORMS OF HEBREW POETRY


            Now whether, as Sievers maintains, Genesis i.

is as strictly metrical as Psalms, Proverbs or

Isaiah xl.-lxvi., or whether, as has been commonly

assumed, Genesis i. is plain, unadorned and un-

metrical prose, between Genesis i. on the one

hand and the passages just cited from Psalm

xxxiii., Isaiah xlv. and Proverbs viii. there are

these differences: (1) whereas Genesis i. is carried

along a single line of narrative, the other passages

are, in the main at least, carried forward along

two lines, parallel to one another in respect of

their meaning, and of the terms in which that

meaning is expressed; (2) whereas Genesis i.

consists in the main of connected clauses so that

the whole may be represented by a single line

rarely broken, the other passages consist of a

number of independent clauses or sentences, so

that they must be represented by lines constantly

broken, and at fairly regular intervals, thus--

            ===                 ===                 ===

            Stated otherwise, as contrasted with the

simpler style of Genesis i., these other passages are

characterised by the independence of their succes-

sive clauses or short sentences, and the repetition

of the same thought or statement by means of

corresponding terms in successive short clauses or

sections. Where repetition and what may be

termed parallelism in its fullest and strictest sense

occur, a constant breaking of the line of narrative

or statement is the necessary consequence: a


PARALLELISM: A RESTATEMENT                     55


thought is expressed, or a statement made, but

the writer, instead of proceeding at once to ex-

press the natural sequel to his thought or the next

statement, breaks off and harks back in order to

repeat in a different form the thought or state-

ment which he has already expressed, and only

after this break and repetition pursues the line of

his thought or statement; that is to say, one line

is, as it were, forsaken to pursue the parallel line

up to a corresponding point, and then after the

break the former line is resumed. But the break

in the line and the independence of clauses may

occur even where there is no repetition of thought

or correspondence of terms; just as breaks

necessarily occur occasionally in such simple

narratives as that of Genesis i. The differences

between the two styles here shade off into one

another; and everything ultimately depends on

the frequency and regularity with which the

breaks occur. Where the breaks occur with as

much regularity as when the successive clauses

are parallel to one another, we may, even though

parallelisms of terms or thought between the

clauses are absent, term the style parallelistic,

as preserving one of the necessary consequences

of actual parallelism.

            But not only is the question whether a passage

belongs to the one style or the other, so far as it

depends on the recurrence of breaks and the con-

sequent independence of the clauses, one of degree;




the question whether two such independent lines

are correspondent or parallel to one another is

also at times a question both of degree and of

exact interpretation. To return to the passages

already cited; when the Psalmist writes :


            He gathered as into a flask the waters of the sea,


and then adds,


            He put into treasure houses the deeps,


it is clear that at the end of the first line he breaks

the straight line of continuous statement: the

second line adds nothing to the bare sense, and

it carries the writer no further forward than the

first; the two sentences thus correspond strictly

to two equal and parallel lines: where the first

begins the second also begins, and where the first

ends there also the second ends: each line records

exactly the same fact and the same amount of

fact by means of different but synonymous terms.

And the same is true of the two lines,


            For he spake and it came to pass,

               He commanded and it stood sure.


We can without difficulty and with perfect pro-.

priety represent these two couplets thus

                        ===      ===

But what are we to say of,


            I made the earth,

                        And man upon it I created ?


This is certainly not the simplest form of putting

the thought to be expressed : the terms " made "


PARALLELISM: A RESTATEMENT                     57


and "created" are synonymous, and the whole

thought could have been fully expressed in the

briefer form, "I made the earth, and man upon

it."  But have we, even so, completely delimited

substance and form, the thought to be expressed

and the art used in its expression ? Probably

not ; the writer continues:


            My hands stretched out the heavens,

                        And all their host I commanded.


Here we cannot simply drop a term as in the

previous lines and leave the sense unimpaired;

but the correspondence of thought between the

two sets of statements may yield a clue to the

essential thought of the whole; as the first two

lines mean no more than this: I created the earth

and its inhabitants; so the second means simply

this: I created the heavens and their inhabitants.

But have we even yet determined the funda-

mental thought of the passage? Did the writer

really mean to express two distinct thoughts in

each set of lines? Was he thinking of the crea-

tion of man as something independent of the

creation of the earth? Did he mean to refer

first to one creative act and then to a second and

independent creative act? Or did he regard

the creation of man as part of the creation of

the earth, so that his lines are really parallel state-

ments, a parallelism, to wit, of the part with the

whole, and not successive statements? This

seems to me most probable; his thought was:




Yahweh created the heavens and the earth; but

instead of expressing this in its simplest form by

a sentence that would properly be represented by

a single continuous line, he has artistically ex-

pressed it in a form that may once again, though

with less complete propriety, perhaps, than in the

case of the couplet from Psalm xxxiii., be ex-

pressed by two groups of parallel and broken


                =====       =====


            f the thought of man and the host of heaven

had a greater independence than this view recog-

nises, we must still treat the statement (which is

not, like Genesis i., the continuous statement of

successive acts) not as a continuous line, but as

a line broken at very regular intervals, thus

though, if we wished diagrammatically to bring

out the similarity in the verbal cast or grammati-

cal build of the clauses rather than the independ-

ence of the thought, we might still adopt the


                        ======         =======

            efore leaving this diagrammatic description

I merely add, without illustrating the statement,

that a poem rarely proceeds far along two parallel

lines each broken at the same regular intervals,


======  ====== ===== ====== ====== =====

Either the two lines are broken at different points,


PARALLELISM: A RESTATEMENT                     59


or one is for the time being followed to the neglect

of the other, thus—

===== ===== --===  ==--   -----    -----   =====


            I pass now by a different method to a more

detailed examination of parallel lines, and of the

degree and character of the correspondence

between them. Irrespective of particles a line

or section to which another line or section ap-

proximately corresponds, consists of two, three,

four, five or six words, very seldom of more.

Complete parallelism may be said to exist when

every single term in one line is parallel to a term

in the other, or when at least every term or

group of terms in one line is paralleled by a corre-

sponding term or group of terms in the other.

Incomplete parallelism exists when only some of

the terms in each of two corresponding lines are

parallel to one another, while the remaining

terms express something which is stated once

only in the two lines. Incomplete parallelism

is far more frequent than complete parallelism.

Both complete parallelism and incomplete paral-

lelism admit of many varieties ; and this great

variety and elasticity of parallelism may perhaps

best be studied by means of symbols, even though

it is difficult to reduce all the phenomena to

rigidly constant and unambiguous symbolic

formul. I have already elsewhere1 suggested

that the varieties of parallelism may be con-


            1 Isaiah ("International Critical Comm."), p. lxvi.




veniently described by denoting the terms in the

first line by letters—a . b . c, etc.—and those in

the second line by the differentiated letters—

a' . b' . c', where the terms, without being identical

(in which case a . b . c would be used for the

second line as well as for the first), correspond,

or by fresh letters—d . e . f, where fresh terms

corresponding to nothing in the first line occur.

The simplest form of complete parallelism is

represented by           a .  b

                                    a'.   b'.

here each line consists of two terms each of which

corresponds to a term in the corresponding posi-

tion in the other line. Examples are

                        bqfyb MqlHx

           lxrWyb Mcypxv


            I-will-divide-them1  in-Jacob,

                        And-I-will-scatter-them in-Israel.—Gen. xlix. 7c.d.


                        tvnlHh-Nm Hygwm

           MykrHh-Nm Cycm  

            He-looketh-in at-the-windows,

                        He-glanceth through-the-lattice.


Cant. ii. 9 (the same chapter contains several other examples).

                        fvmwm ytyvfn

           tvxrm ytlhbn

            I-am-bent-with-pain at-what-I-hear,

                        I-am-dismayed at-what-I-see.—Isa. xxi. 3.


            1 Where the suffix in one line corresponds to a noun in the other it

may sometimes be convenient to represent the suffix by an independent

symbol. If both suffixes were so represented here the scheme would be

                                                a .b .c

                                                a'.b .c'.


PARALLELISM: A RESTATEMENT                     61


                        Mhyfwp vbr yk

           Mhytvbwm vmcf

            For their-transgressions are-many,

                        Their-backturnings are-increased.—Jer. v. 6.

            Hear Thy-servant,

                        And-give-ear-to my-petition.—Apoc. Bar. xlviii. 12.


            Complete parallelism between lines each con-

taining three terms will be represented by

                                    a .  b .  c

                                    a' . b' . c'

Examples are--

                        Nyym Mynyf ylylkH

           blHm Mynw Nblv

                        Red-are his-eyes with-wine,

                           And-white-are his-teeth with-milk.—Gen. xlix. 12.


                        vdbxy hvlx tmwnm

           vlky vpx Hvrmv

                        By-the-breath of-God they-perish,

                           And-by-the-blast of-his-anger are-they-consumed. —Job.

iv. 9.

                        vHlmn Nwfk Mymw-yk

           hlbt dgbk Crxhv

                        For the-heavens like-smoke shall-vanish-away (?),

                           And-the-earth like-a-garment shall-wax-old.—Isa. li. 6.


            More frequent than the fundamental scheme

as given above and just illustrated are variations

upon it, of which examples will be given below.

            Complete parallelism of lines with four terms

each, the terms being symmetrically arranged,

will be represented by


                                    a . b .  c . d

                                    a'. b' .  c'. d'




An example is--

                        hmH bywy jr hnfm

           Jx hlfy bcf rbdv

            A-soft answer turneth-away wrath,

                But-a-grievous word stirreth-up anger.--Prov.:xv. 1.


            This scheme occurs not infrequently in anti-

thetic proverbs, and Proverbs xv. contains several

other examples; but it is rare elsewhere. Varia-

tions on this scheme also will be given below.

            Where the parallel sections consist of more

than four terms, and sometimes when they con-

tain as few as four terms, each section tends to

break up into two of those independent clauses

which we have seen to be in part the necessary

consequence of parallelism, and in part a common,

even when not a necessary, accompaniment of

the style distinguished from simple narrative.

For example, Isaiah xlix. 2 is one of the nearest

approximations to the scheme,


                        a .  b .  c . d .  e . f  

                        a' . b' . c' . d' . e' . f'


but here the last two terms in each section stand

independent of the foregoing ; thus:

And-he-made my-mouth as-a-sharp sword : in-the-shadow

                                                            of-his-hand he-hid-me;

And-he-made-me1 into-a-polished arrow: in-his-quiver he-



            1 The suffix me (b') is here parallel to the independent term my

mouth (b); and so is the suffix his in his quiver to the independent term

his hand: in this case, however, I have represented shadow of his hand

under the single symbol (e).

            PARALLELISM : A RESTATEMENT                    63


Such a combination of clauses is commonly

termed "alternate parallelism" and is said to

consist of four lines, of which the third is parallel

to the first and the fourth to the second. This

may be a convenient description: but the main

point is that, within the main independent

sections indicated by the parallelism, other

almost equally independent breaks giving rise

to subordinate independent clauses occur. This

fact is emphasised in many specimens of Arabic

"rhymed prose"; in the passage already cited

on pp. 42 f. from Hariri, almost all the parallel

sections fall into two independent clauses; and

it is these independent, but, from the point of

view of the parallelism, subordinate, sections that

rhyme with one another ; that is to say, similarity

of rhyme connects, while emphasising their dis-

tinction, the shorter independent clauses which

are commonly not parallel to one another, and

change of rhyme marks off the well-defined longer

sections which are regularly parallel to one

another. It is interesting to observe that in the

lines cited from Isaiah xlix. it is the entire parallel

periods and not the subsections that rhyme with

one another, though in view of the irregular use

of rhyme in Hebrew this may be a mere accident-

            ynixAybHh vdy lcb hdH brHk yp Mwyv

     ynirAytsh vtpwxb rvrb CHl ynmywyv

In the illustrations of parallelism which have



been given so far not only has there been com-

plete correspondence, term by term, between the

parallel lines, but each corresponding term in

the second line has occurred in the exactly corre-

sponding position in the second line. But in any

considerable passage Hebrew writers introduce

in various ways great variety of effect, a far

greater variety, I believe, than was commonly

sought or obtained by Arabic writers. These

varieties of parallelism can be readily and con-

veniently shown by a use such as I have suggested

of symbols. I proceed to classify and illustrate

some of the chief classes of variations on the

fundamental schemes which have been already

described and illustrated.



            Variety is attained by varying the position of

the corresponding terms in the two lines.

            In the simplest form of parallelism, which

consists of lines containing two terms only, only

one variation is possible from the scheme,


                                    a . b

                                    a' .b'

of which several illustrations have already been

given. This of course is

                                    a .  b

                                    b' . a'

and this variation occurs very frequently, e.g.—

            PARALLELISM: A RESTATEMENT         65


                        Jskk hnwqbt Mx

           hnwpHt MynmFmkv

            If thou-seek-her as-silver.

               And-as-for-hid-treasures search-for-her.—Prov. ii. 4.


                        hdWh yxct-lx

           yklt-lx jrdbv

            Go-not-forth into-the-field,

                And-by-the-way walk-not.—Jer. vi. 25.


Further examples will be found, for example,

in Deuteronomy xxxii. 16, xxxiii. 9 d, e.

            As the number of terms increases the greater

becomes the possibility of variety and the number

of actual variations; thus


                                    a .  b .  c

                                    a' . b' . c'

can alternate with

                                    a  . b . c

                                    a' . c' . b'


or any of the other four possible permutations.

Of the variation just given, Proverbs ii. 2 is an



                        jnzx hmkHl bywqhl

           Hnvbtl jbl hFt

            So-that-thou-incline unto-wisdom thine-ear,

                 (And-) apply thine-heart to-understanding.


The same variation of order, but with the repeti-

tion instead of a variation of the second term of



the first line at the end of the second line (-i.e.

b instead of b'), occurs in Job xxxii. 17


                        yqlH ynx-Jx hnfx

           ynx-Jx yfd hvHx

            Will-answer I also my-part,

                 Will-declare my-knowledge I also.


            An example may be found in Deuteronomy

xxxii. 30 a, b of

                                    a  . b . c

                                    b' . a' . c'

                        Jlx dHx Jdry hkyx

           hbbr vsyny Mynwv

            How should one pursue a-thousand,

                 Or-two put-to-flight ten-thousand.


The same poem also contains four examples

(Deuteronomy xxxii. 3, 18, 23, 38) of the scheme


                                    a  . b  . c

                                    c' . a' . b'


It may suffice to cite v. 18 (reading hwt for



                        hwt ddly rvc

           jllHm lx Hkwtv

            The rock that-bare-thee thou-wast-unmindful-of,

                 And-forgattest the God that-gave-thee-birth.


            Another example of this scheme may be found

in Proverbs v. 5.

            The tendency in poetry to give the verb its

normal (prose) position at the beginning of the

first line, but, in order to gain variety, to throw

            PARALLELISM: A RESTATEMENT         67


the verb to the end of the second line,1 renders

the two remaining variations of the fundamental

scheme, viz.--

                                    a  . b  . c

                                    b' . c' . a'


                                    a  . b  . c

                                    c' . b' . a'

very frequent, though of course both of these

schemes may also arise from other causes.2

Examples of the former of the two schemes just

given are--

                        rfym hyrx Mkh Nk-lf

           Mddwy tvbrf bxz

            Therefore shall-slay-them a-lion out-of-the-forest,

                A-wolf of-the-steppes shall-spoil-them.—Jer. v. 6.


                        jlm ynpl rdhtt-lx

           dmft-lx Mylvdg Mvqmbv

            Glorify-not-thyself in-the-presence of-the-king,

                 And-in-the-place of-great-men stand-not.—Prov. xxv. 6.


            Four further examples may be found in

Proverbs ii. 5, 8, 10, 20. See also e.g. Job iii.

6 b, c; Amos v. 23; Isaiah xi. 6 a, b, lx. 16 a, b;

Judith xvi. 10 (the last couplet in the passage

cited above, p. 25).


            1 The alternative of throwing the verb to the end of the first line,

and giving it the normal (prose) position in the second line, thus bringing

the two verbs together, is much less frequent. But a good example of

this is Deut. xxxii. 38 : see also vv. 3 and 18 in the same chapter.

            2 As e.g. in Job iv. 17.



Examples of

                                    a  . b . c

                                    c' . b' . a'


                        Mym vlfwb ddm-ym

           Nkt trzb Mymwv

            Who hath-measured with-the-hollow-of-his-hand the waters,

                 Or-the-heavens with-a-span hath-regulated?—Isa. xl. 12.


                        fbw Jymsx vxlmyv

           vcrpy Jybqy wvrytv

            That thy-barns may-be-filled-with plenty

                 And-that With-new-wine thy-vats may-overflow.

                                                                                    —Prov. iii. 10.


See also e.g. Isaiah xl. 26 c, d, 27 c, d; Amos v. 7;

Psalm iii. 8 c, d.

            The possible variations on

                                    a . b . c . d

                                    a'. b' . c'. d'

are of course much more numerous ; the actual

examples are far fewer, partly because complete

parallelism over these longer periods is much

rarer, partly because these parallelisms in four

terms occur particularly in Proverbs, and proverbs,

being complete in themselves, do not call for the

variety which is naturally enough desired in a

long continuous passage. It may suffice to refer

to one variation : when the first line begins with

a verb and its object, immediately following, is

expressed by an independent term, and the desire

for variety throws the corresponding clause to

            PARALLELISM: A RESTATEMENT            69


the end of the second line, the scheme naturally

produced is

                                    a  . b . c  .d

                                    c' . d'  .a' .b'

as for example in

                                    vyp Fbwb Crf1 hkhv

                fwr tymy vytqw Hvrbv

            And-he-shall-smite the-violent1 with-the-rod of-his-mouth,

                And-with-the-breath of-his-lips shall-he-slay the-wicked.

                                                                                                —Isa. xi. 4.


            Another way of obtaining variety is to use in

the second line two or more terms which, taken

together, are parallel in sense to a corresponding

number of terms in the first line, though the

separate terms of the one combination are not

parallel to the separate terms of the other com-

bination. In its extreme form parallelism of this

variety consists of two entire lines completely

parallel in sense but with no two terms taken

separately parallel to one another.2 Denoting

correspondence as before by a . a', etc., and the

number of terms above one in which particular

corresponding ideas are expressed by a figure

attached to the letters, the kind of schemes that

occur are

                                    a2 . b

                                    a'2 . b'


            1 Reading Crf for Crx, the earth.

            2 See e.g. Gen. xlix. 15 c, d, 20 ; Ps. xxi. 6 ; Job iii. 10, 23, iv. 14.



For example

                        ylvq Nfmw hlcv hdf

           ytrmx hnzxh jml ywb

            Adah and-Sillah, hear my-voice,

                Ye-wives of-Lamech give-ear-to my-word.—Gen. iv. 23.


            Here, too, further variety may be obtained

by varying the position of the corresponding

terms or groups of terms, so that such schemes


                                    a   . b2

                                    b'2 . a'


arise; an example of this is Proverbs ii. 17,

                        hyrvfn Jvlx tbzfh

           hHkw hyhlx tyrb txv

            Who-forsaketh the-friend of-her-youth,

                And the-covenant of-her-God forgetteth.


            And another very effective variation arises

when what is expressed by two terms in the first

line is expressed by one in the second line, which

in turn has two other terms corresponding to one

in the first: one such variation is

                                    a2 . b

                                    a'  . b'2

which is exemplified by Genesis xlix. 24,


                        vrwq Ntyxb bwtv

           vydy yfrz vzpyv

            And-his-bow abode firm,

                 And-the-arms of-his-hands were-agile--


where the two words Ntyxb bwtv, abode firm, taken

            PARALLELISM: A RESTATEMENT                     71


together are parallel to vzotv, were agile, and the

single term vtwq, his-bow, to the two terms yfrz.

vydy, the-arms of-his-hands, taken together.

            An example of

                                    a . b . c2

                                    a . c' . b'2

is afforded by Job iii. 17,

                        zgr vldH Mfwr Mw

           Hk yfygy vHvny Mwv

where vHvny, are-at-rest, corresponds to to zgr vldH,

cease from raging, and the single term wicked to

the phrase Hk yfygy, which is compound in Hebrew,

though it is represented by the single word weary

in E.V.

            Once more in Deuteronomy xxxii. 11,

                        vhHqy vypnk wrpy

           vtrbx-lf vhxwy

            He-spread-out his-wings, he-took-him,

                He-lifted-him-up upon-his-pinions,


the single term vtrbx-lf, upon-his-pinions, at the

end of the second line is parallel to the two terms

vypnk wrpy, he-spread-out his-wings, at the beginning

of the first line, taken together, and the scheme is


                                    a2 . b

                                    b' . a'

            Further examples of some of these or similar

schemes will be found in Deuteronomy xxxii.

22 c, d, 35 c, d; Psalms ii. 2 a, b, 9, lxviii. 10;



Proverbs xv. 9; Job iii. 25, iv. 4, xxxiii. 11;

Canticles ii. 3 c, d, 12.

            Occasionally one or other of the compound

parallel phrases is interrupted by the insertion

of another parallel term in the midst of it ; so,

for example, in Psalm vi. 6,

                        jrcz tvmb Nyx yk

           jl hdvy ym lvxwb

            For there-is in-death no-remembrance-of-thee;

                In-Sheol who shall-praise thee?

death and Sheol are parallel terms, and the phrase

there is no remembrance of thee to the interrogative

phrase, which is equivalent to a negative state-

ment, who shall praise thee? But in the first

line the parallel term is inserted bet1 Teen the two

parts of the parallel phrase.



            The third main method of introducing variety

into parallelism and avoiding the monotonous

repetition of the same scheme consists in the adop-

tion of various forms of incomplete parallelism.

            The variety of effect rendered possible by this

method is immense, except in the shortest

parallels consisting of two terms only : with

these the fundamental variations are reduced

to two, viz.—

            PARALLELISM: A RESTATEMENT         73



                                    a  . b

                                    a' . c


            Examples of these are-

                        Mykrb ynvmdq fvdm

           qnyx-yk Mydw hmv

            Wherefore did-the-knees receive-me,

               And-why the-breasts that I-should-suck (Job iii. 12),


                        yntqzHh hrc

           hdlvyk lyH

            Anguish hath-seized-me,

               Pangs as-of-a-woman-in-travail (Jer. vi. 24),


unless we prefer to treat the former of these

examples on the ground of the differentiation of

the interrogative particles as an example of


                        a . b . c

                        a'. c' . d


and the latter example as

                                    a . b


            The latter kind of ambiguity frequently arises.

            Further variety is obtained when variations

corresponding to those illustrated under I. and II.

are combined with incomplete parallelism : this

frequently happens, especially when one at least

of the parallel members contains more than two

terms. But before giving illustrations of such

variations it will be convenient to point out that

74                    FORMS OF HEBREW POETRY


incomplete parallelisms fall into two broad classes

which may be distinguished as incomplete parallel-

ism with compensation and incomplete parallelism

without compensation. If one line contains a

given number of terms and another line a smaller

number of terms, the parallelism is generally1

incomplete; such incomplete parallelism may

be termed incomplete parallelism without com-

pensation; but if the two lines contain the same

number of terms, though only some of the terms

in the two lines are parallel, the lines may be said

to constitute incomplete parallelism with com-

pensation. Thus such schemes as


                        a  . b . c

                        a' . b'


                        a . b . c


are incomplete without compensation ; whereas

such schemes as

                        a . b . c

                        a' . d .c'

are incomplete parallelism with compensation.


            1 Not invariably; for such schemes as

                                                a2 . b

                                                a' . b'

give to the two lines an unequal number of terms, and yet the parallelism

may be said to be complete. See e.g. Lam. ii. 11, cited below, p. 97.

            PARALLELISM: A RESTATEMENT         75


I now give illustrations of different schemes

of both types.



            Incomplete parallelism without compensation.

                        hnwxrbk jyFpw hbywxv

                hlHtbk jycfyv

            I-will-restore thy-judges as-at-the-first,

                 And-thy-counsellors as-at-the-beginning (Isa. i. 26),


is an example of

                                                a . b . c

                                                     b' . c'


and so are Proverbs ii. 18; Canticles ii. 1, 14;

Numbers xxiii. 19' c, d, 24 a, b, xxiv. 5 a, b; Psalm

vi. 2; Deuteronomy xxxii. 7 c, d, 21 a, b, 34.1

                        jrzfb Mymwb bkr

           MyqHw vtvxgbv

            Who-rideth through-the-heavens as-thy-help,

                 And-in-his-dignity through-the-skies (Deut. xxxiii. 26),


            1 A further example of this scheme occurs in the present text of

Hos. vii. 1--

                        Nvrmw tvfrv | Myrpx Nvf hlgnv

                        Revealed are the iniquity of Ephraim

                        And the wickedness of Samaria.

On the second of these lines Harper ("International Crit. Comm.")

remarks : " Here a word is needed to complete the parallelism as well

as the metre." But this is incorrectly put, unless it can be shown that

incomplete parallelism is impossible, or improbable in this connexion ;

and this cannot be done in view of another case of incomplete parallel-

ism (a . b . c a' . c') in v. 3, which Harper retains. Since the line

quoted above and v. 3 are possibly not metrically identical (v. 3 being

perhaps 3 : 3), a metrical consideration in favour of supplying a word

in v. 1 may survive ; but the argument from parallelism is invalid.

76                    FORMS OF HEBREW POETRY


is an example of

                                    a . b . c

                                        c' . b'

and so is Isaiah xlii. 23 a, b.

                        yfcpl ytgrh wyx

           ytrbHl dlyv

            A man have I slain for wounding me,

                  And a youth for bruising me (Gen. iv. 23),

is an example of

                                    a .b . c

                                    a' .    c'

and so is Hosea vii. 3.

                        Mnpg Mds Npgm yk

           hrmf tmdwmv

            For of the vine of Sodom is their vine,

                And of the fields of Gomorrah (Deut. xxxii. 32),

is an example of

                                    a . b . c

                                    a' . b'




            Incomplete parallelism with compensation.

                        ryfwm jtxcb hvhy

           Mdx hdWm jdfcb

            Yahweh, when-thou-wentest-forth out-of-Seir,

                When-thou-marchedst out-of-the-field of-Edom (Jud. v. 4),


is an example of

                                                a . b . c

                                                     b’ . c’2

            PARALLELISM: A RESTATEMENT         77


and other examples are Deuteronomy xxxii.

13 c, d, xxxiii. 23 ; Job iii. 11; Isaiah xli. 26 a, b,

lx. 3.

                        HFb lxrWy Nkwyv

           bqfy Nyf ddb

            And-so-dwelt Israel securely,

                 By-itself the-fountain of-Jacob (Deut. xxxiii. 28),

is an example of

                                    a . b . c

                                         c' . b'2

and other examples are Amos v. 24 ; Proverbs ii.

1, 7 ; Job iii. 20 ; while Isaiah xliii. 3 c, d ex-

emplifies the scheme

                                    a . b . c

                                       c'2 . b'

In Judges v. 26,

                        hnHlwt dtyl hdy

           Mylmf tvmlhl hnymyv

            Her-hand to-the-tent-peg she-stretched-forth,

                And-her-right-hand to-the-workmen's mallet,

will be found an example of

                                    a .b . c

                                    a'. b'2

and another example of the same scheme in

Psalm xxi. 11.

            Examples of compensation by means of a,

fresh term or terms are--

                        xb ynysm hvhy

           vml ryfwm Hrzv

            Yahweh from-Sinai came,

                And-beamed-forth from-Seir unto-them (Deut. xxxiii. 2),



which is an example of

                                    a . b . c

                                    c' . b' .d


                        dvbk hvhyl vmywy

           vdygy Myyxb vtlhtv

            Let-them-ascribe unto-Yahweh glory,

                 And-his-praise in-the-isles let-them-declare (Isa. xlii. 12),

which is an example of

                                    a . b . c

                                    c' . d . a'

            Examples of distichs in which each line has

but one parallel term and two terms non-parallel

are given below (p. 94), and instances of com-

pensation by a fresh term in lines containing two

terms only have already been given above (p. 73).

            I will conclude the present discussion with

two illustrations of the value of a minuter analysis

of parallelism than has hitherto been considered

necessary, and of some such method as I have

been suggesting of measuring or classifying the

various types of parallelism.

            An effective scheme of parallelism that occa-

sionally occurs consists of two lines each contain-

ing three terms but held together by a single

parallel term in each line, these parallel terms

standing one at the end of the first line, and the

other at the beginning of the second. The scheme


                        a  . b . c

                        c' . d . e

            PARALLELISM: A RESTATEMENT         79


Now, if the articulation of the parallelism is not

observed, couplets of this type are reduced to

ordinary prose, or even to nonsense, or at best

feeble repetition ; but if it is properly articulated,

the couplet is an effective form of "synthetic

parallelism" as Lowth would have called it, of

incomplete parallelism with compensation as I

should term it. Examples of this type occurring

in Genesis xlix. 9 (cf. Nunn. xxiv. 9) and Deutero-

nomy xxxiii. 11 are correctly articulated in the

Revised Version:


            He-stooped-down, he-couched as-a-lion,

                 And-as-a-lioness: who shall-rouse-him-up?

            Smite-through the-loins of-them-that-rise-up-against-him,

                 And-of-them-that-hate-him, that they-rise-not-again.


But if the parallelism is not correctly perceived,

and the words otherwise articulated, how un-

satisfactory does the former of these couplets

become! "He stooped down, he couched as a

lion and as a lioness: who shall rouse him up?"

This suggests a comparison with two different

beasts, whereas the parallelism really expresses

comparison with the lion-class, which it denotes

by the use of two synonymous terms. Yet this

very mistaken articulation is found in Numbers

xxiii. 23, both in the Revised Version and, I

regret to say, in my commentary on Numbers.

If we articulate

            Now shall it be said of Jacob and Israel,

            What hath God wrought!



the natural suggestion is that Jacob and Israel

are different entities, which they are not; Jacob

and Israel are here, as elsewhere in these poems

(Num. xxiii. 7, 10, 21, 23 ; xxiv. 5, 17., 18 f.),

synonymous terms belonging to different members

of the parallelism. The proper articulation of

the passage is,

            Now shall it be said of Jacob,

                And of Israel, What hath God wrought!

and it is interesting to observe that this not very

common type of parallelism occurs twice (see

also xxiv. 9) in the oracles of Balaam.

            The strongly marked pause in the middle, and

the marked independence of the last part, of the

second line are characteristic of all the distichs

just cited. If from these observations we turn

immediately to Hosea iv. 13 c, d, we shall prob-

ably conclude that the difficulties which have

been felt with regard to these lines are unreal,

that the emendations which have been proposed1

wholly unnecessary, and that, in respect of

parallelism and structure, the lines closely re-

semble Numbers xxiii. 23, xxiv. 9, and Deutero-

nomy xxxiii. 11; in this case the correct articula-

tion is,

                        hnblv Nvlx tHt

           hlc bvF-yk hlxv

            Under oak and poplar,

                And terebinth: for good is the shade thereof.


            1 See e.g. W. R. Harper, Commentary on Amos and Hosea (" Inter-

national Critical Commentary"), pp. 260, 261.



My second illustration of the advantages of

some method that enables similarities and dis-

similarities of parallelism to be easily detected

and presented is of a different character, and

shows the bearing of these studies on textual


            Psalm cxiv. consists of eight couplets, each of

which, in the present text at all events, shows

one form or another of incomplete parallelism,

for the most part with compensation. The char-

acteristic incompleteness of the parallelism rings

through even a translation :


            1 When Israel went forth out of Egypt,

                        The house of Jacob from a barbaric people,

            2 Judah became his sanctuary,

                        Israel his dominion.

            3 The sea saw it and fled,

                        Jordan turned backward,

            4 The mountains skipped like rams,

                        The hills like young sheep.

            5 What aileth thee, 0 thou sea, that thou fleest,

                        Thou Jordan, that thou turnest back?

            6 Ye mountains that ye skip like rams,

                        Ye hills like young sheep?

            7 At the presence of the Lord tremble, 0 earth,

                        At the presence of the God of Jacob,

            8 Which turned the rock into a pool of water,

                        The flint into a fountain of water.


The scheme in the Hebrew is as follows :

   1 a . b  .  c                2 a . b . c

           b'2 . c'2                    b' . c'




            3 a . b . c                                 6 a  .  b  .  c

               a’       c’2                             a’           c’2

            4 a  .  b  .  c                             7 a .  b .  c  .  d

              a’           c’2                         a .   b’2

            5  a  .  b  .  c                            8  a  .  b  .  c2

                       b’ .  c’2                                b’ .  c’2


            There seems to me strong ground for holding

that this consistent use of incomplete parallelism

was intentional, or, at any rate, if not intentional,

it is at least an unconscious expression of the

writer's general preference—in a word, it is a

stylistic characteristic ; as such 'it ought not

without good reason to be obliterated. For this

reason Dr. Briggs's reconstruction of this Psalm in

the "International Critical Commentary" is open

to grave objection. The emendations proposed

by Dr. Briggs and the effect of them on the paral-

lelism is as follows: (1) he strikes out as glosses

verses 2 and 8, though both verses show the char-

acteristic incomplete parallelism; (2) in verse 7

he deletes ylvH, tremble; then Nvdx becomes con-

struct before Crx, and the expression "Lord of

the earth" becomes parallel to "God of Jacob,"

and the verse as a whole an example of complete


                                    a . b  . c

                                    a . b’ . c’


(3) in verses 4 b and 6 b he inserts UlHA (of which

ylvH in verse 7 is supposed to be a misplaced cor-



ruption), thus again turning incomplete into

regular complete parallelism,


                        a  .  b  .  c

                        a'  . b'  . c'


            Thus merely by a study of the parallelism this

reconstruction is rendered improbable quite apart

from the question whether metre requires any

such changes, or whether Dr. Briggs's is not a

much more prosaic poem than that of the Hebrew


            In the LXX Psalm cxiv. is united with Psalm

cxv. This union has been very generally regarded

as not representing the original text: in addition

to the reasons commonly given for holding that

the division between the two Psalms in the

Hebrew text is correct, we may now add the differ-

ence in the type of parallelism. In cxv. 5-7 we

find three successive examples of complete paral-

lelism, and although elsewhere in the Psalm there

are examples of incomplete parallelism, these are

mostly incomplete parallelisms of a different kind

from those which occur in Psalm cxiv.












                CHAPTER III





























           CHAPTER III





THE Book of Lamentations has played a con-

spicuous part in the constantly renewed discus-

sions of the subject of Hebrew rhythm. Apart

from any analysis of its cause, and without

any exceptional degree of attention, the reader

of the Hebrew text, or even indeed of the English

version, of the Lamentations, perceives some-

thing in the rhythm or cast of the sentences that

is common to practically the whole of the first

four chapters of the book. This same something

that brings these four poems into a common class,

sharply marks there off from the fifth chapter or

poem, and at the same time, too, from the greater

quantity of the poetry of the Old Testament,

though careful examination has discovered not

a little in various books of the Old Testament

that resembles the first four chapters of Lamenta-

tions in the peculiarity in question.

            But though this striking peculiarity is common





to the four poems constituting the first four

chapters of Lamentations, there are other features

that distinguish them one from another—the

differing alphabetic sequences that are followed

by the initial letters of successive divisions of the

poems (P preceding f in ii., iii., and iv., following

it in i.), the differing lengths of the divisions,

the differing degrees of passion, spontaneity and

vividness with which the subject, common to

them all, is handled. These differences have

attracted and received attention; but, so far as

I am aware, the differences in the use of parallel-

ism as between the four poems have not yet

been analysed: and, yet, such differences exist.

Owing to uncertainties of text and interpreta-

tion, it does not seem to me easy or even practic-

able to give exact statistics of these differences;

yet, by the help of a more accurate measurement

of parallelism, such as I have suggested in the

previous chapter, it will, I hope, be possible to

make manifest the existence and general char-

acter of the differences ; and, in any case, by an

examination of these chapters, I hope to carry

further my line of approach to rhythmical ques-

tions through parallelism.

            Though I cannot undertake any compre-

hensive survey of the history of the study of

rhythm in Lamentations, it will be worth while

to refer to two discussions of the subject—that

of Lowth, who was the first to point out and to



attempt to analyse the rhythmical peculiarity

of Lamentations i.-iv., and that of Budde, who,

by a series of contributions to this subject, begin-.

ping with his fundamental article in the Zeit-

schrift far die alttestamentliche Wissenschaft for

1882, has profoundly influenced subsequent in-

vestigation and terminology.

            Lowth devoted his 22nd and 23rd lectures to

the Hebrew elegy, and he returned to some of

the points then discussed in the preliminary dis-

sertation to his Isaiah (vol. i. pp. xxxiv-xliii,

ed. 3). The genius and origin of the Hebrew

elegy, of the kinah or nehi as the Hebrews called

it themselves, he traces to their manner of cele-

brating the funeral rites ; and in particular to

the employment of professional mourners who

sang dirges. The natural language of grief, he

remarks, "consists of a plaintive, intermitted,

concise form of expression": and as in other

arts, so in that of the Hebrew elegy, "perfection

consisted in the exact imitation of nature. The

funereal dirges were, therefore, composed in

general upon the model of those complaints

which flow naturally and spontaneously from

the afflicted heart: the sentences were abrupt,

mournful, pathetic, simple and unembellished.

. . . They consisted of verse and were chanted

to music."1

            Lowth then points out the peculiarity of the

Lectures . . . (ed. Lond. 1787), ii. 123, 127.



first four poems in Lamentations, and remarks:

"We are not to suppose this peculiar form of

versification utterly without design or importance:

on the contrary, I am persuaded, that the prophet

adopted this kind of theme as being more diffuse,

more copious, more tender, in all respects better

adapted to melancholy subjects. I must add,

that in all probability the funeral dirges, which

were sung by mourners, were commonly corm.-

posed in this kind of verse: for whenever, in the

prophets, any funereal lamentations occur or any

passages formed upon that plan, the versification

is, if I am not mistaken, of this protracted kind.

. . . However, the same kind of metre is some-

times, though rarely, employed upon other occa-

sions. . . . There are, moreover, some poems

manifestly of the elegiac kind, which are com-

posed in the usual metre, and not in unconnected

stanzas, according to the form of a funeral dirge."1

            The peculiarities of this elegiac versification

are best summarised in the Isaiah, as follows :

"The closing pause of each line is generally very

full and strong: and in each line commonly,

towards the end, at least beyond the middle of

it, there is a small rest, or interval, depending on

the sense and grammatical construction, which

I would call a half-pause. . . . The conjunction

v . . . seems to be frequently and studiously

omitted at the half-pause : the remaining clause


            1 Lectures, ii. pp. 136, 137.



being added, to use a grammatical term, by ap-

position to some word preceding; or coming in

as an adjunct, or circumstance depending on the

former part, and completing the sentence."1

The parallelism accompanying the versification

of this kind is, according to Lowth, for the most

part of the constructive order,2 which is, as we

have previously seen, Lowth's way of saying that

strict parallelism is at best incomplete, and is

more often entirely absent.

            There is in the passages just cited or summar-

ised a surprising amount of correct and acute

observation or fruitful suggestion. Some sub-

sequent scholars neglected this important part of

Lowth's inquiries, and, in consequence, Ewald,

for example, never clearly saw, as Lowth had

seen, the sharp distinction between Lamentations

i.-iv. and v.

            For our present purpose it will suffice to refer

much more briefly to Budde's important discus-

sions. In the main his advance on Lowth con-

sisted in the detailed working out of two important

points : (1) the nature of the unequal division of

the rhythmical periods ; and (2) the extent to

which the rhythm characteristic of Lamentations

i.-iv. occurs elsewhere in the Old Testament. As

to the division of the rhythmical periods, Budde's

position may be stated thus :—(1) the kinah

rhythm rests on the division of the rhythmical


            1 Isaiah, ed. 3, p. xxxix.              2 Ibid. p. xxxv.



period into two unequal parts of which the longer

part precedes the shorter part; (2) the normal

length of the longer part is three words, of the

shorter two words; (3) but by legitimate varia-

tions a longer part consisting of four words may

be followed by a shorter consisting of (a) three,

or (b) two, words ; (4) the period is never equally

divided;1 if, as sometimes happens, each part

consists of two words, the two words of the first

part are heavier and weightier than the two

words of the second part; (5) between the two

parts of the verse, there is no strict and constant

rhythmical relation beyond the fundamental fact

of inequality of length.

            To some of these metrical questions I shall

return: meantime I proceed to examine the

parallelism of the poems, and I will begin with

the isolated fifth chapter which happens to be

an excellent storehouse of examples of the types

of parallelism occurring in poetry that is free

from the well-marked peculiarities of Lamenta-

tions i.-iv. By comparison with the more ordinary

parallelism of Lamentations v., any peculiarities

in the parallelism of Lamentations i.-iv. may be

the better discerned.

            The majority of the twenty-two verses of

Lamentations v. may be treated as containing

six terms equally divided among the two stichoi

that compose each verse, i.e. each stichos normally


            1 Zeitschr. fur die alttest. Wissenschaft, 1882, pp. 4 f.



contains three terms. Seventeen of these dis-

tichs show strict parallelism between at least one

term in each stichos; of the remaining five dis-

tichs, one (v. 5) is too uncertain to classify, and

two (vv. 8, 16) are best regarded as lacking strict

parallelism. In the two verses or distichs that

still remain (vv. 9 and 10) the stichoi are certainly

not parallel to one another: but these two verses

in their entirety seem to be (incompletely) parallel

to one another: for disregarding the first half of

v. 10, which may be corrupt, we may represent

the parallelism between the two verses thus :


                        a . b . c . d . e . f

                        .    .   .  . d' . e' . f'

If this parallelism of the last parts of these verses

was intentional, it is likely enough that such

naturally parallel terms as vnwpn, our soul (R.V.

lives), vnrvf, our skin, which occur in the first parts

of the verses, were originally more really parallel

than they now are.

            Of the twenty-two distichs, then, contained in

Lamentations v., seventeen at least show parallel-

ism between the stichoi. In five, or, on one

interpretation of v. 12, in six, of these the parallel-

ism is complete:1 in the remaining twelve (or

eleven) incomplete. The several examples may

be classified thus:--


            1 For the meaning of the terms complete and incomplete parallelism

see above, pp. 59, 74.





Form.              Number of Occurrences.                  Verses.

a . b . c                        3                                              4, 13, (17)

a'. b'. c'

a . b . c                        (1)                                           12 (on one inter-

b'. a'. c'                                                                        pretation)

a . b2                           1                                              15

a'2 . b'

a . b                             1                                              22

b' . c'




(1) With compensation.

a . b . c                        4                                              1, 11, 12 (on one in-

a'2 . b'                                                                               terpretation), 20

or similar types

a . b . c                        2                                              6, 7

a' . d . e


(2) Without compensation.

a . b . c                        4                                              2, 3, 14, 18

a'. b'

or similar types

a . b . c . d                   1                                              19

a'          c'2

a . b . c . d                   1                                              21

a'2           e



            The occurrence in this poem of incomplete

parallelisms without compensation raises ques-

tions that must be considered later.

            In turning now to consider Lamentations i.-iv.

we are faced with a difficulty of terminology.

Lamentations iii., as is well known, consists of

sixty-six Massoretic verses distinguished from one

another by the occurrence, at the beginning of

each, of the letter of the alphabet appropriate

to the alphabetic scheme, so that each of the first

three verses begins with x each of the next three

with b, and so forth. Chapters i. and ii., though

they number each but twenty-two Massoretic

verses, contained1 each of them sixty-six sec-

tions of the same length as the Massoretic verse in

iii., and these sections are still easily distinguish-

able, though the letters of the alphabetic scheme

occur at the beginning of every fourth section

only. Chapter iv. consists of forty-four similar

sections. What is the proper term to apply to

these sections : are they lines or couplets, stichoi

or distichs? Are they, as compared with the

stichoi of chapter v., "protracted lines," as

Lowth described them, or, as compared with the

distichs of chapter v., truncated couplets or

distichs, as Budde considers them? These ques-


            1 In the present text, owing to what is generally recognised as

textual expansion (in i. 7, ii. 19), the number of sections is sixty-seven

both in chaps. i. and ii. The R.V. for the most part distinguishes the

sections correctly, but occasionally so divides the verses (e.g. i. 1, ii. 2,

and even iv. 22) as to give them the appearance of consisting of four




tions can best be considered later : I will, for the

time being, use the neutral term section, meaning

by that a Massoretic verse in chapter iii. and the

equivalent sections of the remaining chapters, i.e.

the third of a Massoretic verse in i. and ii., and

the half of such a verse in iv. Similarly, for the

two parts of these sections, the longer first and

the shorter second part, I will use the term sub-


            As the normal number of terms in a verse of

chapter v. is six, so the normal number of terms

in each section of chapters i. and iv. is five. It

follows from this at once that in chapters i.-iv.

the common form of complete parallelism

            a .  b . c

            a' . b' . c'

will not readily1 occur in a normal section, and,

as a matter of fact, it does not, I think, occur at

all in any section, whether normal or abnormal.

This, however, is not equivalent to saying that

complete parallelism between the subsections is

either impossible or actually non-existent in

these poems ; on the other hand complete paral-

lelism actually occurs, though relatively with

much less frequency than in chapter v. An

example is ii. 11:


            1 The force of this qualifying adverb will become clear later. As a

matter of fact, though a , b , c, // a’ . b’. c’ does not occur, a corresponding type

of incomplete parallelism with compensation does occur: see iv. 11.



     yfm vrmrmH |   ynyf tvfmdb vlk

Consumed with tears are mine eyes, |  in a ferment are my


            The scheme is a2 . b | a' . b' ; and it is prefer-

able to regard iii. 4,

He hath worn out my flesh and my skin, |  he hath broken

                                                                                    my bones,

as an example of a . b2 | a' . b' rather than of the

scheme  a . b . c | a' . b'.

            Other examples of complete parallelism in

chapters i.-iv. occurring in sections that are not

perhaps strictly normal are

             vnl vbrx rbdmb |  vnqld Myrhh-lf

Upon the mountains they chased us, |  in the wilderness they

                                                                        lay in wait for us.


            hnfl ynvrh |   Myrvrmb ynfybwh

He hath filled me with bitterness,  | he hath sated me with


            These will be found in iv. 19 and iii. 15; they

are both examples of a . b | a' . b', or, if we

prefer to regard the pronominal suffixes as in-

dependent terms, of a . b . c | a' . b' . c; another

example occurs in iv. 13, and there are perhaps

a few others: but in the 242 sections of chapters

i.-iv. there are but few, if any, more examples

of complete parallelism than in the twenty-two

distichs of chapter v.; or, in other words, com-

plete parallelism is, relatively, about eleven times

as frequent in chapter v. as in chapters i.-iv.



            If, however, the section of chapters i.-iv. be

a "protracted line," we might expect to find

complete parallelism occurring as between the

sections rather than as between the subsections.

As a matter of fact, incomplete parallelism be-

tween the sections is not uncommon in chapters

i.-iv.; it is less common, indeed, than parallelism

between the stichoi in chapter v.; it is, on the

other hand, much commoner than parallelism

between whole verses, of which we noted but one

example, in chapter v. And yet complete paral-

lelism between sections is exceedingly rare, and

in fact, I think, does not once occur. Probably

the nearest approach to complete parallelism

between sections is where four of the five terms

correspond, as in ii. 2 a, b, where the scheme is

                        a . b . c . d . e

                        a' .     c' . d' . e'2

            bqfy tvxn-lk-tx lmH-xlv yndx flb

           hdvhy-tb yrcbm vtrbfb srh

The-Lord hath-swallowed-up unpityingly all the-homesteads


He-hath-thrown-down in-his-wrath the-strongholds of-the-

                                                                                    daughter of-Judah.


            A much greater relative amount of those forms

of what Lowth called synthetic or constructive

parallelism, in which there is a complete absence

of strict parallelism, is another feature of Lament-

ations i.-iv. which sharply distinguishes these

poems (with one exception) from Lamentations v.



Other differences exist as between one or more

of these poems and chapter v.; and these will

appear when we turn, as we must now, to a closer

examination of the parallelism in chapters i.-iv.,

and of the differences in this respect to be dis-

cerned as between these chapters considered


            Budde quotes with approval a remark of De

Wette's that in Lamentations " merely rhythmi-

cal parallelism," another term for Lowth's con-

structive or synthetic parallelism, is most promi-

nent, and that parallelism of thought, when it

occurs, occurs mostly as between the subsections,

i.e. between the clauses or sentences which con-

sist alternately of (as a rule) three and two terms,

not between the sections, which consist, as a rule,

of five terms; put otherwise, this amounts to

the assertion that parallelism in these poems is

chiefly of the general type

                        a . b . c

                        a'. b'

not of the type

                        a . b . c . d . e

                        a'. b'. c'. d'. e'

Budde's only criticism of this is that De Wette

considerably underrates the extent of this

parallelism between the subsections, which we

may briefly term subsectional parallelism. But

neither De Wette nor Budde carried the analysis



of this feature sufficiently far; had they done so

they would have seen that a general statement

such as they make cannot be rightly made with

reference to all the poems indiscriminately. I

hope to show that the statement that " merely

rhythmical parallelism " is most prominent is

substantially true of chapters i. and iii. and

very misleading in reference to chapter ii., and

in a less degree in reference to chapter iv.;

and also that the statement that parallelism,

when it occurs, occurs mostly between the sub-

sections is the very opposite of the truth with

regard to chapter ii., though substantially correct

with regard to chapter iv.

            I will examine chapter iii. first. In a certain

sense the whole of the first eighteen verses or

sections might be said to consist of eighteen

parallel statements of the fact that Yahweh is

chastening the speaker; the first person singular

pronoun appears in each separate verse, and gives

a certain degree of parallelism to them all; and

similarly throughout the poem large groups of

sections express, mainly by a succession of figura-

tive statements, the same thought: but beyond

this general repetition of thought there is seldom

any real parallelism of individual terms or even

of groups of terms. Moreover, there is a feature

of this poem that suggests that some even of th.e

apparent examples of parallel sections are due

more to accident than design; I refer to the fact




that the clearest apparent examples of sectional

parallelism occur between the last section begin-

ning with one letter of the alphabet and the first

section beginning with the next letter;1 thus,

there are throughout the poem no sections more

parallel to one another than, and few as much so

as, the following (vv. 12, 13 ; 48, 49 ; 60, 61),


He hath bent his bow and set me as a target for his arrow;

He hath caused to enter into my kidneys the shafts of his



In streams of water my eye runs down for the destruction

                                                                                    of my people;

My eye hath poured down unceasingly, because there are no



Thou hast seen all the vengeance they took, all their devices

                                                                                    against me;

Thou hast heard all their reproaches (of me), 0 Yahweh, all

                                                            their devices against me.


The first of these couplets consists of the last line

beginning with d and the first with h, the second

of the last line with p and the first with f, the

third of the last with r and the first with w.

            There are not more than about a dozen2

couplets of contiguous sections that are as


            1 The significance of this does not seem to me to be affected by the

fact that in Ps. cxi., cxii. the alphabetic scheme distinguishes each

stichos, not each distich, by successive letters of the alphabet, and

therefore regularly and necessarily gives to parallel stichoi different

initial letters.

            2 The sections that may most reasonably be regarded as more parallel

(though whether always by the intention of the writer is doubtful) to

one another than is almost any section of the poem to any other are :

12, 13; 19 (pointing -10, 20 ; 28, 29, 30 (?) ; 34, 35, 36 (?) ; 40, 41 ;

48, 49 ; 60, 61 ; 64, 65. The italicised numbers are cited above.



parallel to one another as the foregoing, or

indeed that are strictly parallel to one another

at all.

            In about one-third of the entire number of

sections parallelism more or less clear and con-

spicuous between subsections ' occurs ; examples

are vv. 10 (a . b . c2 | a' . b') and 14 (a . b . c

b' . d).--

                        Myrtsmb hyrx | yl xvh brx bd

As a bear lying in wait is he unto me, | a lion in secret places.

           Mvyh-lk Mtnygn | Mymf-lkl qHw ytyyH

I am become a derision to all peoples, | their song all the day.


            Clearly, then, since subsectional parallelism

occurs in considerably less than half, and prob-

ably in not more than a third, of the sixty-six

sections of the poem, and sectional parallelism,

which might have occurred thirty-three times,

actually occurs scarcely a dozen times at most,

"merely rhythmical parallelism" is more con-

spicuous here than real parallelism of thought

and terms; whether subsectional is much or any

more relatively frequent than sectional parallel-

ism depends on the view taken as to the reality

of parallelism in the couplets specified on p. 101

and as to the character of the more doubtful

examples of subsectional parallelism given below.1


            1 The clearest examples of subsectional parallelism occur in the

following fifteen verses : 4, 9, 10, 14, 15, 17, 18, 22, 23, 25, 33, 47, 58,

60, 61. The text of some even of these (e.g. 22, 23, 33) is open to

question: but probably parallelism existed in the original text. More

doubtful examples maybe found in vv. 5, 7, 11, 16, 19, 30, 39, 43, 53, 56, 65.



Chapter ii. differs greatly from chapter iii.

The repetition in chapter iii. of the initial letter

before each of the three sections belonging to it

corresponds to a real independence, as a general

rule,l of the sections in that poem. On the other

hand, the three sections which belong to each

letter of the alphabet in chapter ii., but of which

the first section only is distinguished by beginning

with that letter, are closely connected with one

another ; and this connexion is formally marked

by the frequency with which the entire sections

within the several alphabetic divisions are parallel

to one another. The exact number of these

sectional parallelisms depends on interpretation,

and in some cases on textual questions: but I

believe it may be safely asserted that in a large

majority at all events of the twenty-two alpha-

betic divisions two at least of the three sections

are parallel to one another, and in several all

three sections are so. I should myself put the

number of parallelisms' between two, if not all

three, sections as high as eighteen, if not higher.2

Over against this frequency of sectional paral-

lelism we have to set the relative infrequency

of subsectional parallelism : this latter kind of

parallelism, which might have occurred sixty-six


            1 Vv. 34-36 form an exception.

            2 Absence of parallelism or a near approach to it will be found in

vv. 4, 17, 18, 22, but even this may be partly due to textual corruption.

In most of the remaining verses parallelism is obvious, in all it was

probably intended.



times, actually occurs only a dozen1 times, more

or less, according to the view taken of two or

three doubtful cases.

            Thus it is not true of chapter ii. that "merely

rhythmical parallelism" is more frequent than

real parallelism of thought and term, nor is it

true that parallelism occurs mainly between the

subsections ; quite the reverse: we must, to be

accurate, put the case thus: In chapter ii. real

(though incomplete) parallelism is very frequent;

the fundamental parallelism is between the sec-

tions; but this is occasionally reinforced by an

additional and secondary parallelism between the

subsections, much in the same way that the

fundamental rhymes at the close of the (alternate)

lines of a quatrain are in some English poems

occasionally reinforced by an additional rhyme

in the middle of one or more lines, as often in

Coleridge's Ancient Mariner, e.g.


            The sun came up upon the left,

                        Out of the sea came he!

            And he shone bright, and on the right

                        Went down into the sea.


            The fact is, parallelism in Lamentations ii. is

singularly intricate and skilfully varied. It is

rarely complete either as between sections or sub-

sections, but it is generally clear enough and

sufficient to constitute a real formal connexion


            1 See vv. 4 a (?), 5 b, 6 a (?), 7 a, 9 a (read UrB;wu for rbwv dbx), 10 b, 11 a,

(not 13 a: AV.), 15 c (present text), 17 a, c, 18 c, 20 b, 21 e.



between the three sections of the several alpha-

betic divisions, or at least between two of them,

the remaining section being sometimes not parallel,

as is frequently one stiehos of a tristich in other

poems. Since the nature of the parallelism in

chapter ii. and, consequently, an important formal

difference between chapters ii. and iv. have

hitherto not been clearly observed, I give a few

verses of this poem with a translation and notes

on the parallelism:--


            Nvyc tb tx | yndx vpxb byfyh1 hkyx 1

            lxrWy trxpt | Crx Mymwm jylwh

     vpx Mvyb | vylgr Mvdh rkz xlv

1 How hath the Lord beclouded1 in his anger | the daughter

                                                                                                of Sion!

    He hath cast down from heaven to earth | the ornament of


    And he hath not remembered his footstool | in the day of

                                                                                                his anger.


            Here all three sections are parallel: observe

the daughter of Sion (d 2) || the ornament of Israel

(d' 2) || his footstool (d" 2), and beclouded (a) ||

cast down from heaven to earth (a' 3) || hath not

remembered (a"). Moreover, the unity of the entire

alphabetic division is emphasised by the addi-

tional parallelism in his anger (b) || in the day of

his anger (b' 2) in the first and last sections; a

similar effect is obtained in v. 12 which opens with

Mtmxl, to their mothers, and closes with Mtmx, their

mothers. Variety is obtained not only by varying


            1 Flatly . . . beclouded: read byfyh for byfy, beclouds.



the number of terms by means of which corre-

sponding ideas are expressed, but also very effect-

ively by bringing the object of the verb much

nearer to the beginning in the third section than

in the two that precede : a somewhat similar

effect is obtained in v. 8 (cp. also i. 1).

            There, is no subsectional parallelism in any of

these three sections.

            bqfy tvxn lk tx | lmH xlv yndx flb 2

            hdvhy tb yrcbm | vtrbfb srh

            hyrwv hklmm | llH Crxl fygh


2 The Lord 'hath destroyed unsparingly | all the homesteads

                                                                                                of Jacob;

    He hath pulled down in his wrath | the strongholds of


    He hath brought to the ground, hath profaned | the realm

                                                                                    and its princes.


            Here, again, all three sections are parallel, but

in none is there parallelism between the sub-

sections. This time all the object-clauses stand

at the end of their respective sections and, as in

v. 1, the parallel verbs or verbal clauses Crxl fygh

llH (he hath brought to the ground, hath profaned),

srh (he hath pulled down), flb (hath destroyed) at

the beginning. The additional parallelism of

terms is not as in v. 1 between the first and

third, but between the first and second sections

(unsparingly || in his wrath), unless, indeed, with

Lohr, we emend by transposing the clauses He

hath brought to the ground and in his wrath; then,



as before, the fuller parallelism will be between

the first and third sections.

            Nvyc tb ynqz | vmdy Crxl vbwy 10

            Myqw vrgH | Mwxr lf rpf vlfh

            Mlwvry tlvtb | Nwxr Crxl vdyrvh

10 They sat on the ground dumb—| the elders of Sion;

     Lifted up dust on their head, | were girded with sack-


     They lowered to the ground their head—| the virgins of


            Here in the second section we find subsectional

parallelism; each clause in it mentions one sign

of mourning and grief; parallel to each of these

clauses and to one another are the first clauses of

the first and third sections, but these sections

contain no subsectional parallelism : on the other

hand, the second parts of the first and third

sections are very strictly parallel to one another

(the elders of Sion || the virgins of Jerusalem). But

there is still further and in part rather subtle

verbal parallelism between the sections: note Crxl

(on (to) the ground) in the first and third sections ;

Mwxr and Nwxr (their head) in the second and third

respectively; and the antithesis vlfh (lifted up)

and vdyrvh (lowered) which is emphasised by the

parallelism in a way which it is impossible to

represent adequately in translation: what they

lift up is dust, what they cast down is their heads!

Very clearly, then, sectional parallelism is again

primary; but here it is reinforced by subsectional

parallelism in one of the three sections.



            A correct appreciation of the main and. second-

ary parallelism in this poem may set some ques-

tions of textual interpretation in a new light.

Verse 3 reads,

            lxrWy Nrq lk | Jx yrHb fdg

            byvx ynpm | vnymy rvHx bywh

     bybs hlkx | hbhl wxk bqfyb rfbyv

He hewed off in fierce anger | all the horn of Israel;

He turned backward his right hand | from the face of the foe;

And he kindled in Jacob a flaming fire | which devoured

                                                                                    round about.


Whose is the right hand here referred to, Israel's

or Yahweh's ? It is commonly taken to be

Yahweh's, and there is certainly much to be said

for this view. But the parallelism of the sections,

which certainly exists in any case, would become

still clearer and more complete if the right hand

be Israel's. Then, for the use of the pronoun

only in the middle section corresponding to the

two parallel proper names for the nation in the

first and third sections, there are two exact

parallels in this poem : see vv. 5 and 10.

            In both 4 a and 15 c it is generally admitted

that a word or more has intruded. But which

word or words should we omit? If subsectional

parallelism was primary, and as frequent as :it is

in Lamentations iv. and Isaiah xiv., parallelism

would furnish a strong argument for those 'who

retain rck, as a foe (parallel to as an enemy), in

v. 4, and both the clauses perfection of beauty



and joy of the whole earth in v. 15. But, since

subsectional parallelism is merely secondary and

not very frequent in this poem, such an argument

has little if any weight: and it may certainly be

doubted whether it is nearly strong enough to

justify those who omit vrmxyw, with the character-

istic to in v. 15, in order to retain both the

parallel clauses at the end of the verse without

at the same time keeping a section so long as the

existing text presents.

            Verse 8 is also interesting. Had subsectional

parallelism been primary, the author would

naturally have written--

            Rampart and wall lament | together they languish;

but to gain a closer parallelism with the two pre-

ceding sections, each of which begins with a verb

of which Yahweh is the subject, he avoided what

would have been a more perfect subsectional

parallelism and wrote instead--


He caused to lament rampart and wall; | together they


            By many who refrain from postulating unity

of authorship for the Book of Lamentations,

chapters ii. and iv. at least are attributed to the

same writer. Be this as it may, there is an

appreciable difference, though it has hitherto

been overlooked, in the use of parallelism in the

two poems, just as there is a difference in the

length of the alphabetic divisions. In chapter ii.



sectional parallelism is fundamental and frequent,

subsectional parallelism secondary and relatively

rare : in chapter iv. subsectional parallelism is

relatively more frequent, perhaps even consider-

ably more frequent than sectional parallelism,

though neither type is quite so unmistakably

primary or quite so persistent as the sectional

parallelism in chapter ii. Subsectional parallel-

ism occurs in nearly, if not quite, or even more

than, a half1 of the sections in chapter iv. as corn-

pared with a bare fifth in chapter ii.; on the

other hand, less than half, perhaps scarcely a

third, of the sections are parallel to one another,2


            1 The sections in Lamentations iv. number 44, of which two (v. 15)

are through corruption very uncertain. Subsectional parallelism is

clearest in these 17 sections : 1 a (see below), 2 a, b, 3 a, b, 7 a, b, 8 a, b,

11 a, b, 12 a, 13 a, 16 b, 18 b, 19 b, 21 a. To these should be added the two

similarly constructed sections, 6 a, 9 a, perhaps also 5 a, b (antithetical

parallels), G b, 14 a, 15 a, 21 b, 22 a, b. Subsectional parallelism is at all

events sufficiently frequent to raise the question whether the text of

v. 1 is correct ; subsectional parallelism would indeed be perfect even

in the present text if we ventured to divide the section equally (cp.

R.V.) : but rhythm, as we shall see later, forbids this, and if the text

is sound Dr. Smith (Jerusalem, ii. 270) rightly arranges as follows :

            How bedimmed is the gold, how changed

                        The best of the gold.

I suspect, however, that either (1) :.w' is a gloss (Aramaic ?) on evr,

or (2) that men should be omitted, leaving en: parallel to em as in

Job xxxi. 24. Then we have either

            How bedimmed is the gold,

                        Even the best fine gold,


            How bedimmed is the gold,

                        Changed the fine gold.

            2 The most conspicuous sectional parallelisms will be found in vv.

4, 5, 8, 17, 22 : see also vv. 1, 7, 19, but in these latter verses, as also in

the antithetical sections of v. 3, the sectional parallelism is much less

conspicuous than the synonymous subsectional parallelism in one or,

in most of the verses, in both sections.



and there is little or nothing of that subtle linking

of the sections which occurs in chapter ii.

            In Lamentations i., in spite of the sustained

and well varied parallelism of the first three

sections, strict parallelism is decidedly less frequent

than in either chapter ii. or chapter iv., or even

than in chapter iii. Subsectional parallelism is

perhaps rather more frequent1 than in chapter

ii., where it is infrequent and secondary: but

sectional parallelism is very decidedly less fre-

quent 2 than in chapter ii.: the result is that it

is difficult to select either type of parallelism as

primary ; and the more important fact is that

the form of the greater part of this poem is

independent of strict parallelism.

            It is not surprising that the Book of Lamenta-

tions has driven even unwilling scholars to the

consideration or reconsideration of the question

of metre or rhythm in Hebrew poetry. Budde,

who, like many others, had in 1874, after an

examination of existing theories in regard to

Hebrew metre, rejected them all and expressed

the most thoroughgoing scepticism with regard

to any new theories that might arise, found him-

self eight years later, after a study of Lamenta-

tions, venturing, to quote his own phrase, "on


            1 See vv. 1 (three antithetical parallels), 2 a, c, 3 a, b, 41), c, 5 a, 7 c, d,

13 c, 16 a, b, 18 b, 20 a, c; possibly also vv. 8 a (omit I Nk-lf?), b (omit yk ?),

c, 9 c, 13 a, 22 a.

            2 See vv. 1, 10 a, b, 11 a, b, 12 b, c, 15, 20 a, b: perhaps also 2 b, c,

4a,b, 5a, c, 8.




the dangerous slippery ice"; and it has generally

been admitted that he skated with considerable

skill over the corner of the ice to which he confined


            The challenge lies here: there is a common

and well-marked peculiarity in the 242 sections

that make up the first four chapters of Lamenta-

tions ; it is a rhythmical peculiarity, and yet a

rhythmical peculiarity that cannot be explained

by the parallelism. In putting it thus, I recog-

nise, as I think we well may, that parallelism

might create rhythm, and may even, as a matter

of fact, in the remote past have created the

dominant Semitic and Hebrew type of rhythm

in particular : a habit of expressing a thought

in a given number of terms, and then repeating

it by corresponding terms, would necessarily pro-

duce a certain rhythmical effect: thus, for

example, the habit of expressing thought in the

mould symbolised by

                                    a  . b  .  c

                                    a' .  b' . c'

would produce a rhythm which may be expressed

by 3 : 3 ; and thought expressed in a mould

symbolised by

                                    a  . b  .  c

                                    a'  . b'

would produce a rhythm that may be expressed

by 3 : 2.

            But as soon as parallelism becomes incomplete,




and still more when it becomes merely synthetic,

i.e., strictly speaking, disappears, and yet the

lines retain the same number of words or terms,

obviously the rhythmical relation between the

lines is no longer, even if it was originally, merely

secondary : thus rhythm is no longer a mere

result of parallelism, but an independent desire

for rhythm is at least a contributory cause, if


                                    a  .  b  .  c

                                    a'  . b' .  c’

such schemes as

                                    a . b . c

                                    a'2   .  c'


                                    a  .  b  . c

                                    a'  . d .  e


                                    a . b . c

                                    d  . e . f

constantly alternate, but schemes such as

                                    a  . b . c

                                    a'2 . b' . c'


                                    a  . b . c

                                    b' . c' . d

rarely or never ; or, again, if with schemes such as

                                    a . b . c . d . e

                                    a'. b'. c' . d' . e'




there alternate schemes such as

                                    a  . b . c . d . e

                                    a' . b'2   . d' . e'

but not such as

                                    a  . b  .  c  . d  . e

                                    a' . b'2 . c' . d' . e'

or with schemes

                                    a  . b . c

                                    a’ . b’

schemes such as

                                    a  . b . c



                                    a  . b . c

                                    a’2 . b’

but not such as

                                    a   .  b  .  c

                                    a'2 . b'

            Now, if my analysis is even approximately

correct, what, stated in general terms, are the facts

of the Book of Lamentations, and the questions,

which, once the facts are analysed and classified,

almost necessarily arise? Lamentations iii. con-

tains sixty-six sections unmistakably marked off

from one another by the alphabetic scheme: there

is no complete parallelism between any two suc-

cessive sections: there is incomplete parallelism

between perhaps fifteen groups of two sections:

there is none at all between the rest. Why are




these sections nevertheless of equal length, or at

least even in the present text so closely approxim-

ated to equality of length? Again, these sections

fall into subsections : in some twenty sections the

two subsections are parallel to one another, though

often only incompletely parallel; why alike in

these twenty sections and in the remaining forty

odd sections in which there is no parallelism

between the subsections does the longer sub-

section precede the shorter: why is the ratio

between the two subsections so constant?

Again, why are the twenty-two alphabetic

divisions of Lamentations ii. each divided into

three equal divisions marked off from one another

by a strongly marked division of sense, each

section again into subsections by a less strong but

still clearly marked pause? Why do the sections

so constantly consist of five terms, the subsections

of three terms and two terms respectively, the

shorter regularly following the longer? Why all

this, though, while many of the sections are

parallel to one another, complete parallelism

between sections scarcely, if ever, occurs, and

though in only about a dozen out of the sixty-six

sections does even incomplete parallelism occur

between the subsections?

            The answer to all these questions and the

similar questions which Lamentations i. (with a

difference) and Lamentations iv. provoke has

been increasingly found. by admitting the play




of a rhythmical principle ; and what is called

the Dinah rhythm has accordingly gained recogni-

tion amongst many who still remain sceptical of

other Hebrew rhythms.

            What, then, is really meant by the Dinah

rhythm? A certain ambiguity seems to lurk in

the 'usage of the term. Does it mean five terms

forming a complete sentence with a well-marked

pause after the third? or a succession of such

sentences? If the first sentence of Genesis--

Mymwh-txv Crxh-tx | Myhlx xrb tywxrb—occurred. in

any of the first four chapters of Lamentations,

every one would accept it as a rhythmically

normal line. Is, then, the first sentence in

Genesis an example of kinah rhythm occurring

sporadically in prose, as hexameters occur spor-

adically in the Authorised Version? Scarcely, for

it is probable that those who define kinah rhythm

as verse unequally divided by a pause, and

normally in the ratio 3 : 2, tacitly mean by

kinah rhythm a succession of such verses. And

certainly it was the frequent repetition of such

verses in Lamentations i.-iv. that first drew atten-

tion to the peculiarity of their style or rhythm.

            Five words with a pause after the third is,

even in Hebrew prose, too frequently occurring

and too easily arising a phenomenon to possess

by itself anything distinctive. An hexameter is

a noteworthy phenomenon wherever it occurs ;

five words with a pause after the third are not ;




on the other hand, a dozen or twenty repetitions

of five words with a pause after the third do con-

stitute something as noteworthy as an hexameter.

            Not the sporadic occurrence, but the regular

recurrence of a particular type of word-combina-

tion is apart from, or in addition to, any parallel-

ism that may accompany it, the peculiarity of

Lamentations i.-iv. And yet, as soon as we

frame the conclusion thus, it is necessary, if all

the facts, especially of chapter i., are to be

recognised, to add that the particular type of

word-combination in question falls into two sub-

types; and as soon as we define the sub-types as

consisting respectively of combinations of five

words with a pause occurring after the third, and

combinations of four words equally divided by

a pause, we may at first appear to destroy the

whole theory of a kinah rhythm which we were

attempting to formulate. The actual fact is not

quite so serious as this, for while the normal

section of five accented words, unequally divided,

may contract to four words equally divided, it

probably does not expand to six words equally


            However, whether the facts seriously weaken

the theory or not, the main question at present

is this : is Ludde correct in denying that the

sections in Lamentations were ever (in the original

text) equally divided ? And is his attempt to

maintain the appearance of inequality by calling




two words "heavy" as against two others that

are to be called "light," any better than the

attempt to cover up the absence of parallelism

between two lines by speaking of them as synthetic


            To this question we shall return. Meantime,

I will only say that the theory of light and heavy

groups of words seems to me to suffer shipwreck

on the very first verse of the book : for it is very

difficult to believe that if Myvgb ytbr at the end of

the second section is light, tvnydmb ytrw at the

beginning of the third is heavy. The truth is

rather that Lamentations i. 1 b, c are both lines

of four words equally divided: and Sievers is

probably not far wrong in finding a full half of

the entire number of lines in Lamentations i. to

be of the same nature.1 In any case, Lamenta-


            1 The sections treated by Sievers as containing four accented words

and as being equally divided by the caesura are 1 b, c, 2 b, 4 c, 5 b, c, 6 a,

c, 7 a (to hyrvrmv), c, 8 b, c, 9 b, 10 a, b, 11 a, 12 c, 13 a, b, c, 14 b, e, 15 a, b,

17 c, 18 b, c, 19 a, b, e, 22 b, c; marked as less certain sections of the same

kind are 2 c, 3 b, c, 4 b, 15 c. Sections of this kind are far less frequent

in the remaining poems ; those treated as such by Sievers are : ii.

12 (a, b) c, 14 a, b, c, (19 d) ; iii. 6, 10, 13, 15, 23, 24, 50 (58, 59, 60); iv.

3 b, 5 a, b, 6 b, 13 a, b,14 (a) b, (15 a, b),18 a (b), 20 (a) b, 21(a) b. References

to uncertain examples are enclosed in brackets. It is interesting and

instructive to compare with this classification the examples given by

Budde (Zeitschr. fur die alttestamentliche Wissensehaft, 1882: cp. his

commentary on Lamentations in the Kurzer Handlcommentar, 1898)

of the verses in which the first part contains only two words—these

being, on his theory, " long " or " heavy." Budde cites i. 1 b, c, 4 e,

9 b, 13 c, 14 b,17 c, 18 c, 19 a, b ; ii. 12 b, c ; iii. 15 ; iv. 5 a, 13 b, 17 b.. The

large number of sections treated by Sievers as evenly divided, but not

treated by Budde as containing two words only in their first parts,

consists of lines in which Budde either allows a full word-value to

prepositions or other particles (e.g. i. 8 c, 10 b, 11 a), or emends the text

(e.g. in i. 5 b he inserts xvh after hvhy).




tions i. is of crucial importance in the study of

the kinah rhythm: any one who has sufficient

ingenuity to discover an unequal division in all

its sections need have little fear of being able to

do the same for the three succeeding chapters or

any other passages where the occurrence of some

unequally divided lines suggests to him the

"kinah" rhythm. If, on the other hand, the

occurrence in the present text of Lamentations i.

of equally divided lines of four terms is too

frequent to admit of doubt that some such lines

occurred in the original text, then we may suspect

that the same variations also occurred or may

have occurred in other kinah poems.

            And as a matter of fact the variation is prob-

ably to be found in one of the earliest kinahs that

survive. In Amos v. 2 the prophet's kinah over

the house of Israel is given: it consists of two dis-

tichs, or long lines as we may here by preference call


            lxrWy tlvtb |  Mvq Jsvy-xl hlpn

     hmyqm Nyx  |  htmdx-lf hwFn

Fallen to rise no more is the daughter of Israel,

            Stretched out upon the ground with none to raise her.


            The parallelism resembles the dominant paral-

lelism in Lamentations ii.: it is between the

long lines, not between the parts of these, the

scheme being

                                    a . b2 |  c2

                                    a'2     |  b'2




The first of these two long lines is quite unambigu-

ously divided into two unequal parts : rhythmic-

ally it is 3 : 2; but the second can only be forced

into the same scheme by giving to the preposition

a full stress. If, however, we find other examples

of periods in kinahs that cannot be anything but

2 : 2, we shall certainly do better so to regard the

second period here and to give htmdx-lf but one















                      CHAPTER IV































                                    CHAPTER IV




            THE study of parallelism must lead, if I have so

far observed and interpreted correctly, to the

conclusion that parallelism is but one law or

form of Hebrew poetry, and that it leaves much

to be explained by some other law or form.

Complete and exact correspondence of all the

terms in two parallel lines necessarily produces

the effect of exact or approximate rhythmical

balance. But such complete parallelism is rela-

tively rare in Hebrew poetry; the parallelism

is more often incomplete; and, moreover, along

with lines completely parallel and lines incom-

pletely parallel there frequently occur, also lines

unconnected by the presence in them of any

parallel terms. And yet, alike in the incompletely

parallel, and in the non-parallel couplets, there

will often be found, consistently maintained, the

same kind of rhythm as in those that are com-

pletely parallel. We are thus driven back behind

parallelism in search of an independent rhythmi-






cal principle in Hebrew poetry which will account

for the presence of balance, or other rhythmical

relation, as between two lines in which the

parallelism is not such as necessarily to involve

this balance or other rhythmical relation.

            Some such rhythmical principle, whether or

not its nature can ever be exactly and fully ex-

plained, seems to govern much of the present text

of the Old Testament, sometimes for long con-

secutive passages, as for example in Lamentations

and many parts of Job and Isaiah xl.-lv., some-

times for a few lines only, and then to be rudely

interrupted by what neither accommodates itself

to any rhythmical principle that can be easily

seized, nor produces any rhythmical impression

that can be readily or gratefully received.

The difficulties in the way of discovering and

giving any clear and full account of this principle

are considerable. In the first place, as was

pointed out in the first chapter, no clear tradition

or account of the rhythmical or other laws of

Hebrew poetry has descended to us from the age

when that poetry was still being written. The

remarks of Josephus are interesting, but in them-

selves anything but illuminating. Then we are

faced with serious textual uncertainties in all the

so-called poetical books and in the prophetical

books, and in the ancient poems, such as the song

of Deborah, and the blessing of Jacob, embodied

in some of the narrative books. Feeling, as in my

ELEMENTS OF HEBREW RHYTHM                    125


opinion we ought to do, that much of the poetical

contents of the Old Testament has suffered serious

textual corruption, we might well view with sus-

picion any metrical theory that found all parts

of the existing text equally metrical ; for though

a textual corruption may accidentally at times

have the same metrical value as the original

reading, this is the kind of accident that cannot

happen regularly. On the other hand, a metrical

theory which finds innumerable passages corrupt,

though they show, metre apart, no sign of corrup-

tion, has this disadvantage: given the right to

make an equal number of emendations purely

in the interests of his theory, another theoriser

might produce an equally attractive theory; and

we should be left with the uncertainty of choice

between two alternatives both of which could not

be right, but both of which might be wrong. A

sound metrical theory, then, must neither entirely

fit, nor too indiscriminately refuse to fit, the

present text of the Old Testament. A third

serious difficulty lies in our imperfect knowledge

of the vowels with which the texts were originally

intended to be read. This last difficulty may,

perhaps, always leave a considerable degree of

detail ambiguous, even if the broader principles

of rhythm become clear.

            In spite of these difficulties, how far is it

possible in the first instance to determine the

exact rhythmical relations between, let us say,



the several examples or types of two sections,

sentences, lines, call them what we will, that are

associated with one another by some degree of

parallelism of terms or at least by some similarity

of structure, by being, if not parallel, yet paral-

lelistic? Parallelism both associates and dis-

sociates; it associates two lines by the corre-

spondence of ideas which it implies; it dissociates

them by the differentiation of the terms by means

of which the corresponding ideas are expressed as

well as by the fact that the one parallel line is

fundamentally a repetition of the other. The

effect of dissociation is a constant occurrence of

breaks or pauses, or rather a constant recurrence

of two different types of breaks or pauses: (1)

the break between the two parallel and corre-

sponding lines; and (2) the greater break at the

end of the second line before the thought is

resumed and carried forward in another combina-

tion of parallel lines. And even when strict

parallelism disappears, the regular recurrence of

these two types of pauses is maintained. Thus

there are in Hebrew parallelistic poetry no long

flowing verse-paragraphs as in Shakespearian or

Miltonic blank verse, but a succession of short

clearly defined periods as in much English rhymed

verse and in most pre-Shakespearian blank verse..

Rhyme in English and parallelism in Hebrew

alike serve to define the rhythmical periods; but

the relation between rhyme and sense is much less



close than between parallelism and sense, and

consequently rhyme in English has nothing like

the same power as parallelism in Hebrew to pro-

duce coincidence between the rhythmical periods

and the sense-divisions; accordingly, though

rhyme very naturally goes with "stopped-line"

verse, as it is called, it is also compatible with

non-stop lines; so that non-stop lines and verse-

paragraphs that disregard the line divisions almost

as freely as Shakespearian or Miltonic blank verse

are by no means unknown in English rhymed

poetry. On the other hand, parallelism is,

broadly speaking, incompatible with anything

but "stopped-Line" poetry. Whether or not

there may be in Hebrew a non-parallelistic poetry

in which rhythmical and sense divisions do not

coincide is not, for the moment, the question;

it is rather this: parallelism, even incomplete

parallelism in its various types, offers a very

large number of couplets in which we can be

perfectly certain of the limits of the constituent

lines; how strict, how 'constant, of what precise

nature is the rhythmical relation between these

lines which are thus so clearly defined? If we

can determine this question satisfactorily, we

may obtain a measure to determine whether the

same rhythmical periods occur elsewhere without

coinciding with sense-divisions.

            I have referred to two types of English verse;

but the closest analogy in English to Hebrew



poetry is probably to be found neither in blank

verse nor in rhymed verse, but in the old Anglo-

Saxon poetry, and its revival (with a difference)

in Chaucer's contemporary, the author of Piers

Ploughman. That poetry has one feature which

is no regular, nor even a particularly common,

feature of Hebrew poetry, viz. alliteration; but

that feature, though a most convenient indication

of the rhythm, is absolutely unessential to it.

Apart from the references to this alliteration, how

admirably does Professor Saintsbury's descrip-

tion of this type of English poetry correspond,

mutatis mutandis, to the rhythmical impressions

left by many pages of Hebrew psalms or prophecy.

"The staple line of this verse consists of two halves

or sections, each containing two ‘long,’ ‘strong,’

‘stressed,’ ‘accented’ syllables, these same syl-

lables being, to the extent of three out of four,

alliterated. At the first casting of the eye on a

page of Anglo-Saxon poetry no common resem-

blances except these seem to emerge. But we see

on some pages an altogether extraordinary differ-

ence in the lengths of the lines, or, in other words,

of the number of ‘short,’ ‘weak,’ ‘unstressed,’

‘unaccented’ syllables which are allowed to

group themselves round the pivots or posts of

the rhythm. Yet attempts have been made, not

without fair success, to divide the sections or half-

lines into groups or types of rhythm, more or less

capable of being represented by the ordinary



marks of metrical scansion. . . . A sort of mono-

tone or hum . . . will indeed disengage itself

for the attentive reader . . . but nothing more

. . . the sharp and uncompromising section, the

accents, the alliteration--these are all that the

poet has to trust to in the way of rules sine queis

non. But before long the said careful reader

becomes aware that there is a ‘lucky license,’

which is as a rule, and much more also ; and that

this license . . . concerns the allowance of un-

accented and unalliterated syllables. The range

of it is so great that at a single page-opening,

taken at random, you might find the lines varying

from nine to fifteen syllables, and, seeking a little

further, come to a variation between eight and

twenty-one."1 In Piers Ploughman the verse still

consists of "a pair of sharply-separated halves

which never on any consideration run syllabically

into each other, and are much more often than

not divided by an actual stop, if only a brief one,

of sense";2 but there is a greater approximation,

though only an approximation, to regularity in

the length of the lines: and the first hemistich

(measured of course syllabically, not by its stressed

syllables, which are always equal in number) is

generally longer than the second.3

            As between Anglo-Saxon poetry or Piers


            1 G. Saintsbury, A History of English Prosody, i. 13 f.

            2 Ibid. i. 182.

            3 Cp. ibid. i. 184. Professor Saintsbury gives the well-known open-

ing lines of the poem as an illustration. A briefer specimen from else-



Ploughman and Hebrew parallelistic poetry

these resemblances are certain: (1) the isolated

verse in Anglo-Saxon corresponds to the parallel

distich in Hebrew; (2) the strong internal pause

in Anglo-Saxon to the end of the first parallel

period of the Hebrew distich; (3) there is a

correspondingly great irregularity in the number

of the syllables in successive lines of Anglo-Saxon,

and in successive distichs of Hebrew. Yet

whether the two poetical materials, the Anglo-

Saxon and the Hebrew, agree in what is after

all most fundamental in Anglo-Saxon, viz. the

constant quantity of stressed syllables in a verse,

and the constant ratio of the stressed syllables

in the two parts of a verse to one another remains


where (ed. Wright, i. 6442-6457) may serve for the comparison with

Hebrew poetry made above.

            On Good Friday I fynde • a felon was y-saved,

            That hadde lyvecl al his life with lesynges and with theftc;

            And for he beknede to the Gros, - and to Christ shrof him,

            He was sonner y-saved • than seint Johan the Baptist;

            And or Adam or Ysaye, • or any of the prophetes,

            That hadde y-leyen with Lucifer • many longe yeres,

            A robbere was y-raunsoned • rather than thei alle,

            Withouten any penaunce of purgatorie, • to perpetuel blisse.


            The most famous example in later English literature of rhythm

resting on equality in the number of accented syllables accompanied

by great inequality in the total number of the syllables is Coleridge's

Christabel. The accented syllables in the lines are always four;

the total number of syllables commonly varies, as Coleridge himself

puts it, from seven to twelve, and in the third line of the poem drops

down to four. For reference I cite the five opening lines--

            'Tis the middle of night by the castle clock,

            And the owls have awakened the crowing cock;

               Tu-whit !—Tu-whoo

            And hark, again ! the crowing cock,

               How drowsily it crew.

ELEMENTS OF HEBREW RHYTHM                    131


for consideration; the answer is not immediately

obvious, for Hebrew does not so unambiguously

and conveniently indicate what are the stressed

syllables in a line as does Anglo-Saxon by its

alliterative system. In many Hebrew lines we

cannot immediately see for certain either which,

or how many, are the stressed syllables: what

means exist for ultimately determining these

uncertainties in part or entirely I will consider

later. But first I return to a point already

reached in the last chapter.

            Even parallelism suggests a division of Hebrew

distichs into two broad types of rhythm: in one

of these two types the two parallel lines balance

one another, whereas in. the other the second

comes short of and echoes the first. No great

attention is required in reading Lamentations v.,

or Job xxviii., or many other passages in Job

or the Deutero-Isaiah, or many Psalms, such as,

e.g., li., in order to become aware of the dominance

and,' in some cases, of the almost uninterrupted

recurrence of balance between the successive

couplets of mostly parallel lines; nor, again, in

reading Lamentations ii., iii., iv. to become

aware of the different rhythm produced when a

shorter line constantly succeeds to a longer one.

So far we can get without any theory as to the

correct method, if there be one, whereby these

rhythms should be more accurately measured

or described, or as to the best nomenclature



wherewith to distinguish these differences when

we wish to refer to them. But if we get thus far,

it further becomes clear that, if we admit the

prevalence in Lamentations iv. of a clearly

defined rhythm fit to receive a name of its own,

whether or not the name kinah by which this

rhythm commonly goes be the best term to

define it, then Lamentations v. and Job xxviii.

also have, though a different, yet a no less clearly

defined rhythm whether we give it a name or

not; and of course, if we wish to discuss the

subject, we must find some convenient way of

referring to this rhythm no less than to the other.

            To distinguish these two broad classes of

clearly distinguished types of rhythm I have

suggested the terms balancing rhythm and echoing

rhythm.1 This terminology seems to me free

from some of the objections which attach to the

term kinah as a term for the echoing rhythm,

even if we could discover a good companion

term to kinah to describe the other type. As I

pointed out in the last chapter, kinah rhythm is

really a rather ambiguous term, meaning either

the total rhythmical effect of a poem in which a

particular echoing rhythm is prevalent, or that

particular echoing rhythm even though it be

confined to a single line or period. And one

serious disadvantage of the term kinah rhythm

lies in the ease with which it obscures the fact


            1 Isaiah ("International Critical Commentary"), i. p. lxiii.



that within the same elegy (kinah) or other

rhythmically similar poem more than one type of

rhythm as a matter of fact occurs.

            But whether even echoing rhythm and balancing

rhythm be a satisfactory terminology for the two

broad classes of Hebrew rhythm under which

sub-classes may be found, this broad fundamental

distinction itself is nevertheless worth keeping

clear ; it forms a comfortable piece of solid ground

from which to set out and to which to return

from excursions into the shaking bog or into the

treacherous quagmire that certainly needs to be

traversed before the innermost secrets of Hebrew

metre can be wrested and laid bare.

            In Lamentations v. a balancing rhythm, in

Lamentations iv. an echoing rhythm prevails ;

a rapid reading of the two chapters will suffice

to verify this general statement. But, if the

reader will re-read the chapters with closer

attention to details, he will probably feel that

Lamentations v. 2--

                        Myrzl hkphn vntlHn

           Myrknl vnytb

            Our inheritance is turned unto strangers,

                Our houses unto aliens,


differs not only in respect of its parallelism but

also of its rhythm from most of the other verses

in the same chapter, and also that, while it is

rhythmically unlike most of chap. v., it is



rhythmically like most of Lamentations iv.;

it is, for example, rhythmically unlike Lamenta-

tions v. 13


                        vxWn NvHF MyrvHb

           vlwk Cfb Myrfnv

            Young men bare the mill,

                 And youths stumbled under the wood;

it is, on the other hand, rhythmically like, e.g.,

Lamentations iv. 8--

                        glwm hyryzn vkz

                        blHm vHc

            Her nobles were purer than snow,

                Whiter than milk.


One or two other verses in Lamentations v. may

at first seem ambiguous : are verses 3 and 14,

for example, in balancing or echoing rhythm?

Again, in Lamentations iv., where the echo-

ing rhythm clearly and greatly prevails, a few

verses disengage themselves as exceptions; e.g.

verse 13

                        hyxybn tvxFHm

           hynhk tvnvf

            For the sins of her prophets,

                The iniquities of her priests,


gives the impression of balance rather than echo,

though the entire rhythmical impression is not

quite that which is left by the balancing rhythm

of Lamentations v.

            Thus, without any more detailed examination

or exacter measurement of lines, we reach the

important conclusion, which a close study of



Lamentations i. abundantly confirms, that the

same poem may contain distichs of different

metrical character.

            But within what limits may or do these and

other differences occur within the same poem ?

If that question is to be answered we must dis-

cover some principle of measurement which will

enable us to determine in less simple cases than

those just cited when the rhythm remains constant

and when it changes, and how.

            Is balance, then, due to (1) equality in the

number of syllables in the two lines, and echo to

inequality in the number of syllables ? If this

be so, then Lamentations v. 3,

                        bx Nyx vnyyh Mymvty

           tvnmlxK vnytvmx

            Orphans were we, without father,

                 (And) our mothers (were) as widows,


is in balancing rhythm, the number of syllables

in each line being eight.

            Or (2) is balance due to the sum of the metrical

values of all syllables in each line being the same,

even though the number of the syllables differs?

The number of syllables in a Latin hexameter

varies; but the sum of the metrical values of

the syllables must always be equivalent to six

spondees. If this were the true account of

Hebrew rhythm, it would become necessary to

determine what syllables are metrically long,

what short.



            Or (3) is balance due to equality in the number

of stressed or accented words or syllables in the

two lines, echo to the presence of a greater number

of stressed syllables in the first line, and a smaller

number in the second? If so, is there no limit

to the number of unstressed syllables that each

stressed syllable can carry with it? If there is a

limit, what is it? Is it no wider than in Christa-

bel? or is it as wide as, or wider than, in Anglo-

Saxon poetry?

            Of these three possibilities, the first two seem

to me to have been ruled out in the course of

discussion and investigation concerning Hebrew

metre. I confine myself to some discussion of

the third.

            It is just possible that some of the ancients

had analysed the laws of Hebrew poetry suffi-

ciently to detect the essential character of the

stressed syllables. The interesting suggestion

has been thrown out1 that the author of Wisdom,

who certainly attempted to naturalise parallelism

in Greek, also attempted a new Greek rhythm

on the model of the Hebrew by making the

parallel periods in Greek contain the same

number of accented syllables. Then, again, in

the opinion of some the difficult passage in Origen

which refers to the subject of Hebrew metre

implies an appreciation of the stressed syllables.2


            1 Encyclopaedia Biblica, col. 5344.

            2 Origen's scholion has already been cited above, p. 12 n. The

subject of the scholion is Psalm cxix. 1—



Be this as it may, there has certainly been an

increasing agreement among modern students of

this subject, particularly under the influence of

Ley,l to find in the stressed words or syllables the

"pivots or posts," to use Professor Saintsbury's

phrase, of the Hebrew rhythm.

            But allowing this, what is the limit—for there

surely must be some limit—to the number of

unstressed syllables that may accompany each

or any of the stressed syllables? Again, is there

any law governing the position of the stressed

syllable in relation to the unstressed syllables

that go with it?

            Taking the first of these two questions first:--

Does a single word extending beyond a certain

given number of syllables necessarily contain more

than one stress ? or is such a word ambiguous,

capable of receiving two, but capable also of

receiving only one stress ? And is the actual

number of unstressed syllables that may accom-


            1           jrk ymt yrwx

            hvhy trvtb Myklhh

which contains six fully stressed words and is rendered in the LXX--

                        Maka<rioi oi[ a@mwmoi e]n o[d&?,

                           oi[ poreuo<menoi e]n no<m& kuri<ou.

which contains six accents. Ley (Zeitschr. fur die AT. Wissenschaft,

1892, pp. 212 ff.) argues that one of the things which Origen is struggling

to express is that in this particular verse we find the unusual phenome-

non of text and translation containing the same number of stressed

words and consequently the same rhythm.

            1 Julius Ley, Die metrischen Formen der hebraischen Poesie, 1866;

Grundzuge der Rhythrnus, des Vers- and Strophenbaues in der hebraischen

Poesie, 1875 ; Leitfaden der Metrilc der hebraischen Poesie, 1887.



pany a stressed syllable neither less nor more

than the number of syllables in the longest

Hebrew word with inseparable attachments such

as a preposition at the beginning and a suffix at

the close?  In other words, is the general rule :

one word, one stress, to which words of more than

a certain number of syllables, say four, so far

form an exception that they may receive a second

stress? Or, to put it otherwise, in such longer

words may the counter-tone as well as the tone

count as a full stress ? I incline to the opinion

that by the rule that words of a certain length

may, but do not necessarily, receive a double

stress, we at least approximate closely to an

actual law of Hebrew rhythm. But there is a

second question : does every single word receive

a stress, or, as in several lines of Christabel,

may we in Hebrew poetry have not only several

syllables but also more words than one to each


            We obtain some light on both these questions

from certain characteristics of the Massoretic

punctuation, and on the second of them from

Assyrian analogy also. The effect of makkeph

in the Massoretie system is to render unaccented

any word which is thus joined to a succeeding

word. We may believe that the principle of

the Massoretic makkeph corresponds to a principle

in the ancient language without accepting every

particular use of makkeph in the Massoretic text



as corresponding to the intention of the original

writers. Nothing is more probable than that the

negative particle xl, conjunctions liken, and

other particles were frequently toneless: but

were they so regularly? If not, and if also we

cannot unquestioningly follow the Massoretic

punctuation, then an element of uncertainty

arises as to the number of stressed syllables in a

given line; for example, do the two lines in

Isaiah i. 3,

                        fdy xl lxrWy

           Nnvbth xl ymf

                        Israel cloth not know,

                           My people doth not perceive,


contain each three stresses (as in MT), or each

but two ? We cannot determine this off-hand.

If, indeed, we lay down the principle that two

stressed syllables must not immediately follow

one another, then the two xl's must be mak-

kephed, for in each line the syllable that precedes

xl is stressed; but it is decidedly dangerous

to lay this down as a. rigid principle, in spite Of

the strong tendency in MT to use makkeph in

order to avoid such concurrences. Modern Pales-

tinian popular songs, which have much that is

analogous to Hebrew poetry, according to the

express testimony of Dalman,l admit the con-

currence of two tone-syllables. And the import-


            1 "Zuweilen stossen auch zwei betonte Silben unmittelbar auf

einander," Palastinischer Diwan, p. xxiii.



ance of xl (not) in the two lines above cited

(for the antitheses to the two lines that precede

depend on it) rather strongly indicates that it

there received the stress in each line.

            But there are other combinations of words

that are frequently makkephed in the Massoretic

text ; for example, constructs and genitives.

Again the question arises : were such combina-

tions regularly read with a single stress ? if not,

has the MT always preserved a correct tradition

of the intention of the original writer ? We are

thus faced with another group of uncertainties.

These can perhaps be reduced by observing that

in MT there is a far greater tendency to makkeph

construct and genitive if the construct case is

free from prefixed inseparable particles such as

prepositions or the copula ; so, e.g., in Lamenta-

tions iv. 9 we find brH-yllH with, but brh yllHm

without makkeph.

            The Massoretic punctuation rests partly on

an ancient tradition, partly on an exegetical

theory, partly on an accommodation of the text

to a recent mode of reading it. It is valuable,

therefore, to have such principles as that the

negative particles are normally, and construct

cases often, toneless, supported by Assyrian


            In the Zeitschrift fir Assyriologie for 1895

(pp. 11 ff.) Zimmern published an interesting

Assyrian inscription, a poem as it appeared to be,



though since, as Dr. Langdon informs me, neither

Zimmern himself nor any one else has yet

succeeded in making a consecutive translation,

it may be in reality a succession of disconnected

verses written out in illustration of scansion.

In any case the important point is that here we

seem to have visualised a mode of scansion that

throws light on the composition of the feet or

rhythmical units in Assyrian, for these verses

are divided by longitudinal lines into four sections,

and by latitudinal lines into groups of eleven.

The longitudinal lines mark off into separate

compartments the four stressed syllables or words

with their accompanying unstressed syllables,

which here, as in most Assyrian and Babylonian

poetry, compose the line.

            I will briefly summarise the statements made

by Zimmern at the time, based on his first

examination of this document; these were ampli-

fied in a later article, to which reference will be

made below. According to Zimmern, then, the

following metrical facts are attested by these

scansion tablets:

            (1) Normally there is to one word, one stress;

but (2) the relative pronoun (monosyllabic in

Assyrian), the copula, prepositions, the negative

particles la and ul, and the optative particle lu

receive no stress, but go with the following word

to form a single-stress group of syllables; so

also (3) the status constructus and the genitive



generally receive but one stress — on the other

hand, if the second substantive has a pronominal

suffix they receive two; (4) two particles and a

word, or one particle and a word with a pro-

nominal suffix, form single-stress groups; (5) two

words expressing closely related ideas form a

single-stress group—e.g. abi u banti; (6) a voca-

tive may be inserted without being reckoned

in any of the four stress-groups that compose

the line.

            Though we make the most of the suggestions

from both sources, the Massoretic punctuation

of the Hebrew text and the scansion of the

Assyrian tablets, we shall still be left with a fair

range of uncertainty, and many lines of Hebrew

poetry will occur in which, judged by themselves,

the number of stresses will remain ambiguous.

And that ambiguity will be still further increased

when we attempt to determine what single words,

if any, may receive two stresses; here again some

light is cast on the possibility of such double

stress by the Massoretic punctuation; for as

the effect of makkeph is to bring two or more

words under one tone, so the effect of metheg is

to indicate the presence in the same word of two

tones, of a counter-tone in addition to the main

tone. But there is no probability that all the

counter-tones marked by metheg, such, for example,

as the first syllable in forms like UlF;qA, really

received a stress; and for this theory of double-



stressed words we receive, I think, no very

helpful analogy from Assyrian.

            The question, then, arises : Can we discover

a more accurate method of determining the

limits of what may accompany a stressed syllable?

It is the attempt to answer this question that

occupies in the main the attention of recent

theorisers on Hebrew metre, and it is in the

attempt to answer it that they diverge from one


            The popularity which for a time was enjoyed

by Bickell's1 system has waned in favour of that

of Sievers, which has the advantage of being very

much more elaborately and systematically worked

out. I propose very briefly to summarise some

of the chief points in Sievers' system, premising

at the outset that if it could be held to be estab-

lished it would (1) greatly reduce, though not

entirely eliminate, lines of ambiguous measure-

ment; and (2) give for every line, regarded by

itself independently of its association with any

other line, a clear rhythmical definition.

            In connexion with the present discussion the

two fundamental laws of Sievers' system can,

perhaps, best be stated thus: (1) the number

of unstressed syllables that may accompany a


            1 Gustav Bickell, Metrices biblicae regulae exemplis illustralae (1879);

Carmina Veteris Testamenti metrice (1882); Die Die/dung der Hebrder

(1882). The English reader will find a useful summary of Bickell's

system in W. H. Cobb, A Criticism of Systems of Hebrew Metre, pp.




stressed syllable must never exceed four, and only

in a particular type of cases may it exceed three.

Corollary: every word containing more than

five syllables must have two stresses. (2) The

stressed syllable regularly follows the unstressed

syllables that accompany it; and more than a

single unstressed syllable may never, follow the

stressed syllable that it accompanies.

            Using the term anapaest not of course of a

combination of two short followed by a long

syllable, but of two unstressed syllables followed

by one that is stressed, Sievers claims that the

Hebrew rhythm rests on an anapaestic basis,

and that the normal foot is

                                    x x __

examples of such feet being MykirAd;, UlF;q;yi, yneB;-lfa

Possible variations of the normal foot are--

                        (1) x x x __

                        (2) x __, and even

                        (3) __

            Moreover, since the stress may fall on a syllable

which with an additional and secondary short

syllable corresponds to an original single syllable,

as in the segholates, further variations are x x  _’_ x,

x x x _’_ x , etc., an example of such feet being



            1 After Sievers had indicated his theory in outline, Zimmern (Zeit-

schrift fur Assyriologie, xii. 382-392) returned to the examination of the

scansion tablets referred to above (p. 140 f.), and found that between



            If this theory be entirely sound, or even if it

closely approximate to the truth, it will consider-

ably diminish the range of uncertainty that must

remain so long as we leave entirely undetermined

the limits of the unstressed syllables that may

accompany a stressed syllable. This may be

illustrated by an example: how many stressed

syllables are there in each of these lines in Psalm

i. 1,

                        dmf xl MyFH jrdbv

                        bwy xl Mycl bwvmbv


            The question turns on the treatment of xl;

was it stressed or unstressed? The Massoretic

punctuation leaves the negative in each line

disunited from the verb and therefore capable at

least of being stressed; and Dr. Briggs1 in calling

the lines tetrameters certainly allows a stress

to each xl. I think it may be urged against this


two stressed syllables at least one, generally two, and not rarely three

unstressed syllables occurred, but never or quite rarely more than


            It may be worth while adding here that Dalman (Palastinischer

Diwan, p. xxiii, with footnote) has found that, in the modern Palestinian

(Arabic) poems that follow not a quantitative but an accentual system,

one to three, and occasionally four, unstressed syllables occur between

the stressed syllables. The value of these Palestinian analogies lies

in the fact that we are dealing not with speculations as to how a written

poem was or could be pronounced, but with the manner in which hither-

to unwritten poems were actually read to the editor who committed

them to writing.

            1 It so happens that I have mainly referred to details in Dr. Briggs'

work with which I disagree ; the more reason, therefore, that I should

recall the fact that in the subject with which I am now dealing Dr.

Briggs was a true pioneer, and that he was one of the first writers in

English to insist on the fundamental importance in Hebrew prosody

of the stressed syllable.



that xl has nothing like the need of emphasis

and stress here that it has in the lines previously

cited from Isaiah i. 3, where fdy xl is antithetic

to fdy in the previous distich. I should therefore

think it most probable that the lines were three-

stressed and not four-stressed ; but apart from

the' bearing of the rest of the Psalm on the question

we cannot determine the point unless we are

justified in calling in such a theory as that of

Sievers. Now it is perfectly true that even on

that system monosyllabic feet are possible, and

that xl in particular at times, as in Isaiah i. 3,

stands by itself as a foot; but if the anapaest is

the basis of the rhythm, we cannot naturally

divide each of the two perfectly normal anapaests

fdy-xl and bwy-xl into a monosyllabic and a

dissyllabic foot; on Sievers' theory the only

natural way of reading the two lines is with

three stresses; they are, to use Dr. Briggs'

terminology, trimeters, not tetrameters.

            Sievers' theory, then, if established, would

reduce the number of lines which, measured. with

exclusive reference to the stressed words or

syllables only, are ambiguous. Is the theory,

then, as a matter of fact, so firmly establish ed on

perfectly certain data that it does actually

diminish the number of uncertainties that are

left when we attempt to count stressed syllables

simply without very closely defining either the

position which such stressed syllables must



occupy, or the number of unstressed syllables

which may accompany them? I doubt it. I

cannot here undertake any examination or criti-

cism of Sievers' long and exhaustive exposition

of his theory; nor can I examine the arguments,

worthy as most of them are of the closest atten-

tion, by which he supports certain theories of

vocalisation on which his metrical system rests.

But these theories, however much may be said

for some of them, are not all of them as yet so

certainly established as to allow the metrical

system, which in part suggests them, but which

also certainly rests upon them, to furnish a

sufficiently sure instrument for eliminating the

uncertainties that arise when we measure a

Hebrew text by the stressed syllables only. The

degree of uncertainty which the theory would

remove is largely counterbalanced by the in-

security of the basis on which it rests.

            In illustration of what I have just said it must

suffice to refer `to a few classes of the conjectural

vocalisation adopted by Sievers, all of which are

more or less essential to the smooth working out

of his system.

            (1) Partly on general phonetic grounds, partly

from actual features of the Massoretic vocalisa-

tion, such as the alternative forms of the type

MykilAm;.la and MykilAm;la, and the complete abandon-

ment of the reduplication and also of the following

syllable in such inflexions as Brijlzu from NroKAzi, tOnyog;wi



from NOyGAwi, Sievers infers that regularly when,

owing to inflexion, the full vowel after a re-

duplicated consonant is lost, the reduplication

and also the vowel that followed it were entirely

lost also; and that, for example, Myklml was

always pronounced lamlachim in three syllables,

never lammelachim in four, and yhyv always waihi

in two syllables (cp. ydeymi not ydey;.mi), and never

wayehi in three syllables.

            (2) Again, the consonantal text of the Old

Testament distinguishes two forms of the second

person perfect alike in the masculine and the

feminine. The second person masculine is gener-

ally of the form tlFq, more rarely of the form

htlFq, and again the feminine is generally tlFq,

and more rarely ytlFq. According to the received

vocalisation, the masculine, however spelt, was

pronounced katalta, and the feminine katalt.

Sievers, however, treats both the rarer forms

htlFq and ytlFq as trisyllabic, pronouncing them

katalta and katalti respectively; and he treats the

more frequent form tlFq, alike whether masculine

or feminine, as dissyllabic, pronouncing it katalt.

            (3) Certain pronominal forms were originally

pronounced with a syllable less than in MT ; thus

MT 1.T7, pausal j~d,yA, has replaced j`dAyA;  cp. such

forms in Origen's Hexapla as hxalax = j~l,kAyhe, bax=

j~b;, and in Jerome goolathach = j~t,lA.xuG. And it is

also argued that the endings hA-,, hA-u were once




It will be seen from the foregoing examples

that the tendency of Sievers' vocalisation is to

reduce the number of syllables below the

number produced by the received system. Con-

sequently what I stated as the first funda-

mental law of his metrical system, viz. that

not more than four unstressed syllables may

under any circumstances accompany one stressed

syllable, often means not more than five stressed

syllables counted according to the received


            One other of Sievers' theories with regard to

the pronunciation of Hebrew poetry must also

be noted; it works in an opposite direction, and

is designed to supply unstressed syllables when

their absence would be too keenly felt. Sievers

admits monosyllabic feet, but he abhors the

concurrence of two stressed syllables; he calls

to his aid the analogy of singing: as in singing

a single syllable is sung to more than one note by

virtually repeating the vowel sound, so Sievers

postulates that when tone-syllables appear to

follow one another immediately the long tone-

syllable was broken up into two in pronuncia-

tion; e.g. in such circumstances xl was pro-

nounced not lo, but lo-o, and lvq not kol, but

ko-ol, and the metrical foot is in each case not

__ but x __.

            Two things seem to me to gain probability

from Sievers' exhaustive discussion, even though



the elaborated system rests on too much that is

still uncertain or insecure: (1) the natural basis

of Hebrew rhythm is anapaestic rather than

dactylic; this is really an obvious corollary from

the regularity with which the Hebrew accent

falls on the last syllable of words, and the in-

frequency of detached monosyllables, and earlier

metrists also have for the most part detected a

prevalence of anapaestic or iambic rhythm in

Hebrew; (2) in the union of two or more words

under one stress, and in the distribution of long

words among two stress groups we should be

guided by the principle that the stress groups

within the same period are likely to be not too

dissimilar in size and character; and in general

it is safer to proceed on the assumption that

particles like yk, lf, etc., rarely receive the stress

unless for some reason an actual sense-emphasis

falls upon them.

            The sum of the whole matter is that we are

left with an instrument for the measurement of

rhythm capable of doing some service, but much

less delicately accurate, or much less clearly

read, than we could wish. With this instrument

we must work at the difficult question, which I

have so far merely indicated, but which I shall

examine more closely in the next chapter: What

limits, if any, are set to the number of different

rhythms that may be introduced into the same


ELEMENTS OF HEBREW RHYTHM                    151


            In concluding the present chapter I will

consider one further possible, and even probable,

service which it appears to me that parallelism

may render in reducing the element of uncertainty

in determining the rhythm of particular lines.

In Anglo-Saxon, alliteration clearly distinguishes

three of the stressed syllables in a line, leav-

ing only the fourth outwardly undistinguished;

Hebrew has no such outward indication of this

all-important element in the rhythm; in par-

ticular all particles, all construct cases, and

some other types of words are rhythmically

ambiguous; in any given line they may be

stressed or they may not. What I suggest is

that parallel terms tended at least to receive the

same treatment in respect of stress or non-stress.

I will give one or two illustrations of the value

of this law if its probability be admitted. If we

take by itself the line (Isa. i. 10),

                        Mds ynycq hvhy rbd vfmw

            Hear the word of Yahweh, ye judges of Sodom,


we may certainly be in doubt whether hvhy rbd

received one stress or two, and whether the whole

line was read with four stresses or five. Sievers

gives it but four, and thereby in its context, as

I believe, treats it wrongly. I suggest that rbd

(word) ought to receive the same metrical value

as its parallel term trvt (law) in the completely

and symmetrically parallel line or period that



follows, and that we should read both periods

alike with five stresses

                        Mds ynycq hvhy rbd vfmw

           hdmf Mf vnyhlx trvt vnyzxh

            Hear the-word of-Yahweh, judges of-Sodom,

            Give-ear-to the-law of-our-God, people of-Gomorrah„

                        xFvH yvg yvh

           Nvf dbk Mf

                        Ah! sinful nation,

                        People laden with iniquity.

This Sievers reads thus-

                        xFvH-yvg yvh

           Nvf dbk-Mf


and so far observes the rule which I am suggesting

that he leaves both the parallel terms yvg and Mf

unstressed; on the other hand, Nn1rr and its

parallel Nvf dbk do not receive the same treat-

ment, though they are quite capable of so doing.

A more probable reading of the lines will be


                        xFvH yvg-yvh

           Nvf-dbk Mf


                        xFvH yvg |  yvh

           Nvf-dbk Mf

I take as a last example an apparent exception

to the law. Lamentations i. 1 reads—



            Mf ytbr ryfh |  ddb hbwy hkyx

            Myvgb ytbr |  hnmlxk htyh

            sml htyh |  tvnydmb ytrw

How cloth she sit solitary, |—the city (once) great in popula-


She is become like a widow, | she that was great among the


She that was mistress over provinces, | she hath been (set)

                                                                                    to forced labour.

            Budde suspected ryfh, the city, in the first

line on the ground that at present the second

half of the first line contains three stresses,

whereas it should only contain two. Sievers

removes the ground for suspicion by treating

Mf-ytbr, great in population, together as a single

stress. At first this seems, by making ytbr,

great, unstressed, to give a term in the first

line a metrically different character from that;

of corresponding terms, ytbr and ytrw, mistress,

in the second and third lines. But the parallelism

of in the first line with ytbr in the second

and ytrw in the third is, as a matter of fact, not

complete; the real parallel in the first line to

ytbr, great, in the second line and ytrw, mistress,

in the third is not ytbr by itself but Nf ttbr, great

in population, i.e. populous, which, so taken

together, is also an antithetic parallel to the

single-stressed word ddb, solitary, in the first half

of the line; it is only when taken together

that the words Mf ytbr express the idea in the

mind of the writer, viz. the populousness of the



city, whereas ytbr in the second and irnm in the

third line sufficiently express by themselves the

ideas of the "great lady" (in antithesis to "the

widow.") and "the princess"; Myvgb, among the

nations, and tvnydmb, over provinces, respectively

serve merely to amplify the two ideas. The

distinction between Mf ytbr and Myvgb ytbr is

shown grammatically by the difference in con-

struction; and the writer probably allowed him-

self to repeat the same word 'inn in the two lines

instead of using two different and synonymous

terms on the same kind of principle as that of

the well-known law of Arabic poetry that the

same word may be repeated in the course of a

poem as the rhyme word, provided that the word

is used on the two occasions with some difference

of meaning.

            Thus, perhaps, a close examination of Lamenta-

tions i. 1 confirms, rather than reveals an excep-

tion to, the law which I have suggested, and

incidentally shows that ryfh is not merely metric-

ally possible, which Budde had denied and which

is all that Sievers claimed, but metrically required.















                        CHAPTER V































                                    CHAPTER V




HEBREW rhythms fall into two broad classes

according as the second line of the successive

distichs is equal in rhythmical quantity to, and

therefore balances, the first line, or is less in

quantity than, and so forms a kind of rhythmical

echo of, the first line. Distichs in which a shorter

first line is followed by a longer second line are

relatively speaking so rare1 that in a first broad

division they may well be neglected; and we

may classify the great majority of rhythms not

merely as distichs consisting of equal or unequal

lines, but, so as to bring out the regular and more

striking difference between them, as balancing

and echoing rhythms respectively.

            But before we can discuss the question of the

extent to which, or the sense in which, strophe

may be said to be either a regular or an occasional

form of Hebrew poetry, it becomes necessary to

subdivide these two broad classes of rhythms


            1 Examples are given below, pp. 176-182.




which have hitherto mainly engaged our attention,

and then to consider to what extent different

rhythms may enter into one and the same poem.

This subdivision must be carried through by

applying a measure which, as I have pointed

out in the previous chapter (p. 150), is less

accurate than we could desire, and leaves us with

corresponding uncertainties which must not be

forgotten. Even when we may be certain of the

general class into which a particular distich may

fall we may remain uncertain of its exact measure-

ment; for example

                        fdy xl lxrWy

           Nnvbth xl ymf

a distich which occurs in Isaiah i. 3, is certainly

a distich of equal lines (balancing rhythm) : but

whether each line contains three or only two

stressed words is, as we have already seen (p. 139),

in some measure uncertain.

            Whether the unit in Hebrew poetry is the line

or the distich has been much discussed; regarded

from the standpoint of parallelism, it is obviously

the distich that is the unit; the single line in this

case is nothing; it is incapable of revealing its

character as a parallelism. On the other hand,

it is rhythmically just as easy to measure a single

line as to measure a distich ; and at times it is

necessary so to do: for, as there alternate with

distichs that consist of parallel lines distichs that

            VARIETIES OF RHYTHM   159


contain no parallelism, so occasionally there

alternate with these distichs single lines or mono-

stichs, and also tristichs in which one of the three

lines may or may not be parallel to the other

two. For these non-parallel isolated stichoi, or

the third stichoi of tristichs, measurement of the

line becomes necessary.

            At the same time, unless an anapaestic rhythm

such as Sievers claims to discover, or other

rhythm equally well defined, can be shown to

prevail within the lines, these isolated stichoi

owe their rhythmical character, so far at least

as we can discern or measure it, to the fact that

they contain the same number of stressed syllables

as the halves of the distichs among which they


            Thus in any case the distich remains so char-

acteristic of Hebrew poetry that it is better, so

far as possible, even in a rhythmical classifica-

tion, to measure and classify by distich rather

than stichos: though the stichos when isolated

will of course call for measurement too.

            Distichs consist of (i.) those in which the lines

are equal; and (ii.) those in which one line

(generally the second) is shorter than the other.

            The first class of distichs subdivides into

(a) distichs with two stresses in each line, for

which we may use the formula 2 : 2 ; (b) distichs

with three stresses in each line (3 : 3) ; and (c)

distichs with four stresses in each line (4:4).



Of these three types of balancing rhythm the first

and third are intimately connected: for four-

stress lines are commonly divided into two equal

parts by a caesura, and the pause at the caesura

is often strong enough to justify, regard being

had to rhythmical grounds alone, treating each

period of four stresses as a distich of two-stress

lines. Any isolated group of two periods of four

stresses is best classified as a single distich of

four-stress lines, or two distichs of two-stress

lines, according as parallelism occurs between the

clauses or sentences of two stresses or of four

stresses. But in view of this intimate connexion

it is not surprising that combinations of two

two-stress clauses or sentences, and combinations

of two four-stress sentences, occur in the same

poem. Such a mixture of rhythms, if in such

case we are right in speaking of a mixture of

rhythms at all, exactly corresponds to the fact

that, in the same kinah or elegy, parallelism

sometimes occurs between the two unequal

sections of three and two stresses respectively,

and sometimes does not; in the latter case we

may, if we will, speak of a line of five stresses, and

in the former of a distich in which a two-stress

line follows a three-stress line; but the line in

the one case and the distich in the other are

rhythmically identical, since each contains five

stresses ; there is no real change in the rhythm,

though the change in the parallelism introduces

            VARIETIES OF RHYTHM   161


a markedly different effect 1 which it is well to

render as manifest as possible.

            If, at least where parallelism commonly takes

place between sections of three and two stresses

respectively, we more properly speak of a distich

of unequal lines than of a line of five stresses,

then clear examples of distichs of two-stress lines

are those which interchange with the 3 : 2 distichs

in Lamentations i., iii., iv. : as, for example,


                        Myrvrmb ynfybwh

                hnfl ynvrh

            He hath filled me with bitterness,

                 He hath sated me with wormwood.


            However we choose to term them, combinations

of parallel clauses of two stresses do, as a matter of

fact, interchange within the same poem with dis-

tichs of four-stress parallel lines : so, for example,

in 2 Samuel i. 22--

                                    MyllH Mdm

                Myrbg blHm

     rvHx gvwn-xl Ntnvhy twq

     Mqyr bvwt-xl lvxw brHv

            From the blood of the slain,

                From the fat of the mighty,

            The bow of Jonathan turned not back,

                And the swore of Saul returned not empty.


For are we not forced by the parallelism to place

a much greater pause between the first two sets


            1 Cp. e.g. Isaiah i. 10 f., 18-20, 21-26, and see Isaiah (" International

Critical Commentary "), p. Ixvi (Introduction, § 54) ; see also ibid.

pp. 4 f., 26, 31.



of two words than between the next two sets, at

the end of the first of these four lines than in the

middle of the third or fourth line? And are not

the two short parallel periods really separated

by almost as strong a pause as the two longer

ones that follow? If we call the two longer ones

a distich of four-stress lines, why not the two

shorter ones a distich of two-stress lines? Does

not the passage really consist of two distichs

rather than of a single tristich (cp. R.V.) of three

four-stress lines ?

            For another example of this combination we

may turn to Isaiah xxi. 3—1

                        hlHlH yntm vxlm Nk-lf

           hdlvy yryck ynvzHx Myryc

                fmwm ytyvfn

                tvxrm ytlhbn

            Therefore filled arc my loins with writhing,

                 Pangs have seized me as of a woman in travail.

            I am bent (with pain) at what I hear,

                 I am dismayed at what I see.


Here the first two periods must be regarded as a

distich of four-stress lines : the lines cannot be

subdivided into distichs of two-stress lines as

which so much of the rest of the poem may be,

and, indeed, is best read.2


            1 Cp. Isaiah, pp. 348 f.; also my article, "The Strophic Division of

Isaiah xxi. 1-10, and xi. 1-8," in the Zeitschr. fur die AT. Wissenschaft,

1912, pp. 190 if.

            2 The existence of two-stress lines in Isa. xxi. 1-10 is, indeed, denied

by Lohmann. In the Zeitschr. fur die AT. Wissenschaft, 1912, pp.

49-55, he had urged, and in reply to my criticism (contained in the

article mentioned in the previous footnote) he maintains (in the same

            VARIETIES OF RHYTHM   163


            Which is the best way to divide the Hebrew

text, or even an English. translation, though this

at least should as far as possible be divided

according to the parallelism, often becomes a

delicate question. For example, does

                        tvklmm vFm Myvg vmh

            (Ps. xlvi. 7) Crx gvmt vlvqb Ntn

consist of one distich of four-stress lines incom-

pletely parallel to one another (so R.V., v. 6)? or

of two distichs of two-stress lines, the lines in the

first distich being completely parallel, the lines

in the second not parallel at all? Thus--

            Nations were in tumult,

                 Kingdoms were moved;

            He uttered his voice,

                 The earth melted.


            If Psalm xlvi. 7 be treated as a single distich,

then the first line of the distich is marked by an

internal and secondary parallelism; and it is


journal, 1913, pp. 262-264), that the whole of this poem except vv. b

and 9 originally consisted of four-stress periods, and that vv. 8 and 9

consisted of five six-stress periods, each equally divided by a double

caesura into three two-stress sections. But this theory rests on textual

emendations that appear to me to lack support independent of the

theory itself. I should not very confidently maintain that v. 10 must

be in its original form; but it is surely very precarious criticism to

argue that because the words n'rsrr nos are absent from the LXX

in v. 5, therefore two other words in the same verse, viz. hvtw lvbx,

were absent from the original text., and that the words absent from the

LXX were present in the original text. Nor again can the words,

"eating, drinking" be dismissed as "trivial." It is distinctly more

probable that the princes were bidden to rise after the banquet had

begun rather than while the tables were still being laid. But while in

this detail I differ from Lohmann, I repeat what I said in my article,

that his discussion is in the main a valuable criticism of Duhm's mis-

taken treatment of Isa. xxi. 1-10.



to be observed generally that the well-defined

caesura which regularly occurs in four-stress

periods renders it particularly easy for the halves

to receive such secondary parallelism, and so to

assume, when isolated, an appearance of greater

independence. Whatever view we take of par-

ticular examples, whether we break them up

into distichs of two-stress lines or distichs of

four-stress lines, the rhythm remains essentially

the same, and our only problem is how best to

do justice to other formal elements in the poem

which differentiate what are, in the last resort,

rhythmically identical periods. There is nothing

that is peculiar to Hebrew poetry in this particular

kind of uncertainty which is produced when,

within a rhythm that remains constant, another

poetical form is irregularly followed. A popular

metre with English poets in the sixteenth century

was the " poulter's " measure, in which lines

of twelve syllables alternate with lines of a

“poulter's” dozen, i.e. of fourteen syllables;

these long but unequal lines rhymed.l Divide

the twelve-syllable line of the poulter's measure

in half, and the fourteen-syllable line into lines

of eight and six syllables respectively, supply the

four short lines thus produced with two sets of


            1 Four lines of Grimald in Tottel's Miscellany (ed. Arber, p. 110)

may serve as an example :

            Of all the heavenly gifts that mortal men commend,

                 What trusty treasure in the world can countervail a friend?

            Our helth is soon decayed; goods, casual, light and vain;

                 Broke have we seen the force of power, and honour suffer pain.

            VARIETIES OF RHYTHM   165


rhymes instead of one so that they rhyme alter-

nately, and the form of the typical short metre

of our hymn-books is the result. But in some

cases the origin of short metre asserts itself, and

within the same hymn the first and third lines

sometimes rhyme and sometimes do not; as,

for example, in these two consecutive verses of

Wesley's translation of Gerhardt's hymn

            Give to the winds thy fears,

               Hope and be undismayed;

            God hears thy sighs, and counts thy tears,

               God shall lift up thy head.


            Through waves and clouds and storms

                He gently clears thy way:

            Wait thou His time; so shall that night

                Soon end in joyous day--


and so throughout the hymn, though in no

regular alternation, we may observe rhymed and

unrhymed first and third lines. Rhythmically

the two long lines of the old poulter's measure

and the four short lines of modern short metre

'are identical: where rhymes regularly mark off

the shorter periods, it is obviously convenient

to make this prominent by dividing into four

lines ; but where the first and third sections

only occasionally rhyme, either course might be

adopted : and so with a Hebrew poem in which

parallelism sometimes, but not invariably or

even predominantly, exists between the halves

of successive periods of four stresses.

            Yet, clearly allied as 2 : 2 and 4 : 4 are, at



times it makes some difference whether we treat

the passage as in the one form or the other;

the main difference lies here, that in ambiguous

cases we shall naturally give to the separate lines

of what we regard as a distich of two-stress lines

a greater independence than if we were to regard

these two-stress clauses as merely parts of a

single four-stress line. I take as an example

Psalm xlviii. There are in this Psalm, as is

well known, some difficult phrases and some

doubtful text, but the presence of several short

parallel clauses, enough, I think, to be charac-

teristic of the poem, is certain: on the other

hand, in the present text there is no single clear

case of parallelism between four-stress periods.

This being so, verse 4 (RN., v. 3) ought, I believe,

to be taken not as a single four-stress line (R.V.),

but as a distich 2 : 2; it consists of two independ-

ent parallel lines--

                        hytvnmrxb Myhlx

                bgwml fdvg