Anul III, Numărul 2/2005


The Study of Classical Hebrew Poetry: Terminological Issues

Silviu TATU


When terminology is concerned there have been three main solutions offered so far. The first one used to compare Hebrew poetry (henceforth HPy) with POETRY from other world literatures more or less cognate with Hebrew and evaluate them from the perspective offered by the POETRY theory at use at a given time. The second solution belongs to the other extreme, since it refutes any foreign tenor upon the Hebrew literature, preferring the original Hebrew terminology. In between, a more ponderable solution offered to prefer neutral terms, since the Graeco-Roman terminology was indeed alien to the Hebrew literature, but Hebrew original terminology was not known anyway. Therefore even terminology of Hebrew inspiration, is as good as any other one when original meaning of literary categories is concerned. Subsequently, we will proceed to survey them and propose the terminological set we embrace for the present research.


The classical solution

Originally owing much to Classical Graeco-Roman literary criticism, this method influenced for centuries the perception of HPy and the possibility of mutual interchange was traced as if the older influenced the more recent one. Nevertheless, the terminology of the recent one was imposed on the older one. Thus, terms such as stich and hemistich (Gr. stichos) were involved to define a line of a verse and a half line respectively. A colon (Gr. kolon, Lat. colon) named a section of a metrical period in quantitative verse, consisting of two to six feet and in Latin verse having one principal accent. Most frequently, verses were composed in tetrameters, pentameters and hexameters. Later on, other names came into use such as limb or member. A structure of two cola (pl. of colon) made a bicolon and a three cola structure made a tricolon.

Later on, when Jewish literature flourished under the beneficent wind of change brought by the Arab culture, a new POETRY, which was inspired mostly from Greek via Arab translations of the great Greek classics, tended to be created. Jewish scholars encompassing gaons (i.e. leaders of the great universities in Sura, Pumbedita, and later in Baghdad, Damascus, Cairo), poets, medical doctors, and court educators used to read the Greek philosophers in the original language. Therefore, we can safely admit the great influence Greek culture had on medieval Jewish culture. Consequently, a poem came to be called piyyut (Gr poietria), and its author paytan (Gr poietes), i.e., a poet.

The works of these great Jewish poets and critics were eventually read by Christian scholars, who considered them as their legacy of POETRY theory. Thus is explained the interest Robert Lowth had shown in HPy, namely as an indirect effect of reading the Jewish literary critics, such as Azariah de'Rossi (16th cent.), via Johannes Buxtorf who translated and incorporated de'Rossi's POETRY theory in his critical studies.[1] Consequently, when Lowth rediscovers parallelism, he names it according to the Graeco-Roman critical studies to which he was so indebted, i.e., parallelismus membrorum (Lat. 'parallelism of members').[2]

Recent criticism of HPy refers to different lexica, indebted to the linguistic approach promoted by each biblical scholar in accordance with his personal skills. The imminent danger is that whenever lacking the proper background, new terms, categories, paradigms and taxonomies might be proposed, at large of a rare confusing character. It is not our interest to present them all or even some of them, in order to support this conclusion. Nevertheless, some will appear clearly in the present paper. Unless necessary we will refrain from inventing new terminology.


The vernacular solution

Another solution was that offered by Jewish authors who tried to emancipate POETRY theory of HPy from its Graeco-Roman parallels. As early as the Middle Age, there were some who suggested that new terms were to be prefered: bayit for a verse of poetry, composed of two hemistichs, of which the first was called delet, and the second soger. Pismon stand for refrain (Berlin 1991, 173-4; cf Fishoff 1972).

Trying to solve the thorny issue of justifying the presence of POETRY in the middle of 'PROSE' texts, as part of the solution offered, Weitzman opts for new terminology, presumably neutral this time, both in terms of accuracy and slippery references to modern criticism. Although narrative has already been accepted by biblical scholars as equivalent to 'PROSE', 'fiction', or 'history', it might still perpetuate a sense of complexity which was not necesarily intended by the original authors. Therefore, the elementary term story is preferred. As far as POETRY is concerned, the search for neutrality is not that simple. However, song might make a perfect pair to story, according to Weitzman. The term was previously advanced by Geller (1982, 70), who opposed it to the elaborate, however unnatural, presention of Hebrew verse that O'Connor proposed. Inspite of its supposed neutrality, both song and story have their coverage in Hebrew terminology (cf. toledot and šîr).

It is suggested that the difference started in aesthetics, the song being the expression of subjectivity, but it was perfected as such by the final editor responsible for the version that got through the final stage of Scriptural canonization, named shortly scripturization (Weitzman 1997, 3, 13).[3] However, there is a small problem with this distinction between terms. Just as story covers different genres, so song refers to more than one genre. On one hand, form critics will prefer to include in the category of story the toledot narratives from Genesis, the heroic narratives from Exodus, Judges, Samuel and Kings, the genealogical lists from Ezra-Nehemiah, Chronicles, and the legal texts from Leviticus, Numbers or Chronicles.[4]

On the other hand, textlinguists will speak of paragraphs of one of the following natures: narrative, predictive, expository, and hortatory (Longacre 1989, 59-63).[5] As far as POETRY is concerned this differentiation seems very unlikely to have been clarified so far, due to the difficulty in understanding the Hebrew terms used by the original author/editor to define their own work: hdh, mzmwr, mšl, mktm, mśkyl, šgywn, šyr, šyrh, tplh, thlh, possibly even khzwn, or mśh, etc. (Schökel 1988, 8-10). The lack of theoretical ground to relate the 'poetic genres' among themselves is also sensed by Kugel (1984, 110) as the major disruption of the discussion about the PROSE-POETRY distinction. Since our study has the duty of evaluating the style of different poems, we are not constrained from challenging previous terminologies, but in searching to rediscover the meaning of literary terms in an ancient literature such as Hebrew, we prefer the most neutral terms possible to start with.


The midway solution

Hrushovski (1972, 1200) favoured the term line for the main unit of the poetic utterance, and verset for its half-line constituents, a complex relationship of parallelistic nature holding the parts together.[6] Kugel also preferred leaving aside concepts inspired by Graeco-Roman rhetoric, such as stich, hemistich, colon, semicolon, bicola, tricola, membrum, and promoted neutral terms such as line or clause (A and B), larger units being called binary rhymes, when two clauses are involved, or ternary rhymes, when a supplementary third clause appears. Clause has no syntactic reference for Kugel, as he tenaciously rejects O'Connor's theory.[7]

Korpel and de Moor (1988, 1-4) advanced the idea that the basic unit of Ugaritic and HPy is the colon, the unit of the oral rendering of the poem which developed into a couplet only later. However, the smallest building block is the foot (i.e. a word containing at least one stressed syllable). O'Connor (1980) referred to a half-line as a line and promoted it as the basic unit of his syntactic theory of Hebrew verse. Postulating the line as the basic unit of POETRY, O'Connor bears the guilt for the forbidden atomization of Hebrew verse. Such a narrow definition of a Hebrew line becomes acceptable only if strict metre could be established for Hebrew prosody, but O'Connor did not seem to have believed in it (Geller 1982, 71). Against O'Connor's theory of syntactic verse comes the growing consensus among the sholars that the monocolon was not the original structural unit of HPy in particular and Semitic poetry in general, and the 'gapping' device so much involved in parallelism, in many instances the verb itself, or the subject (the basic units of a sentence) is missing.

Once believed to be uniquely characteristic of HPy, parallelism is nowadays accepted not only as a traditional rhetorical device present in some classic literatures, but also as a common feature of literary writing. Whether quantitative in terms of metre and elements of clause structure, or qualitative in terms of phonetics and semantics, parallelism fascinated Biblical commentators even before Bishop Lowth advanced his theory. Though traced in many world literatures, its presence in Hebrew literature surpasses its usage in world literature. Since Hebrew verse hardly can exist without it, it was inferred that in ancient literatures such as Hebrew and Ugaritic, parallelism is not a mere rhetorical device but an organizing principle, which ties together two clauses in a meaningful unit called a line.[8] Consequently, if not rhetorically, parallelism should be considered either syntactically or semantically. Furthermore, the syntactical theory of parallelism O'Connor promoted did not seem to gain overall scholarly support. It would be very difficult to prove that parallelismus membrorum functions at the syntactic level in HPy. His theory was applied first on prophetic texts, which are well known as being inspired by the true HPy found outside the Canonical oracles.

Baruch Margalit proposed a more coherent terminology, well defined and tested as a working theory both in Hebrew and in Ugaritic literature. Briefly, these are some basic definitions to be considered:      1. The verse represents the main unit of POETRY; it corresponds to the sentence in PROSE, meaning that it is at once a grammatical and a prosodic entity[9] . The basic verse-type is the binary structure. Monoline and ternary structure are its structural variations.      2. The verse is subdivided by principal caesura(e) into verse-lines (A and B).      3. The verse-line is subdivided by blanks into verse-units, mostly co-extensive with the individual vocable or lexeme. There is a tendency towards limiting the number of verse-units per verse-line to three, but there are exceptions. The number of verse-units/verse is called the 'valence' of that verse. The pattern of a verse according to its corresponding lines valence is represented as the numerical sum of the verse-units in the component lines (e.g., 4 + 3; 3 + 3; 3 + 2). Scansion is called the process of determining verse valence.      4. The strophe is a conventionally structured sequence of verses in indeterminate quantity. The basic verse pattern (the valence of most verses) represents the strophic 'theme' (Margalit 1980, 219-28). [10]

The process of scansion is basically a heuristic one, a solution being proposed and then its counter-offers weighted. As Kaddari (1973, 168) noticed, complex semantic units might appear (compound words, "series", "expressions"), "essential semantic links" (single semantic units whose constituent parts do not have any exact parallels) may be present as well. A practical way to check the results is by identifying the theme of the semantic field as a whole. Units might stand in thematic relation although particular words might stand alone without explicit counterparts (Kaddari 1973, 170).

Nevertheless, Kaddari made use of the grammatical-syntactic structure of the utterance; it is examined in order to validate the parallelism traced down semantically. It is considered "a prerequisite for the proper establishment of a parallelism in regard to any word or phrase." (Kaddari, 1973, 172) By doing that, he actually discarded everything did not fit the synonymous and antithetic categories, those which were previously named synthetic parallels. After such steps are completed, the semantic critic identified not more than parallels sharing a common semantic field. Further finer description will disclose the relationship between semantic units belonging to a common field.[11]

The result might be double-checked by the Masoretic accents (te'amim), on which one would assign not only an intonational value, but, more importantly, a grammatical value. Intended by their authors, the Tiberian Masoretes, to facilitate the reading (by a proper intonation) and understanding of the Holy Scriptures (by connecting/disconnecting words and clauses), when accents happen to contradict the fluency of syntax, they might as well provide a stumbling block on the way to easing our text perception (Kugel 1981, 109-16). Thus, one would consider carefully the links and divisions suggested by Masoretes and proceed accordingly.



Alter, Robert. 1998. The Poetic and Wisdom Books. Pp. 226-240 in The Cambridge Companion to Biblical Interpretation. Edited by John Barton. Cambridge: University.

__________. 1990. The Art of Biblical Poetry. Edinburgh: T&T Clark. [Previously printed by Basic Books, New York, 1985].

Berlin, Adele. 1985. The Dynamics of Biblical Parallelism. Bloomington, Ind.: Indiana University.

___________. 1991. Biblical Poetry Through Medieval Jewish Eyes. ISBL. Bloomington and Indianopolis, Ind.: Indiana University.

Dawson, David Allan. 1994. Text-Linguistics and Biblical Hebrew. JSOTSupl, 177. Sheffield: Academic.

Fishoff, E. 1971. Piyyut. Columns 574-602 in vol. 13 of Encyclopaedia Judaica. 16 vols. Jerusalem: Keter.

Gaster, T. H. 1969. Myth, Legend and Cult in the Old Testament. New York, N.Y.: Harper & Row.

Geller, Stephen A. 1982. Theory and Method in the Study of Biblical Poetry. JQR 73:65-77.

Hrushovski, Benjamin. 1971. Prosody, Hebrew. Columns 1195-1240 in vol. 13 of Encyclopaedia Judaica. 16 vols. Jerusalem: Keter.

Kaddari, M. Z. 1973. A Semantic Approach to Biblical Parallelism. JJS 24:167-175.

Korpel, Marjo C. A. and Johannes C. de Moor. 1988. Fundamentals of Ugaritic and Hebrew Poetry. Pp. 1-61 in The Structural Analysis of Biblical and Canaanite Poetry. Edited by Willem van der Meer and Johannes C. de Moor. JSOTSup, 74. Sheffield: JSOT.

Kosmala, Hans. 1964. Form and Structure in Ancient Hebrew Poetry (A New Approach). VT 14:423-45.

Kugel, James L. 1981. The Idea of Biblical Poetry : Parallelism and Its History. New Haven, Conn.: Yale University.

____________. 1984. Some Thoughts on Future Research into Biblical Style: Addenda to "The Idea of Biblical Poetry." JSOT 28:107-17.

Longacre, Robert E. 1989. Joseph: A Story of Divine Providence. A Text Theoretical and Textlinguistic Analysis of Genesis 37 and 39-48, Winona Lake, Ind.: Eisenbrauns.

Margalit, Baruch. 1980. A Matter of Life and Death: A Study of the Baal-Moth Epic (CTA 4-5-6). AOAT 206. Kevelaer: Butzon and Bercker, Neukirchener-Vluyn: Neukirchener.

O'Connor, Michael Patrick. 1980. Hebrew Verse Structure. Winona Lake, Ind.: Eisenbrauns.

Schökel, Alonso. 1988. A Manual of Hebrew Poetics. SubBib, 11. Roma: PIB.

Weitzman, Steven. 1997. Song and Story in Biblical Narrative: The History of a Literary Convention in Ancient Israel. Bloomington and Indianopolis, Ind.: Indiana University.



[1] Buxtorf the Younger translated substantial passages from Azaria de'Rossi and incorporated them in his book Liber Cosri (Basel 1660) under the title Mantissa Dissertationum (pp. 415 ff). Robert Lowth translated them into English as part of the Preliminary Dissertation of his translation on Isaiah, Isaiah, A New Translation, with a Preliminary Dissertation and Notes, 2nd ed. London 1779, pp. xli-xlvii [Kosmala (1964, 425 ns. 1 and 2); cf also Berlin (1991, 142)].

[2]Before Lowth had the great illumination on the structure of HP verse, other Jewish authors are credited for laying down the premises toward its discovery. Kugel (1981, 156-9) implies here Theodore of Mopsuestia, who had proved to be the first one to discover the working of biblical parallelism, as he was commenting on Hosea 5:9, where he implied the category diaeresis (Lat for 'dividing in two'). Berlin

[3]Royalties are paid to Herman Gunkel, the first one to suggest an aesthetic function of POETRY incorporated into prose (1967, 2 n.c) and T. H. Gaster, who suggested pragmatic functions of the POETRY paragraphs, such as relieving tedium, securing audience participation and marking logical pauses in the narrative (Gaster 1969, 240-1).

[4] These are rather broad categories which are made of subservient types, in Eissfeldt's own terms, speeches, sermons, prayers, records of various natures, myths, fables, sagas, legends, reports, autobiographies, visions, dreams, sayings, proverbs, songs, poems to name just some of them (Eissfeldt 1965, 9-127).

[5]Using a matrix organized on the basis of three parametres instead of only two as Longacre did, Dawson refines the four types of texts into eight: NARRATIVE DISCOURSE has two types, prediction and story, PROCEDURAL DISCOURSE has two types, how-to-do-it, and how-it-was-done, BEHAVIOURAL DISCOURSE has also two types, exhortation and eulogy, and EXPOSITORY DISCOURSE has two types, budget proposal and scientific paper. Some can be traced back to Longacre, but it is so obvious that terms need to be defined more carefully because what story tends to assume as a catch all term is defining just one out of eight categories in textlinguistic terminology (Dawson 1994, 98).

[6] Alter agrees with Hrushovski (recently Harshav) both on terminology and the definition and role of parallelism (Alter 1998, 227-8). For a more in depth presentation of Alter's perception of HP see (Alter 1990).

[7] Although he avoids speaking of clauses per se, preferring more simple terms as A and B, in his later "addenda" Kugel (1984) regards A and B as clauses. By defining them as clauses Kugel loses the neutrality he was so eager to see prevailing in his work. Implying a subject and a predicate, this syntactic term is also not appropriate, because as Kugel himself proves, sometimes these types of 'clauses' are, interestingly enough, made of secondary units only.

[8] A clear history of parallelism might be outlined by a complementary reading of O'Connor (1980), Kugel (1981), and Berlin (1985). Its presence in Hebrew literature was never challenged except its role as the distinctive marker for POETRY as opposed to prose.

[9] Prosody is a term that assumes metre, whose presence particularly in Ugaritic literature has long been a serious bone of contention.

[10]Launched in 1975 (Studia Ugaritica I: Introduction to Ugaritic Prosody, Ugarit-Forschungen 7: 289- 313), the method was applied extensively to Ugaritic literature in A Matter if Life and Death: A Study of the Baal-Mot Epic (CTA 4-5-6) [AOAT 206; Kevelaer: Butzon and Bercker, Neukirchener-Vluyn: Neukirchener, 1980], and in The Ugaritic Poem of Aqht: Text, Translation and Commentary [Berlin, Walter de Gruyter, 1989].

[11]There are two main situations semantic units might found themselves: sometimes they are in a parallel relation although they regularly occur separately, outside the sphere of parallelism; other times, semantic units appearing regularly together even in non-parallel contexts, are split into the cola of a parallel verse. Semantic units belonging to the first category would develop three types of relations: synonymous, heteronomous, and partially antynomous. Semantic units belonging to the second category could create two types of relations: composites of linked phrases are divided into two co-ordinated parallel cola, and a series of co-ordinated types is split in parallelism (Kaddari 1973: 172-4). Despite of all these ramifications, the author is sober enough to accept that there are poetical units which do not display semantic parallelism, and non-poetical texts which can be constructed in the style of semantic parallelism as well.


© Studia theologica 2003-2010
ISSN 1583-5367