Twenty-seven years later, D. N. Freedman approached Mitchell Dahood and convinced him to begin work on the Psalms for the Anchor Bible.[2] As an outgrowth of that work, Dahood independently made the same observation as Möler, naming the pattern "double-duty modifier," and demonstrating its presence in both Hebrew and Ugaritic[3]. His three volume work on the Psalms precipitated a great deal of scholarly debate on the Ugaritic method. His small 1967 paper on the double-duty modifier also precipitated the tiny bit of literature that has since been published on the Pivot Pattern as such. Dahood first defined the double duty modifier as "a phrase, sometimes just a divine name or title in the vocative case, suspended between the first and third cola of a verse and simultaneously modifying both of them." He suggested the term "double duty modifier" because it would place this new observation in a category related to the double duty suffixes, prepositions, vocatives, interrogatives, etc. that had been recovered and which were receiving increased attention. He noted that the recovery of this device improves the syllable count in cases previously considered bicola and that the sense of the verse is affected. Dahood is also responsible for presenting the Pivot Pattern as a tricolon on the page, with the modifier printed on its own line between the other two lines.[4]

Perhaps unaware of Dahood's discovery, A. M. Habermann in 1972 identified the pattern as "incomplete parallel bicola" and gave the example of Psalm 114:7. As suggested by the title, the pattern was considered a bicola and written on two lines.[5] G. B. Gray elaborated upon this concept in his 1972 book The Forms of Hebrew Poetry . Also perhaps unaware of Möler or Dahood, he calls the construction "incomplete parallelism without compensation." He includes examples from Deuteronomy and Lamentations.[6]

Wilfred G. E. Watson took note of Dahood's discovery and pursued it. In a 1976 article, he attributed the original observation to Dahood and coined the term Pivot Pattern. He also noted that the construction is identical with Gray's "incomplete parallelism without compensation." Watson prefers the term Pivot Pattern to double-duty modifier and two-way middle because it describes the structural unit as a whole and does not focus on the central modifier. The central modifier, says Watson, may at times not even be essential to the meaning of the verse.[7] Like Dahood, Watson described the structure as a tricolon with the pivot printed in the middle as a separate line. It was Watson who demonstrated that this pattern was present in Akkadian, Sumerian, Arabic and Greek, as well as in Hebrew and Ugaritic.[8] Watson defined the Pivot Pattern as, "the central member of a tricolon, though metrically independent, (which) belongs semantically to the first and third cola alike."

Watson took up the 1958 method of R. Austerlitz in "Ob-Ugric Metrics: The Metrical Structure of Ostyak and Vogul Folk-Poetry." By identifying two or more lines that comprise a single metrical unit, the remaining lines can then be placed into some sort of classification. Watson then proceeded to classify a metrical pattern, the bicolon (though there are a few exceptions) and compiled a list of texts having this pattern. Those with a central pivot are then classified according to a 2X2 matrix. The pivotal term is identified as grammatically essential or nonessential and the outer cola are identified as being set in parallel or repetition. Thus, pivotal terms may be classified as PeP (two parallel outer cola with a grammatically essential pivot), PnP (two parallel outer cola with a grammatically nonessential pivot), ReR (two repetitive cola with an essential pivot) and RnR (two repetitive cola with a nonessential pivot). In 1976, Watson noted that the relationship of the Pivot Pattern to other devices, particularly chiasm and enjambment, was in need of exploration.[9]

In a 1976 article, Dahood continued to use the term double-duty modifier and noted that the construction had also been called the "two-way middle." He noted that it occurs in modern poetry as well and gave as examples Shakespeare's Sonnets and John Donne's Corona. He did not mention the term Pivot Pattern. This article provides three examples: Psalms 86:12 and 109:14, which Dahood uses elsewhere, and II Samuel 1:21, which appears only here.[10]

Pierre Auffret took up Watson's terminology in 1978. Two of his three examples of the Pivot Pattern are set in chiasm similar to what Möler had observed. In Jonah 2:10, for example, he sets the pivot in the middle of a five line chiasm with the resulting syllable count of 7:4:5:4:7.[11] Watson's term was also used in 1978 by R. Althann who contributed one example of the pattern in a text (Jer 4:11-12) that had previously been considered prose. His article traces the discovery of the Pivot Pattern to Dahood, whom he calls the "pioneer," and attributes most of the subsequent work to Watson.[12]

By 1981, Watson had recognized Möler's article as being the first to identify the pattern and modified his view of the pattern from that of a tricolon to that of a bicolon. As Möler's article before it, Watson's article in Welsh's book is a grammatical handbook in itself with respect to chiasm. Again, like Möler, Watson now sees one possible Pivot Pattern (a-b-p-b-a) as a species of chiasm. On page 119 of that article, he gives an example from Psalm 59:2 (which would be repeated in his 1984 textbook). He footnotes this example on page 161, note 13, by saying that this example follows Dahood's Psalms, but is careful to note that Dahood did not recognize this particular example.[13]

In line with Gray and Habermann, Adele Berlin wrote a 1992 article on parallelism noting that it may or may not have compensation. No mention is made of Dahood's double duty modifier or Watson's Pivot Pattern as being related to this pattern.[14]

Watson's subsequent textbooks on Hebrew Poetry in 1984 and 1994 have cemented the use of his term.[15] Watson modified his definition in 1984, stating that the Pivot Pattern is "a couplet where the expected final word is not expressed as it is implied by the last word (or words) of the first line." He also writes, "the pivot pattern can be defined as a couplet with final silent stress " (italics his). This makes the Pivot Pattern a species of ellipsis and seems to have been influenced by the Gray and Habermann definitions. Consistent with this understanding, and deviating from Dahood, Watson now prefers to print the construction as a bicolon with the pivotal word or phrase attached to the first line. Additionally, as implied by Auffret's example, Watson notes that the Pivot Pattern is related to chiasm, specifically the aba-monocolon and the ABA tricolon. It is also related to staircase parallelism. Watson then gives several examples to illustrate his definition, drawing on Psalm 57 as well as Akkadian and Ugaritic examples.[16]

Watson retains his classification of the Pivot Pattern according to four types: 1) the pivotal word as essential to the syntax with repetitive lines, 2) the pivot word as essential to the syntax with parallel lines, 3) the pivot word as nonessential to the syntax with repetitive lines, and 4) the pivot word as nonessential to the syntax with parallel lines. He does not, however, repeat his previous abbreviations (PeP, etc.). He offers CTA 3D iv 83-84 as an example of (2) above: "The coming of his sister, Baal does see, the approach so swift of his father's daughter," where the pivotal "does see" is essential to a construction of two parallel lines. As an example of (3) above, Watson offers Gilgamesh XII 89: "I cannot tell you, my friend, I cannot tell you," where the pivotal "my friend" is not essential to the construction which is a single line repeated.[17] Watson further notes that the pattern may occur in prose (e.g., Ruth 2:19) and may combine with other patterns, such as gender matching patterns and chiasm.[18]

Despite the fact that Watson had discovered Möler by 1976, perhaps while researching chiasm for Welsh's book, his 1984 bibliography shows Dahood as the earliest reference. This is corrected in his 1994 book, which gives Möler's observation some discussion along with credit for being the first to recognize the pattern.[19]

One or two additional examples may have been isolated by Alter in 1985. Working with Lamech's speech in Genesis 4:23-24, he notes that the beginning and ending of this poetic piece is marked by exact syntactic parallelism with the verbs ytrma and qy doing "double duty for both halves of the line." He states that this is a common pattern of ellipsis and that the double duty term is most often a verb.[20] His emphasis, however, is not the Pivot Pattern but syntactic parallelism.

Further examples were easily found by the members of the Hebrew Poetry Seminar held in Chicago in 1996 of which the first draft of this paper was a part. In addition to Joel 4:20 discussed above, I offered Joel 2:21 as another example. Hunn Choi, in reviewing this paper, identified five additional example from the Psalms. Ahida Cama-Calderon found an additional example in Psalm 61:6 while researching word pairs. The unpublished discoveries of the Seminar are included in an index appended to this paper.

Development and Function
Only Watson has attempted to discuss the development and function of the Pivot Pattern. Of the two, something more definitive can be said about development in those cases wherein parallel texts can be compared. Isaiah 37:31, for example, is believed to have been derived from the Phoenician Eshmunazor Inscription (11-12) which says: