Pe before Ayin in Biblical Pre-Exilic Acrostics

This article was adapted from a longer article in my book Esther Unmasked: Solving Eleven Mysteries of the Jewish Holidays and Liturgy (Kodesh Press, 2015), pp. 207-230. Earlier versions of this article were published in Biblical Archaeology Review, July-August 2012, pp. 47-50, andJournal for the Study of the Old Testament 38 (2014), pp. 471-85

The third chapter has three verses for each letter; the fifth chapter, although not an acrostic, consists of exactly twenty-two verses (i.e., the same number as letters of the Hebrew alphabet). The following is a listing of all the acrostics in the Bible: Lam. 1-4, Prov, 31:10-31,and Psalms 9-10, 25,34,37,111,112, 119 and 145. 

This acrostic is made up of three verses for each letter, making the divergence here even starker; I have cited only the initial verses of each triple acrostic above.

Even though these are two chapters in the modern counting, they are one psalm and part of a single acrostic.

The acrostic that spans chapters 9 and 10 is much less even with respect to the average unit length for each section than are the other acrostic psalms. This also suggests that the acrostic is textually corrupt. See Peter C. Craigie, Psalms 1-50 (Word Bible Commentary, vol. 19; Waco, TX: Word books, 1983), p. 129. The Dead Sea and Nachal Chever texts of limited portions of chapters 9 and 10 closely parallel the Masoretic text, and do not preserve an earlier, less corrupt version; see Discoveries in the Judaean Desert, XXIII (Oxford: 1998), pp. 54 and 66, and XXXVIII (Oxford: 2000), pp. 148-151. Nor is the LXX is of any help in restoring the original acrostic.

The suggestion that פיהו was the first word of the pe verse, and עיניו the first word of a subsequent ayin verse was made long ago. See Charles Augustus Briggs and Emilie Grace Briggs,A Critical and Exegetical Commentary on the Book of Psalms,vol. 1 (ICC; Edinburgh: 1906), p. 70. It is still widely accepted today; see, e.g., Demsky, p. 18, Hanan Eshel and John Strugnell, “Alphabetical Acrostics in Pre-Tannaitic Hebrew,”Catholic Biblical Quarterly 62 (2000), pp. 441-458 [443]; Ron E. Tappy, P. Kyle McCarter, Marilyn J. Lundberg, and Bruce Zuckerman, “An Abecedary of the Mid-Tenth B.C.E. from the Judaean Shephelah,” Bulletin of the American Schools of Oriental Research 344 (2006), pp. 5-46 (26).

  See Discoveries in the Judaean Desert, XVI (Oxford: 2001), p. 234. I have included above just the opening words of each verse, which is the same as in the MT. There are some differences in the verses themselves as well, but a discussion of this would take us too far afield.

These manuscripts are hundreds of years earlier than the earliest Hebrew texts of Proverbs. In the other main witness to the text of the Septuagint, Codex Alexandrinus (5^th century), the ayin verse precedes the pe verse, most likely, reflecting an attempt to make the LXX conform with the MT.

Editor’s note:  Aside from sorting and alphabetizing words, which as far as I know is not attested in antiquity, it is only for didactic reasons that alphabets need to have a particular order. MZB

See, for instance, discussion of the Ugaritic cuneiform abecedary (13^th century B.C.E.) found in 1933 at Beit Shemesh, in Wayne Horowitz, Takayoshi Oshima, and Seth L. Sanders,Cuneiform in Canaan and the Land of Israel: Cuneiform Sources from the Land of Israel in Ancient Times (Jerusalem: 2006), pp. 157-60; Seth L. Sanders, The Invention of Hebrew(Urbana: University of Illinois Press, 2009), 90-96.

I am calling this writing “Hebrew” because it was likely composed by an Israelite. But paleographers would not use the term “Hebrew” here, since it was probably not until the 9^thcentury B.C.E. that a “Hebrew” script evolved that was distinct from the neighboring Phoenician script, where such writing originated. The same issue of terminology arises with respect to the find at Tel Zayit.

According to most scholars, the first four lines on the Izbet Sartah potsherd seem to be just random letters (and seem not to have been inscribed by the same person who wrote the abecedary). The whole potsherd may simply reflect an exercise by two students. See Moshe Kochavi, “An Ostracon of the Period of the Judges from ‘Izbet Ṣarṭah ,” Tel Aviv 4 (1977), pp. 1-13; Aaron Demsky, “A Proto-Canaanite Abecedary Dating from the Period of the Judges and its Implications for the History of the Alphabet,” Tel Aviv 4 (1977), pp. 14-27; Demsky and Kochavi, “An Alphabet from the Days of the Judges,”Biblical Archaeology Review Sept.-Oct. 1978, pp. 23-30; and Kochavi and Demsky, “Alphabeit bi-Khetav Proto-Canaani me-Izbet-Tzartah,” Kadmoniot 11 (1978), pp. 61-67.

Editor’s note: Akkadian is always written from left to right.  It is  unknown why the right to left order, in which right-handed people tend to smudge the letters that they have just written, was established for Hebrew and related languages.  The suggestion that this was done by a lefty is an often suggested, but unlikely. MZB

Also noteworthy is that the ḥet precedes the zayin.

Kochavi, p. 12, and Demsky and Kochavi, p. 24. Three occupational strata were found at Izbet Sartah, but the potsherd was found in a silo that was unattached to any house, so it cannot be dated through stratigraphy. The earliest stratum probably dates to the late 13^th century B.C.E. and the latest probably to the 10^th cent B.C.E.

For example, the site contains the typical Israelite storage silos and pillar-courtyard style houses. See William G. Dever,Who Were the Early Israelites and Where Did They Come From? (Grand Rapids and Cambridge: 2003), p. 83. Kochavi, p. 3, writes that the identification of Izbet Sartah as an Israelite site “is not open to question.” See also Israel Finkelstein, The Archaeology of the Israelite Settlement (Jerusalem: 1988), pp. 31-33.

See Ze’ev Meshel, Kuntillet ‘Ajrud (Ḥorvat Teman): An Iron Age II Religious Site on the Judah-Sinai Border(Jerusalem: 2012), pp. 102-03, for the most accurate photo, sketch and transcriptions. Above these three abecedaries is another, largely incomplete, that does not include pe or ayin.

Meshel, pp. 66-69.

See Ron E. Tappy, et al. “An Abecedary,” 26. The Tel Zayit abecedary departs from the traditional order in three other instances: 1) vav precedes he, 2) ḥet precedes zayin as it does in the Izbet Ṣarteh abecedary, and 3) lamed precedes kaf. (As to the last, evidence indicates that the author of the abecedary realized this was a mistake.)  This all suggests that it was likely a student exercise.

Tappy, et al, pp. 6-7, 22, and 42. It should be noted that in the unlikely event that the Izbet Sartah, Kuntillet ‘Ajrud, and Tel Zayit abecedaries were inscribed by non-Israelites, these abecedaries would still be relevant. The Israelites utilized the same alphabet as their neighbors in Canaan, and the order of the letters within the alphabet would presumably have been the same.

See Martin Heide, “Impressions from a New Alphabetic Ostracon in the Context of (Un)Provenanced Inscriptions: Idiosyncracy of a Genius Forger or a Master Scribe?” pp. 148-182, in Meir Lubetski, ed., New Seals and Inscriptions, Hebrew, Idumean and Cuneiform (Sheffield: 2007). Although the ostracon is unprovenanced, the prevailing view is that it is genuine. Supposedly, it came from the debris of the Temple Mount.

Other Hebrew First Temple period abecedaries have been found, but they are much shorter. For example, אבגדה was found incised on one of the steps of the palace at Lachish. Other finds in ancient Israel include: אבגד, הוזחט (see the note below), וזח, and זחט.

There is an eighth century B.C.E. inscription from Lachish that Emil Puech believes follows the ayin/pe order. But André Lemaire rejects Puech’s readings as unsubstantiated. See David Ussishkin, The Renewed Archaeological Excavations at Lachish (1973-1994) (Tel Aviv: 2004), vol. IV, pp. 2116-2117. (This section of the work was written by Lemaire.) According to Lemaire, the letters הוזחט are readable on the middle of the first line, and the second line includes צקר, followed immediately by אב. But the letters preceding צקר on the second line are unclear.

Over time, scholarly awareness of the prevalence of the order is increasing. Ryan Byrne calls the pe-ayin the “prevalent (if not dominant)” order of the time. See Ryan Byrne, “The Refuge of Scribalism in Iron I Palestine,” Bulletin of the American Schools of Oriental Research 345 (2007), pp. 1-31 (5).

Byrne, p. 4.

See Christopher Rollston, “Non-Provenanced Epigraphs I: Pillaged Antiquities, Northwest Semitic Forgeries, and Protocols for Laboratory Tests,” Maarav 10 (2003), pp. 135-93 (160-62).

See Byrne, pp. 4-6.

Ancient Semitic languages are traditionally divided into three branches: Eastern, Western, and Southern. The Eastern Semitic languages (i.e., Akkadian, Babylonian, and Assyrian) were written in Sumerian cuneiform, which was not an alphabetic cuneiform script. Southern Semitic languages (i.e., South Arabian, Ethiopian, and Arabic) have a different alphabetic order altogether.

See Manfried Dietrich, Oswald Loretz, and Joaquín Sanmartín, eds., The Cuneiform Alphabetic Texts from Ugarit, Ras Ibn Hani and Other Places, (Münster: 1995, second ed.), pp. 490-495.

See Rainer Degen, “Ein aramäisches Alphabet von Tell Halaf,” in Rainer Degen, Walter W. Müller & Wolfgang Röllig,Neue Ephemeris für semitische Epigraphik, vol. III (Wiesbaden: 1978), pp. 1-9.

Nahum Sarna, “Psalms, Book of,” Encyclopaeida Judaica 13:1309: “the earliest collection is undoubtedly Book I, or rather Psalms 3-41 within it.”

צעקו – בשובם מרשעתם ויש אומרים כי זה הפסוק שב אל שועת הצדיקים, והנכון כאשר דברתי והעד קרוב

Also, we would have expected it to be stated earlier, before the acceptance of the outcry was described in verse 18.

See, Hermann Gunkel, Die Psalmen (fifth edition; Göttingen: Vandenhoeck & Ruprecht, 1968), 142. The earliest scholars that I  am aware of to make this suggestion are the German Bible scholar, Hermann Hupfeld, see discussion in Julius Wellhausen, The Book of the Psalms: Critical Edition of the Hebrew Text (Leipzig: 1895), 81, and the Jewish scholar, Solomon Judah Rapoport, Erekh Milin 2 (Warsaw: 1914), 42-43. See also, Samuel R. Driver, An Introduction to the Literature of the Old Testament (4th edition; Edinburgh: 1892), 346. The suggestion is even considered by the traditional scholar, Amos Hakham, in his Da’at Mikra commentary (ad loc.).

Although many scholars see this poem as dating to the Persian period, since certain descriptions sound as if they reflect Persian customs, the evidence for the original pe-ayinorder pushes in the direction pre or early exilic.

I am fairly sure that I am the first to make this suggestion about Ps. 25. It is possible that the suggestion has been made before about Ps. 37.

The Dead Sea texts include a text of the first two words of 25:15, but there is no text of the surrounding verses.

This is unlike the acrostics of chapters 9-10 and 37, in which there is (usually) an acrostic verse followed by a verse that begins with any letter.

Very few Hebrew words began with vav in the Biblical period. The concordance of Solomon Mandelkern (Lipsiae: 1896) lists only a handful: וו (hook), וזר (only at Mishlei 21:8), and ולד. This last parallel might suggest that these acrostics were composed at an early stage when the letter vav was not consistently viewed as deserving a place in a Hebrew acrostic. A possible support for this point comes from an 8^th or 7^th century B.C.E. Aramaic text from Deir ‘Alla (the Biblical Succoth, on the east bank of the Jordan River) that records the following sequence of letters: aleph, bet, gimmel, dalet, zayin, and ḥet. Both he and vav were left out. See J. Hoftijzer and G. van der Kooij, Aramaic Texts from Deir ‘Alla (Leiden: 1976), pp. 267 and 285.  Another piece of evidence for the antiquity of Psalm 25 is that it is textually corrupt: it is missing a kof, and has a resh twice. (Some characterize it as also missing a bet, but probably the second word of 25:2 [בך] counts as the bet, with the first word, אלקי, being considered only an introduction.)

There are no Dead Sea texts of the relevant verses in chapter 37. Neither are the relevant verses extant in the pesheron chapter 37, at 4Q171.

^{תהלים לז:כחa} כִּי יְ-הוָה אֹהֵב מִשְׁפָּט
וְלֹא יַעֲזֹב אֶת חֲסִידָיו

^{Ps 37:28a} For YHWH loves what is right,
He does not abandon His faithful ones.

^לז:כחb לְעוֹלָם נִשְׁמָרוּ
וְזֶרַע רְשָׁעִים נִכְרָת.

^37:28b They are preserved forever,
And the children of the wicked will be cut off.

The opening of v. 28b reads as if it connects to the previous clause, which is cumbersome and not the way parallelism generally works in biblical poetry.

The process is likely even more complicated here. The LXX actually has both texts, since the previous phrase “εἰς τὸν αἰῶνα φυλαχθήσονται” reflects the Hebrew, לעולם נשמרו. Editor’s note: One way to explain this is that this phrase was added in later, when a scribe familiar with the Hebrew text reflected in the MT noticed it was missing from the Greek. ZIF

Note how the missing word “lawless” parallels the word “wicked” in the next phrase, and also how the LXX’s verb ש.מ.ד ( “banish” or “destroy”) in place of the MT’s ש.מ.ר (“preserve” or “protect”) parallels the verb “cut off” in the next phrase. For a defense of this reading, see the discussion in Naphtali Herz Tur-Sinai, Peshutto shel Mikra, vol. 4a, ad loc.; also Meyer Gruber’s gloss (ad loc.) in the Olam HaTanach series. Also note the missing לעולם which became attached in the Greek to its reproduction of the Hebrew לעולם נשמרו. Perhaps in the original form of the verse, which only had the עולים נשמדו clause, the “forever” was part of this phrase and it read עולים לעולם נשמדו, “the lawless shall be banished forever.”

The Greek text here is from Codex Alexandrinus, which reads ανομοι δέ εκδιωχθήσονται; Codex Vaticanus and Codex Sinaiticus have the added phrase: αμωμοι εκδικηθησ, meaning “and the blameless shall be avenged,” which would not yield anayin verse, but something like תּמִימִים יִקּוֺמוּ.

Admittedly, chapter 37 is unlike chapters 25 and 34 in that it includes a verse that begins with vav and does not add an end verse beginning with pe. Also, each letter of the acrostic (for the most part) spans two verses. Nevertheless, a striking parallel between chapters 34 and 37 is that the samekh verse in both begins with sur me-ra va-aseh tov. Also, chapter 34 ends with כל החסים בו, while chapter 37 ends with כי חסו בו. Also, the superscription לדוד (without anything else) is found at chapters 25 through 28, and at chapters 34, 35 and 37, perhaps indicating some relationship between these chapters. (In the other books, it is found in MT only at chapters 103, 138, and 144.)