"Revisiting the Ammonite Ostraca," MAARAV 22 (2018): 45-77

MAARAV 22.1-2 (2018): 45–77 REVISITING THE AMMONITE OSTRACA Matthieu Richelle FACULTÉ LIBRE DE THÉOLOGIE ÉVANGÉLIQUE Few studies of the script of Ammonite ostraca have been published during the last three decades.1 E. Puech has devoted one important article to Ammonite palaeography, with some new readings and drawings of several ostraca, as well as palaeographic charts.2 G. van der Kooij’s doctoral dissertation on early West-Semitic scripts3 includes a short study of the script of four ostraca from Hisban4 (two of which are written in Aramaic script, as we will see). A subsequent article of his on the Transjordanian scripts only took into account one of them.5 The most detailed study on the topic to date is from F. M. Cross; in fact, 1 For the formal and semi-formal Ammonite script, see L. G. Herr, “The Formal Scripts of Iron Age Transjordan,” BASOR 238 (1981) 21–34 ; cf. also L. G. Herr, The Scripts of Ancient Northwest Semitic Seals (HSS 18; Missoula: Scholars, 1978). 2 E. Puech, “L’inscription de la statue d’Amman et la paléographie ammonite,” RB 92 (1985): 5–24. See also his remarks on Ammonite script in “Approches paléographiques de l’inscription sur plâtre de Deir ʿAllā,” in The Balaam Text from Deir ʿAlla Re-evaluated: Proceedings of the International Symposium held at Leiden 21–24 August 1989 (J. Hoftijzer and G. Van fer Kooij, eds.; Leiden/New York/København/Köln: Brill, 1991): 221–238. 3 G. van der Kooij, Early North-West Semitic Script Traditions: An Archaeological Study of the Linear Alphabetic Scripts up to c. 500 B.C.; Ink and Argillary (Ph.D. dissertation, Univ. of Leiden, 1986): 47–50, 299–300 and Fig. 4. 4 In current nomenclature, these are A1, A2, A4 and A5. 5 That is, A1: G. van der Kooij, “The Identity of Trans-Jordanian Alphabetic Writing in the Iron Age,” in Studies in the History and Archaeology of Jordan III (A. Hadidi, ed.; Amman: Dept. of Antiquities of Jordan, 1987): 107–121. 45 46 MAARAV 22.1–2 (2018) this remains the foundational work on Ammonite semi-formal cursive palaeography.6 However, several factors make it necessary to reassess this subject. First, some recently discovered items (mostly from Tell Jalul and Tell al-ʿUmeyri) should now be taken into account. Second, even regarding ostraca published several decades ago, several corrections have been proposed since the studies mentioned above.7 Third, digitized photographs enable us to prepare more precise palaeographic charts. Fourth, there has been hesitation with regard to the classiications of several inscriptions. Against this background, the aim of the present article is threefold: (1) Delineating the corpus of ink-ostraca which can really be regarded as written in a distinctive Ammonite script. (2) Revisiting the readings of some ostraca. (3) In light of (1) and (2), revisiting the palaeography of Ammonite ink-ostraca on a letter-by-letter basis.8 1. DELINEATING THE CORPUS Let us begin with an inventory of ostraca and inscribed potsherds discovered in the Ammonite territory, site-by-site.9 6 F. M. Cross, “Ammonite Ostraca from Tell Hisban” in Leaves from an Epigrapher’s Notebook: Collected Papers in Hebrew and West Semitic Palaeography and Epigraphy (HSS 51; Winona Lake: Eisenbrauns, 2003): 70–94. In this article, I am following the terminology “semi-formal cursive” for ostraca, and “formal cursive” for seals and bullae (see C. A. Rollston, Writing and Literacy in the World of Ancient Israel: Epigraphic Evidence from the Iron Age [Archaeology and Biblical Studies 11; Atlanta: SBL, 2010]: 96–97). 7 C. Rollston, “Heshbon A4 [=Heshbon 2],” BASOR 350 (2008): 87–89; M. Richelle and M. Weigl, “Hisban Ostracon A1: New Collation and New Readings,” ADAJ 53 (2009): 127–138; M. Richelle, “Notes épigraphiques sur l’ostracon numéro 3 de Tell el-Mazar,” in Monde sémitique et Bible hébraïque (Semitica 54; T. Römer and M. Langlois, eds.; Paris: Maisonneuve, 2012): 125–144. 8 I am much indebted to several colleagues for allowing me to use photographs for this article. For Hisban ostraca A1 and A3, Tell el-Mazar ostracon 3, and Tell ʿal-Umeyri ostracon 2 (CAI 144): photographs by B. Zuckerman and M. Lundberg, West Semitic Project; courtesy Department of Antiquities, Jordan. For A2: photograph of F. M. Cross, taken by E. Nitowski, communicated to me by W. E. Aufrecht. For Jalul 1 Ostracon, photograph by D. Sherwin, authorization by R. E. Gane and R. E. Younker (Andrews University). Also, I am very grateful to R. E. Gane for drawing my attention to the article he has co-written with C. J. Goulart on Jalul 2 Ostracon, as well as photographs of the two ostraca from Tell Jalul. 9 The reference corpus of Ammonite inscriptions, exhaustive at its date of publication, is W. E. Aufrecht, A Corpus of Ammonite Inscriptions (Ancient Near Eastern Texts and RICHELLE: REVISITING THE AMMONITE OSTRACA 47 • Tell Hisban. The excavations have unearthed nine ostraca and inscribed sherds, published by Cross in various articles10 and in an excellent synthesis.11 They are designated as A1 to A9. Among these texts, six are written in ink (A1 to A6, dated by Cross from the sixth century b.c.e.) and three are incised (A7 and perhaps A8 from the seventh century b.c.e., A9 from the ifth century b.c.e.). • Tell el-Mazar. Three inscribed potsherds and four ostraca have been published by K. Yassine and J. Teixidor.12 In the following, they will be referred to as Mazar 1 to 7, after their order of publication in the editio princeps. According to the latter, Mazar 1, 2 and 6 probably date from the Hellenistic period, while Mazar 3, 4 and 5 are ascribed to the irst half of the sixth century b.c.e. and Mazar 7 to the ifth century b.c.e. • Tell el-ʿUmeyri. Four items have been discovered (which I will designate as ʿUmeyri 1 to 4). In 1992, Larry G. Herr published an inscribed potsherd which comprises only two assured letters.13 In 1997, T. K. Sanders published an ink ostracon comprising ive lines, with only a few legible letters; Studies 4; Lewiston: Mellen, 1989), completed by a subsequent article: “Ammonite Texts and Language,” in Ancient Ammon (Studies in the History of the Ancient Near East 17; B. Macdonald and R. E. Younker, eds.; Leiden: Brill, 1999): 163–188. I am very grateful to Prof. Aufrecht for his help in writing this article, in particular for sending me a version of the (forthcoming) second edition of his Corpus. 10 F. M. Cross, “An Ostracon from Heshbon,” AUSS 7 (1969): 223–229; idem, “Heshbon Ostracon II,” AUSS 11 (1973): 126–131; idem, “Ammonite Ostraca from Heshbon: Heshbon Ostraca IV–VIII,” in Heshbon 1973 (R. S. Boraas et al., eds.; Berrien Springs: Andrews Univ., 1975): 1–20; idem, “Ostracon from Heshbon,” AUSS 7 (1969): 223–229; idem, “Heshbon Ostracon XI,” AUSS 14 (1976): 145–148; idem, “An Unpublished Ostracon from Hesban,” in The Archaeology of Jordan and Other Studies (S. H. Horn, ed.; Berrien Springs: Andrews Univ., 1986): 474–489. See also idem, “The Ammonite Ostraca from Tell Hesban,” in Hesban: After 25 Years (D. Merling, ed.; Berrien Springs: Andrews Univ., 1994): 169–174. 11 Cross, “Ammonite Ostraca from Tell Hisban” (n 6); idem, “Ammonite Ostraca from Tell Hesban,” in Hesban 12: Small Finds: Studies of the Bone, Iron, Glass, Figurines, and Stone Objects from Tell Hesban and Vicinity (P. J. Ray, ed.; Berrien Springs: Andrews Univ., 2009): 29–56. 12 K. Yassine and J. Teixidor, “Ammonite and Aramaic Inscriptions from Tell el-Mazār in Jordan,” BASOR 264 (1986): 45–50. 13 L. G. Herr, “Epigraphic Finds from Tell el-ʿUmeiri during the 1989 Season,” AUSS 30 (1992): 187–200, esp. 195–196, 200; idem, “Epigraphic Finds from Tell el-ʿUmeiri during the 1989 Season,” in Madaba Plains Project 3: The 1989 Season at Tell el-ʿUmeiri and Vicinity and Subsequent Studies (L. G. Herr et al., eds.; Berrien Springs: Andrews Univ., 1997): 323–330. 48 MAARAV 22.1–2 (2018) he dated it to the mid-sixth century.14 In 2000, Herr published another engraved inscription from the 1992 excavations season.15 Lastly, C. A. Rollston has published another ink ostracon from the late seventh to sixth century b.c.e. where one can read the element šlm.16 • Tell Jalul. Two ostraca have so far been unearthed: an ink ostracon with six well-preserved lines (Jalul 1), published in 2008 by R. E. Gane and ascribed by him to the sixth century b.c.e.,17 and a fragment of another ink ostracon18 (which I will designate as Jalul 2), recently published by C. G. Goulart and R. E. Gane, dated by them to the seventh–sixth centuries b.c.e.19 • Amman Citadel. An incised potsherd (CAI 77) was discovered in 1972.20 Seven very short incisions were published by R. H. Dornemann in 1983.21 Among several unpublished short inscriptions, a two-letter word written in ink on a jar has been discovered during excavations by the Ecole Biblique et Archéologique Française de Jérusalem at the Amman Citadel. • Khirbet Umm ad-Dananir. A very short incised ostracon was found in 1987.22 14 T. K. Sanders, “An Ammonite Ostracon from Tall al-ʿUmeyri,” in Madaba Plains Project 3 (n 13): 331–336. This ostracon was already mentioned by L. G. Herr, “What Ever Happened to the Ammonites,” BAR 19 (1993): 26–35, esp. 32. 15 L. G. Herr, “The Inscriptions,” in Madaba Plans Project 4: The 1992 Season at Tall al-ʿUmeyri and Subsequent Studies (L. G. Herr et al., eds.; Berrien Springs: Andrews Univ., 2000): 248–251, esp. 250. 16 In L. G. Herr and D. R. Clark, “Madaba Plains Project: Excavations at Tall alʿUmeyri, 2008,” ADAJ 54 (2010): 51–72, esp. 70–71. Since the submission of this article, another ostracon has been published at ʿUmeyri and it should be read lpnn[ according to L. G. Herr, “The Objects,” in Madaba Plains Project 6: The 1996 and 1998 Seasons at Tall al-‘Umayri and Subsequent Studies (L. G. Herr et al., ed.; Berrien Springs: Andrews Univ., 2014): 409–410. 17 R. E. Gane, “Jalul Ostracon I,” BASOR 351 (2008): 73–84. 18 The ink is gone but the shape of the letters is nevertheless apparent. C. G. Goulart and R. E. Gane, “Three Epigraphic Finds from Tall Jalul, Jordan,” BASOR 365 (2012): 27–32. 20 F. Zayadine, “Recent Excavations on the Citadel of Amman,” ADAJ 18 (1973): 17– 35, esp. 31–32. 21 R. H. Dornemann, The Archaeology of Jordan in the Bronze and Iron Ages (Milwaukee: Milwaukee Public Museum, 1983): 103, 261; ig. 68. 19 22 P. E. McGovern, “Beqʿah Valley,” Liber annuus studii biblici franciscani 37 (1987): 385–388; idem, “The Baqʿah Valley Project 1987: Khirbet Umm ad-Dananir and el-Qeṣir,” ADAJ 33 (1989): 123–136, esp. 125. RICHELLE: REVISITING THE AMMONITE OSTRACA 49 • Sahab. An engraved potsherd with only one sign (perhaps taw) has been discovered.23 • Tell es-Sahadiyeh. An incised, very short ostracon was published in 1988.24 • Tell Abu Kharaz. An incised potsherd (with three letters) has been discovered (notice that the reading proposed in the initial publication has to be corrected to rqy[).25 • Tell Jawa. An ostracon, dated ca. 600 b.c.e. by P.-E. Dion, was published in 2002;26 it is badly preserved and apparently written in the Aramaic language. Regrettably, most of these discoveries only consist of very short engraved inscriptions. The outcome is that the only (published) ink ostraca are: Hisban A1 to A6; Mazar 3, 4, 5, 7; ʿUmeyri 2 and 4; Jalul 1 and 2; and the Jawa ostracon. Yet establishing such a list is not the end of the matter. As is well known, the Ammonite script derives from the Aramaic one,27 but local scribes developed idiosyncratic features which clearly appear, for instance, in the inscription engraved on the Tell Siran bottle. Although J. Naveh regarded the script used by the Ammonites as an 23 M. M. Ibrahim, “Third Season of Excavation at Sahab,” ADAJ 20 (1975): 73. J. N. Tubb, “Tell es-Saʿidiyeh: Preliminary Report on the First Three Seasons of Renewed Excavations,” Levant 20 (1988): 23–73, esp. 31, 33. 25 P. M. Fischer et al., “Tell Abu Kharaz: A Bead in the Jordan Valley,” NEA 71.4 (2008): 144–145; M. Richelle, Le royaume d’Israël dans la première moitié du VIIIe siècle avant notre ère: Analyse critique des sources épigraphiques, bibliques et archéologiques (Ph.D. dissertation, École Pratique des Hautes Études, 2010): 247–249. According to G. Galil (“A New Look at the History of Jabesh-Gilead [Tell Abu al-Kharaz] in the Light of New Archaeological and Epigraphic Data,” UF 46 [2015]: 106), one should read rqy[n], to be identiied with biblical Rezîn, the last king of Damascus. This is technically possible, but of course impossible to prove in view of the fragmentary state of the inscription, and one wonders why the name of this king would be inscribed here. 26 P.-E. Dion, “The Ostracon from Building 800,” in Excavations at Tall Jawa, Jordan, vol. 2: The Iron Age Artefacts (P. M. M. Daviau, ed.; Leiden: Brill, 2002): 268–275. 24 27 A. Millard, “Geschichte der Alphabetschrift,” in Sprachen aus der Welt des Alten Testaments (H. Gzella, ed.; Darmstadt: Wissenschaftliche Buchgesellschaft, 2012): 23. 50 MAARAV 22.1–2 (2018) Aramaic one,28 Cross preferred to speak of a distinct “national” script.29 As correctly noted by Rollston, the evidence is “slight,” but “it supports Cross’s position,”30 although of course a precise deinition of what should be labelled “national script” still eludes us.31 But the Ammonite writing tradition probably remained in close contact with Aramaic script and was inluenced by it,32 which renders the task of classifying inscriptions somewhat complicated. For the present discussion, this has three implications. First, some scholars would add to the list above one or two other items. Indeed, Puech believes that the Nimrud ostracon is written in Ammonite script.33 However, while it is acknowledged that the personal names listed in this text are (“ethnically” speaking) Ammonite, the script is generally regarded as Aramaic.34 Likewise, although it is not an ostracon, it is necessary to mention the Deir ʿAlla plaster inscription because it is a large ink inscription and, if Cross35 and Puech36 are correct in regarding it as written in the Ammonite script, it would be important to include it in the palaeographic charts and in the discussions concerning the development of the Ammonite script. This is, of course, a dificult subject which cannot be resolved here. In any case, the script of this text stands in so close 28 J. Naveh, “The Scripts in Palestine and Transjordan in the Iron Age,” in Near Eastern Archaeology in the Twentieth Century: Essays in Honor of Nelson Glueck (J. A. Sanders, ed.; New York: Doubleday, 1970): 277–283, esp. 280; reproduced in J. Naveh, Studies in West-Semitic Epigraphy (Jerusalem: Magnes, 2009): 3–9, esp. 6; idem, Early History of the Alphabet: An Introduction to West Semitic Epigraphy and Palaeography (Jerusalem, Magnes, 1987): 109. 29 Cross, “Ammonite Ostraca from Tell Hisban” (n 6): 76; Millard (n 27): 23. 30 C. A Rollston, “Northwest Semitic Cursive Scripts of Iron II,” in “An Eye for Form”: Epigraphic Essays in Honor of Frank Moore Cross (J. A. Hackett, and W. E. Aufrecht, eds.; Winona Lake : Eisenbrauns, 2014): 207 n. 10. 31 For a recent discussion of the slightly diverging views developed by Naveh and Cross, see C. W. Tyson, The Ammonites: Elites, Empires, and Sociopolitical Change (1000–500 BCE) (Library of Hebrew Bible/Old Testament Studies 585; London: Bloomsbury, 2014): 100–104. 32 G. van der Kooij, “Book and Script at Deir ʿAllā,” in The Balaam Text (n 2): 239–262, esp. 250–255. 33 E. Puech, “L’inscription” (n 2): 13. 34 See P. Bordreuil, “Les noms propres transjordaniens de l’ostracon de Nimroud,” Revue d’Histoire et de Philosophie Relligieuses 59 (1979): 313–317; J. Naveh, “The Ostracon from Nimrud: An Ammonite Name-List,” Maarav 2 (1980): 163–171, reproduced in Naveh, Studies (n 28): 313–321; 77 n. 55; S. Aḥituv, Echoes from the Past: Hebrew and Cognate Inscriptions from the Biblical World (Jerusalem: Carta, 2008): 384. 35 Cross, “Ammonite Ostraca from Tell Hisban” (n 6): 70. 36 Puech, “L’inscription” (n 2): 12; “Approches” (n 2): 221–238. RICHELLE: REVISITING THE AMMONITE OSTRACA 51 a relation to the Aramaic script tradition37 that it is perhaps more prudent in the present study not to draw conclusions from it concerning what we should regard as features of a distinctive Ammonite palaeography. Second, and conversely, H. Hübner thinks that the Hisban ostraca should be ascribed to the Moabite corpus, and among the seven inscriptions from Tell el-Mazar, he includes only Mazar 7 in the Ammonite ostraca.38 This view, however, has remained isolated in recent scholarship. Furthermore, the presence of the relative pronoun šin in A1 and Mazar 3 (see below) corroborates their Ammonite classiication. Third, some of the ostraca discovered in the Ammonite archaeological sites listed above are most certainly written in the Aramaic script. Indeed, Cross noted that the script of A4 to A6 is virtually indistinguishable from the Aramaic script.39 Puech reached the same conclusion for A4 and A5 and mentioned “aramaizing tendencies” in A3.40 It is all the more signiicant given that A4 (line 4: bny, plural construct), A5 (lines 1, 3, 4) and A6 (lines 1, 3, perhaps 4) use bn (not br) to mean “son of.” Similarly, Mazar 6 is written in Aramaic script from the ifth century b.c.e. In fact, if we are to follow Cross’s analysis, the distinctive Ammonite semi-formal cursive script has been abandoned in favor of Aramaic script in the second half of the ifth century. In the same way, P. Bordreuil considers that c. 500 b.c.e. the Ammonite script has been deinitively “aramaized.”41 While I think that some progress can be made in the detailed description of the development of some letters in the Ammonite writing tradition, there is no reason to challenge the consensus concerning the main lines of chronological evolution. It would be interesting to compare it to what we know of the political changes in the Ammonite territory around the sixth century b.c.e.,42 but that is another story. For the present purpose, let us notice that the same phenomenon probably occurs in Jalul 2, which 37 J. Naveh, “The Date of the Deir ‘Alla Inscription in Aramaic Script,” IEJ 17 (1967): 256–257, reproduced in J. Naveh, Studies (n 28): 206–208; Aḥituv, Echoes (n 34): 434; Millard, “Geschichte” (n 27): 23 (“mit lüssiger aramaïscher Hand”). 38 U. Hübner, “Die ersten moabitischen Ostraca,” ZDPV 104 (1988): 68–73; idem, Die Ammoniter: Untersuchungen zur Geschichte, Kultur und Religion eines transjordanischen Volkes am 1. Jahrtausend v. Chr. (Abhandlungen des Deutschen Palastina-Vereins 16; Wiesbaden: Harrassowitz, 1992): 31–33. 39 Cross, “Ammonite Ostraca from Tell Hisban” (n 6): 54. 40 Puech, “Approches” (n 2): 231. 41 P. Bordreuil, “Perspectives nouvelles de l’épigraphie sigillaire ammonite et Moabite,” in Studies in the History and Archaeology of Jordan III (n 5): 284. 42 O. Lipschits, “Ammon in Transition from Vassal Kingdom to Babylonian Province,” BASOR 335 (2004): 37–52. 52 MAARAV 22.1–2 (2018) uses bn but is written with Aramaic letters.43 Incidentally, it seems possible to suggest a dating of this ostracon that is slightly different from that of the editio princeps (seventh–sixth centuries b.c.e.). Since the best parallel to bet (with a totally opened head) is found in A6 (line 2), and since the scribes moved from Ammonite to Aramaic script at that period, I would tentatively date Jalul 2 to the end of the sixth century or possibly later (there are not enough letters to be more precise). Similarly, as Herr has noticed, two identical seal impressions from Tall al-ʿUmeyri containing the word ʿmn are written in Aramaic script.44 Finally, Rollston has noted that it is very dificult to decide whether the script of ʿUmeyri 3 is Ammonite or Aramaic.45 In sum, while the corpus of ink ostraca which can be regarded as written by Ammonite scribes is larger, the published ink ostraca which seem to exhibit distinctive Ammonite script features are the following: A1, A2 and A3; Mazar 3, 4 and 5; ʿUmeyri 2 and Jalul 1. 2. REVISITING THE READINGS While the Amman Citadel inscription and the Tell Siran bottle have been the subject of many studies challenging the interpretations proposed in their editio princeps, it is worth noting that the initial publication of the Ammonite ostraca have generally been followed without signiicant change in subsequent scholarly works. In this section I would like to point out some diverging readings for three ostraca: A1, Mazar 3 and Jalul 1.46 2.1. Ostracon A1 from Heshbon (see Pl. II) Cross published this ostracon in 1975.47 He was followed without signiicant modiication by most scholars.48 In 1985, Puech proposed a few 43 Goulart and Gane, “Three Epigraphic Finds” (n 19): 31. Herr, “Epigraphic Finds from Tell el-ʻUmeiri during the 1989 Season,” in Madaba Plains Project 3 (n 13): 327. 45 C. Rollston in Herr and Clark, “Madaba Plains Project” (n 16): 70. 46 In the case of A1 and Mazar 3, I will focus on the readings and summarize what I have proposed in two articles (Richelle and Weigl, “Hisban” [n 7] and Richelle, “Notes épigraphiques” [n 7]), to which I refer the reader for detailed explanations of the translation and philological aspects. 47 Cross, “Ammonite Ostraca from Heshbon: Heshbon Ostraca IV–VIII” (n 10). 48 E.g., Aufrecht, Corpus (n 9): 214–215; idem, “Ḥesbân Ostracon A1,” in The Context of Scripture, vol. 3: Archival Documents from the Biblical World (W. W. Hallo and K. 44 RICHELLE: REVISITING THE AMMONITE OSTRACA 53 corrections,49 but they have not been followed by Cross in his recent new re-publication.50 In 2009 M. Weigl and the present author proposed a new collation, based on direct examination of the text at the Amman Citadel Museum and on the use of new photographs.51 Text and translation: 1) [l]mlk. ʾkl 20+8[ To mlk, grain: 28 2) wṣʾn 9 and small cattle: 9 3) wlndbʾl bn nqmʾl k ̇[sp and to ndbʾl son of nqmʾl, sil[ver 4) lzʾ-[b]ṅ ʾlt ̇mk bt 10+2 ʾk ̇[l To zʾ-[so]n of ʾltmk, bath: 12, gr[ain 5) lẏʾ̇[ ]k ̇ṗʾ̇t.wrḥbt 2 w[ To ẏʾ ̇[ 6) lbʿš[ʾ ]ksp 20+20 wš ntn-[ To bʿš[ʾ], silver: 40 and what he gave [ 7) yn 20+2 wṣʾn 10 wbbt[ wine: 22 and small cattle: 10 and merchandise/sheep [ 8) yn 8 wʾkl 6 wine: 8 and grain: 6 9) lytb dšʾ ʾkl 20+4 To ytb hay, grain: 24 10) wyn 9- and wine: 9- 11) wrḥbt 3 and jars: 3 ] and jars: 2 and [ Line 3. Following Cross, most scholars read the patronym nʿmʾl. However, van der Kooij noted that we should probably read qop instead of ʿayin.52 In my view, this is even deinite: the qop here is formed by Lawson Younger, Jr., eds.; Leiden: Brill, 2003): 202 (CoS 3.84); K. Jackson, The Ammonite Language of the Iron Age (HSM 27; Chico: Scholars, 1983): 51–52; Aḥituv, Echoes (n 34): 371–372 (he only changed a numerical sign, reading 50 instead of 40 at line 6). See also lately Tyson, The Ammonites (n 31): 88, who does not seem to be aware of the republication of the ostracon by the present author and Weigl (see below). 49 Puech, “L’inscription” (n 2): 13–14; drawing in 16, ig. V. 50 Cross, “Ammonite Ostraca from Tell Hisban” (n 6). 51 Richelle and Weigl, “Hisban Ostracon A1” (n 7). 52 Van der Kooij, Early North-West Semitic Script Traditions (n 3): 48. 54 MAARAV 22.1–2 (2018) two vertical and symmetrical curved strokes, giving an ellipsoidal shape to the letter (ig. 1), and clearly differing from the occurrence of ʿayin in line 6. Hence the personal name nqmʾl. Figure 1: qop Line 4. Compare: • Cross: lz[ ]m ʾlt nkʾt 2 10+ʾk[ • Puech: lzʾb bn(?)ʾlt.nkʾt w/zrḥ ḃ[t • Richelle/Weigl: lzʾ-[b]ṅ ʾlṫmk bt 10+2 ʾk̇[l At the beginning of the line, the reading is very uncertain because the surface is abraded by scratches and the ink has faded. The most important correction concerns the middle of the line (ig. 2): there is no separator after the sequence ʾlt, and what has been taken for a nun is in reality a mem. Indeed, the top of the letter is too wide for a nun but its a mem; besides, one can even see a little stroke above the top, as in the shape of the mem in the preceding line. Hence the personal name ʾltmk. Figure 2: Line 4 As for the end of the line, Cross’s reading is the most probable. Puech was probably misled by the presence of an incrustation under the penultimate letter that gives it the appearance of a ḥet on some photos, and his reading seems to be inluenced by the next line. RICHELLE: REVISITING THE AMMONITE OSTRACA 55 Line 5. Compare: nkʾt 2 ʾrḥ bt 2 w[ • Cross: l[ ] • Puech: lyʾ [š.] nkʾt w/zrḥ bt 2 w[ • Richelle/Weigl: lẏʾ̇ [ ]k̇ṗʾt.wrḥbt 2 w[ At the beginning of the line, Puech’s additional readings seem correct. Then, where he follows Cross in reading nkʾt, we should rather probably read k̇ṗʾt (ig. 3) because the top of the irst letter is too wide for a nun but its a kap; moreover, this top enlarges to the left. The second letter does not exhibit the U-shaped top of a nun, but it perfectly its a reading pe. Figure 3: Line 5 detail In the next sequence of letters (ig. 4), Cross read the numerical sign two followed by an ʿayin, but Puech has rightly seen only one letter, hesitating between waw and zayin because of the close resemblance between the shapes of these letters. However, since his article only concerned palaeography, he proposed no new translation, so he did not draw all the conclusions from his correct epigraphic correction. The latter implies that we cannot read ʾrḥ bt 2, translated “two-year cow” by Cross, any more. Moreover, there is no separation between ḥet and bet. Rather, it is possible to read the word rḥbt (“amphora, jar,” after the Ugaritic)53 preceded by the preposition waw. The same situation occurs in line 11. Figure 4: Line 5 detail 53 G. del Olmo Lete and J. Sanmartin, A Dictionary of the Ugaritic Language in Its Alphabetic Tradition, Part 2 (HdO 67.1; Boston: Brill, 2003): 737. 56 MAARAV 22.1–2 (2018) Line 6. Here a direct examination conirmed beyond any doubt Puech’s reading of the relative pronoun in the form š, not ʾš.54 The letter taken for a ʾalep by Cross is in fact a waw (ig. 5). The same form occurs on an Ammonite seal (WSS 876) and (see below) in Mazar 3. Figure 5: waw Line 7. At the end of the line, Cross read lbbt but Puech hesitated between a lamed and a waw. As a matter of fact, the correct reading probably is waw (ig. 6): there is no downstroke, and the curving of the bottom left angle of the lamed is apparent. As a result, instead of the word lbbt (“wheat germ”), we can understand wbbt as the conjunction followed by a substantive, perhaps bbt, a sort of merchandise, after the Akkadian babtu.55 One could perhaps also think of Akkadian bibbu, “a kind of wild sheep.” Figure 6: Line 7 detail 2.2 Tell el-Mazar Ostracon 3 (see Pls. III–IV) Here again, the editio princeps (by Yassine and Teixidor) has been followed without signiicant change in subsequent studies. However, it 54 Note that some grammatical studies of the Ammonite dialect have (understandably) followed this erroneous reading: W. R. Garr, Dialect Geography of Syria-Palestine, 1000– 586 B.C.E. (Philadelphia: Univ., 1985): 85; I.-S. A. Yun, “The Transjordanian Languages during the Iron Age II,” UF 37 (2005): 741–766, esp. 751. 55 “Handelsgut” according to W. von Soden, Akkadisches Handwörterbuch, vol. 1 (Wiesbaden: Harrassowitz, 1965): 95; perhaps more precisely “an amount of staples, inished goods or merchandise outstanding (i.e., not at hand at the time of accounting but whose delivery or payment is expected with certainty in the near future” (CAD 2:10–13). RICHELLE: REVISITING THE AMMONITE OSTRACA 57 is possible to propose no less than seven corrections, as well as the decipherment of a few new letters.56 Recto: 1) ʾmr.plṭ. ʾmr.lʾḥh.lʿbdʾ[l.ḣ57] Message from plṭ58: “say to his brother, to ʿbdʾ[l]: 2) šlm ʾ-t.wʿt.šḋḃrt ʾtk[ ] 3) k ̇[ṙ]ṫẇ.lṁšʿrt.lšbṣ. nʿr[t] 4) ẇʿṫ ̇.ṫn lplṭ.ḋẏ[ḣ] 5) l yšb.b ʾ. “Are you well? And now, (regarding) what I discussed with you [ ], they have c[u]t wool (?) to weave, the young wo[men]59. And now, give plṭ what [he] needs [ ] he will give back.60 Go . . .”61 Verso: 1) ]kl[ ]all(?)[ 5) ] ʾḥh[ ]his brother[ Line 2. According to Yassine and Teixidor, the third word of the line is šʿ{r}rt. However, the letter read ʿayin is in fact a dalet, because it is open below and has a small tail (ig. 7). Then the editors believed the scribe to have written reš twice; in reality, despite a dark stain, we can read irst a bet (notice the opened top and the downstroke oblique towards the left, clearly different from the following reš). Hence ṧḋḃrt (which in this context can be translated: “what I discussed with you”). 56 Richelle, “Notes” (n 7). The restitution of a he at the end of l. 1 remains uncertain. Although one does ind the formula hslm.ʾt on a Kuntillet ʿAjrud inscription and on a Horvat ʿUzzah ostracon (see for example Aḥituv, Echoes [n 29]: 320 and 351 respectively), it also appears without the interrogative H- (SLM.ʾT) in Phoenician epigraphy (KAI 6, l. 50), as has been pointed out by D. Schwiderski, Handbuch des nordwestsemitischen Briefformulars: Ein Beitrag zur Echtheitsfrage der aramäischen Briefe des Esrabuches (BZAW 295; Berlin: de Gruyter, 2000): 42–43. 58 Or: “plṭ said.” 59 Or: “tow.” 60 Perhaps with the sense “he will pay” (cf. P. Bordreuil and F. Briquel-Chatonnet, “Aramaic Documents from Til Barsib,” Abr-Nahrain 34 [1996–1997]: 101). 57 61 Or: “So and so came,” the subject being named on the verso. 58 MAARAV 22.1–2 (2018) Figure 7: Line 2 detail The last letter of the line is not a nun but a kap, because its head is formed of a thick and uniform stroke, very slightly curved and rising very gently to the left (the whole letter remains close to horizontal); its lower left part is angular (ig. 8). The contrast with the nun in line 4 is telling: the head, not as thick, has a stroke which rises markedly to the left, like a right angle, and its lower left part is rounded. Figure 8: kap Line 3. This line begins with approximately four letters previously unread (ig. 9), perhaps k̇-ṫẇ (the second letter is erased). Of the irst letter, there remains a shaft as well as a thick head which could belong to a kap (a waw seems to be excluded because of the lack of curve to the head). The third letter could be a taw (one can see the greater part of the shaft and at least the beginning of the down stroke to the right). The fourth letter is probably a waw, with the head fairly steeply curved, rather than a kap which one would expect to have a shaft leaning towards the left. RICHELLE: REVISITING THE AMMONITE OSTRACA 59 Figure 9: Line 3 beginning Then Yassine and Teixidor have read lk.šʿrt. Yet instead of a kap followed by a word divider, we should read a mem with a crenellated top (ig. 10); what has been taken for a separator is in fact the left part of the top, clearly linked to the rest of the letter by an ink stroke starting from its lower part. Admittedly, the orientation of the shaft of the other mems of this ostracon is different (descending to the left). But precisely, at the lower part of the shaft of the letter under consideration, the scribe has deviated towards the left, as if to correct himself. (The lower “deviant” segment is probably not a scratch on the surface, but rather a stroke of ink, since is it of the same color and the same thickness of stroke as in the rest of the shaft.) Hence, after the preposition lamed, the word mšʿrt which could be related to Ugaritic šʿrt, a kind of wool.62 Figure 10: mem According to the editors, the next syntagm is lšbt. But instead of a taw we have in reality a ṣade (ig. 11), because the small stroke on the right consists of an angle, the scribe drawing irst of all an ascending then a descending stroke. The root šbṣ proves to be rare in Semitic languages; in Biblical Hebrew, it means “to weave.”63 An allusion to textile activity 62 Del Olmo Lete and Sanmartin, Dictionary (n 48): 799. 63 HALOT, 1401–1402. 60 MAARAV 22.1–2 (2018) here would make sense, because the ostracon has been discovered in the Palace Fort of Tell el-Mazar, which also contained a “weaving room” from the same stratum (III).64 Figure 11: ṣade Finally, the last word is not kʿr[bn] as proposed by Yassine and Teixidor, because the irst letter of this sequence is probably a nun (ig. 12). We can note the presence at the top to the right of the letter of a small thin stroke, which seems to even cross the main stroke. The editors probably took this stroke to be the prolongation of the shaft, of which a thick stroke ascending on the left separates itself, whence the reading kap. But the comparison with the other occurrences of kap indicates too great a contrast in the head: instead of a long thick line, nearly horizontal, we have a much thinner stroke, at a clear angle to the vertical. Conversely, this corresponds well to a nun; one only has to admit that the scribe corrected himself in the stroke and has left a superluous line at the top on the right. nʿr[t], “young ladies,” is perhaps the best reconstruction. While the syntactical position seems unusual, it might be the subject of a verb k-tw (at the beginning of the line), for instance k̇[ṙ]ṫẇ (“they have cut”). Figure 12: nun 64 K. Yassine, “Tell el Mazar, Field I: Preliminary Report of Area G, H, L, and M: The Summit,” ADAJ 27 (1983): 505; “Mazar, Tell el-,” ABD 4.645–646. RICHELLE: REVISITING THE AMMONITE OSTRACA 61 Line 4. At the end of line 4, one generally reads: ʾ [ḥk]. In reality, the lower left curve of the head of a dalet appears, as well as its tails and the opening at the top of the head. In addition, the following letter is not completely erased and the traces seem to indicate a yod (ig. 13). Dy could be the element designating “what is required, enough,”65 always used in construct. Here, on the analogy with Biblical Hebrew, it could be dy[h] with a sufix –h referring to plṭ, hence the meaning: “give plṭ what he needs,” or dy[n], hence “give plṭ what they [the young ladies] need.” Figure 13: yod 2.3 Jalul Ostracon 1 (see Pl. V) This ostracon has been remarkably well published in a recent article by R. E. Gane, and I only suggest two corrections. Text and translation: 1) 20 bn ʾḥʾ 4 1) 20 son of ʾAḥaʾ 4 2) 20(?) ḥlš bn ʾḥ̇tʾb 1 ḷ 2) 20(?) Ḥallaš son of ʾḥtʾb 1 seʾah(?) 3) 20 ḥlʾ 3 3) 20 Ḥalaʾ 3 4) [ ] nmš bn ʾmyʾ 1 ḷ 4) [ ] Nemeš son of ʾUmmayaʾ 1 seʾah (?) 5) 20 hṣlʾl bn ʾdʾ ? 2 ḷ 5) 20 and hṣlʾl son of ʾAddaʾ ? 2 seʾah (?) 6) ʿdl 6) ʿAdal Among the seven letters which have been read bet, ive are absolutely certain: this is the case of the four bets which occur in the word bn (“son of”) in lines 1, 2, 4 and 5, as well as of the letter at the end of line 2 65 HALOT, 219. 62 MAARAV 22.1–2 (2018) immediately preceding the numerical sign 1. In all of these occurrences, the head is open, as in all the other Ammonite ostraca. Conversely, the signs which are read bet in ʾbwtʾb (line 2) and in wbṣlʾl (line 5) in the editio princeps do not have an open top. In my opinion, these readings must be changed. Line 2. Here, moreover, the following letter cannot be a waw. As Gane himself has noted, it would be an “unusual” waw in the Ammonite corpus, and in order to maintain this reading, one must have recourse to parallels in the Palaeo-Hebrew script.66 The attested waw in Ammonite ostraca are totally different, generally formed of a short shaft and a horizontal curved stroke as head. In fact, the present letter is certainly not a waw. What should we read instead of the erroneous sequence bw? Notice that there is an ink connection between what has been read as two distinct letters. In my view, the only possibility is to read it as one letter, a ḥet. Admittedly, the fact that the right leg is incurved to the left is surprising; while acknowledging the dificulty, I see no better reading. The editio princeps proposed here an unusual personal name, ʾbw-tʾb, where tʾb is a verb (cf. Ps 119:40) and ʾbw means “father.” While not impossible, such a form for “father,” with preservation of the case ending, is a little surprising. If we are to read ʾḥtʾb (“my brother has longed for”), the component ʾḥ is absolutely regular (see ig. 14).67 Figure 14: Line 2 detail Line 5. In addition to the fact that the top of the alleged bet is not open, there is not the usual long tail strongly incurved to the left. Furthermore, regarding the preceding letter, the presence of the conjunction waw in the context of line 5 (at the beginning of the line, immediately after the numerical sign twenty) is surprising. In reality, if one considers the two signs read wb together, one obtains the normal shape of the letter he, 66 Gane (n 17): 81–82. It would be very unlikely to analyze ʾḥtʾb as composed with ʾb (‘father’) and an extremely rare Semitic root ʾḥt, attested only in Ethiopic (and Amharic) in a substantive which refers to an illness, “dysenterie” (D. Cohen, Dictionnaire des racines sémitiques ou attestées dans les langues sémitiques [10 fasc.; Paris: Mouton, 1970]: 1.15). 67 RICHELLE: REVISITING THE AMMONITE OSTRACA 63 exactly like in the attested Ammonite occurrences (Mazar 3, line 1 and ʿUmeyri 2, line 1; see the chart in the next section). It comprises two parallel strokes, with a little oblique stroke departing from the left one. As a result, instead of wbṣlʾl, I propose to read hṣlʾl, which is a personal name already attested in Ammonite (CAI 78, line 2; Mazar 7, line 6; probably ʿUmeyri 2, line 2). Figure 15: Line 5 detail 3. REVISITING THE PALAEOGRAPHY The following is not an exhaustive palaeographic study of the Ammonite ostraca. My aim here is to provide precise palaeographic charts, on a letter-by-letter basis, as a tool for further studies, with a few comments. The charts are based on the main ink ostraca: A1 to A3, Mazar 3, Jalul 1, ʿUmeyri 2 (Mazar 4 and 5 are badly preserved). For the sake of accuracy, I will refer to any letter from a given Ammonite inscription by means of the system elaborated by Rollston.68 For example, A1.2.n1 refers to the irst occurrence of the letter N in the second line of ostracon A1.69 Of course, such a system implicitly refers to speciic readings for each text. For A2,70 A3 and ʿUmeyri 2, I refer to the editio princeps. For A1, Mazar 3 and Jalul 1, I refer to the readings proposed in the last section. 68 Rollston, Writing (n 6): 97 n. 1. In addition, I will use abbreviations: Maz for Mazar; Jal for Jalul; Um for ʿUmeyri; Sir for Tell Siran bottle. 69 70 Regarding A2, Puech has proposed a detailed reading for the letters written in the right margin of line 1 and for those allegedly written in the left margin of line 4. With regard to the latter, they do not clearly appear on the photograph available to me and have not been read at all by Cross. As for the former, they clearly are a marginal addition, with only a few of them legible. Puech reads: ]bḥmsʿm(?) and Cross would recognize the word qrt. The only letters I am sure to read are a nun, a reš and a taw, with shapes quite similar to the same letters attested in the main inscription. Thus, in any case, they would not add information to the palaeographic study which follows. 64 MAARAV 22.1–2 (2018) ʾAlep A2.2.a1 A1.1.a1 A1.2.a1 A3.6.a1 A3.5.a1 Jal1.1.a1 Um2.1.a1 Maz3.1.a2 There are two main modes of penning this letter: (1) with a “check” crossing the vertical (A1.1.a1; A2.2.a1; Jal 1.1.a1) or only touching it (Um2.1.a1), and (2) with a side-ways “check” touching the vertical and a little stroke on the left, made separately (A1.2.a1; A3.6.a1; Maz3.1.a2). A3.5.a1 is probably a case of “extreme cursive.” Mode (1) already appears on the Tell Siran bottle; mode (2) is attested in Aramaic semi-formal cursive of the seventh century (Nimrud ostracon) as well as of the sixth century.71 Notice that (2) is typical of ink ostraca and different from what generally appears on seals.72 71 Cross, “Ammonite Ostraca from Tell Hisban” (n 6): 78; J. Naveh, The Development of the Aramaic Script (Jerusalem: Israel Academy of Sciences and Humanities, 1970): ig. 3. 72 On which see Herr, “Formal Scripts” (n 1): 21, or more recently Herr, “Aramaic and Ammonite Seal Scripts,” in “An Eye for Form” (n 30): 182–186. RICHELLE: REVISITING THE AMMONITE OSTRACA 65 Bet A2.3.b1(?) A2.4.b.1 A1.7.b2 A3.7.b1 A3.7.b2 Jal1.2.b1 Jal1.4.b1 Maz3.3.b1 In all of the ostraca, the top of the head of bet is open. In Ammonite formal cursive script, the opening of bet probably occurred around 600 b.c.e.: it is closed on seals from the seventh century bet (e.g., on WSS 857 ca. 700 b.c.e.,73 WSS 858 and 859 from the mid-seventh century b.c.e.)74 and on the Tell Siran bottle ca. 600 b.c.e., while on the seal of Baalis (early sixth century) it is open. There might be a trace of development in the ostraca, in that the head generally comprises two “teeth” as in a “U,” whereas in Jal1.4.b1 (and possibly A2.3.b1 and A2.4.b1) the right one is far less pronounced. This is also the case in Jalul 2. 73 Seal of “Beyadʾel servant of Padaʾel,” the latter most probably to be identiied with Pu-du-il king of Ammon mentioned by Sennacherib in 701 and by Esarhaddon in 675 (WSS 321). 74 Seals of “Adonipilleṭ servant of ʿAmminadab” and “Adoninur, servant of ʿAmminadab,” the latter being certainly the same as the one mentioned by Ashurbanipal in 667 (WSS 321–122). 66 MAARAV 22.1–2 (2018) Gimel A3.7.g1 The only assured occurrence of G is A3.7.g1. The right stroke clearly is longer than the left one, as in the Nimrud ostracon. Dalet A1.3.d1 A1.9.d1 A3.6.d1 Maz3.2.d1 Maz3.1.d1 Jal1.6.d1 Maz3.4.d1 The head of dalet is closed in A1.3.d1 (the spot on the upper-left part of the top is not ink) and A1.9.d1; the situation in A3.6.d1 is not so clear (and perhaps transitional); then the head is open in all the instances on Mazar 3 and even more in Jal1.6.d1. Here we have a patent typological development from A1 to A3 to Mazar 3 and Jalul 1. In the Ammonite seals script the opening appears at the end of the seventh century b.c.e.75 As Rollston notes, the Ammonite series “retains the closed head for considerably longer than the Aramaic series.”76 75 Herr, “Formal Scripts” (n 1): 21. 76 Rollston, “Northwest Semitic Cursive Scripts of Iron II” (n 30): 211. RICHELLE: REVISITING THE AMMONITE OSTRACA 67 He Jal1.5.h1 Um2.1.h1 Maz3.1.h1 Taking into account a previously unnoticed occurrence (Um 1.1.h1), it turns out that the three available attestations of he are formed by (1) two more or less parallel main strokes and (2) a small stroke departing from the middle of the left one at an angle of around 80° (Jal1.5.h1; Um 1.1.h1) or from the top of it at an angle of around 45° (Maz3.1.h1). The absence of any transverse stroke is remarkable and sets this form apart from the mainstream Aramaic writing tradition. It is also very different from the shape attested on seals77 and from the very peculiar form on the Tell Siran bottle. However, we ind almost the same form of he as in the Ammonite ostraca on an Aramaic papyrus from El Hibe,78 where the oblique little stroke on the left here might well be a vestige of the transverse bar, which is present in some instances and shows a tendency to move to the left. In this respect, it seems that Jal1.5.h1 and Um2.1.h1, where the little oblique stroke has moved down, are typologically more advanced than Maz3.1.h1. Waw Um2.3.w1 A1.6.w1 A1.7.w1 A3.5.w1 Maz3.2.w1 77 See the chart in Herr, “Formal Scripts” (n 1): ig. 1a, or WSS 865, 866, 874, 876. 78 Van der Kooij, Early North-West Semitic Script Traditions (n 3): 303. 68 MAARAV 22.1–2 (2018) Waw is formed by a short shaft and a curved head entirely written on its left. This is clearly different from the archaic “Y”-shape occurring on the Tell Siran bottle.79 Zayin A1.4.z1 A3.4.z1 A3.9.z1 Zayin so resembles yod that there are hesitations in some readings, especially in A1. Ḥet A2.4.ḥ1 A1.5.ḥ1 Jal1.3.ḥ1 A1.11.ḥ1 Jal1.2.ḥ1 A3.7.ḥ1 Maz3.1.ḥ1 Um2.4.ḥ1 In contrast to the two-bar ḥet of the Tell Siran bottle, ḥet is here formed by two parallel strokes and one oblique transverse stroke which departs from (1) the left one at about the middle of it (A1.5.ḥ1; A1.11.ḥ1; Maz3.1.ḥ1; Jal1.2.ḥ1; Jal1.3.ḥ1) or (2) at the bottom (A2.4.ḥ1; A3.7.ḥ1; Um2.4.ḥ1). (2), which is the typical Ammonite “inversed-N” shape, might be typologically more advanced than (1) which corresponds to the one-bar ḥet attested on seals from the seventh century.80 79 Cross, “Ammonite Ostraca from Tell Hisban” (n 6): 78. 80 For examples see Herr, “Formal Scripts” (n 1): ig. 1a. RICHELLE: REVISITING THE AMMONITE OSTRACA 69 Ṭet Maz3.1.ṭ1 Maz3.3.ṭ1 Ṭet only occurs on Mazar 3, in an elongated “U” shape crossed by an oblique stroke (this is more or less clear at least in Maz3.3.ṭ1). It can be compared to the shape attested in the Deir ʿAlla plaster inscription or to the semi-formal cursive Aramaic of the sixth–ifth centuries.81 In the Amman Citadel inscription and on Ammonite seals82 the letter was generally closed. Yod Jal1.4.y1 A1.10.y1 A1.7.y1 A3.11.y1 Maz3.5.y1 Here the head of yod is not penned with two parallel bars, but with two little strokes making an angle of around 45° with each other. Sometimes they are minimal (A1.10.y1; Maz3.5.y1); sometimes the lower stroke is prominent (A3.11.y1); sometimes it is the upper stroke (Jal1.4.y1). This is similar to the Aramaic semi-formal cursive of the sixth–early ifth centuries,83 but also to yod on the Tell Siran bottle and on Ammonite seals. 81 Van der Kooij, Early North-West Semitic Script Traditions (n 3): ig. 4; J. Naveh, Development of the Aramaic Script (n 71): igs. 3–5. 82 Herr, “Formal Scripts” (n 1): ig. 1a. 83 Naveh, Development of the Aramaic Script (n 71): ig. 3 n°2, 4, 5, 6; ig. 4 n°1, 2. 70 MAARAV 22.1–2 (2018) Kap A1.1.k1 Um2.1.k1 A1.1k2 A1.6.k1 A1.9.k1 Kap is uniformly penned with a wide curved head, which often enlarges to the left. As we have seen in the former section, in some of its occurrences it has been erroneously read as nun. The shape is well attested in Aramaic semi-formal cursive from the seventh century (Nimrud ostracon) to the early ifth century,84 but the fact that the head enlarges to the left might be a vestige of the triangular head attested in Ammonite inscriptions (Tell Siran bottle and seals85). Lamed A2.4.l.1 A1.9.l2 Jal1.3.l1 Um2.1.l1 A3.4.l1 Maz3.3.l2 As noted by Rollston,86 the Ammonite corpus contains occurrences of lamed with an angular hook as well as with a rounded hook. For instance, one observes both forms on the Amman Citadel inscription, although the rounded hook is more common. The same holds true with ostraca, but the use of broadstrokes to pen the hook makes it sometimes dificult to characterize it. 84 85 86 Ibid., ig. 3 n°1, 5; ig. 4 n°2. Herr, “Formal Scripts” (n 1): ig. 1a n°36. Rollston, “Northwest Semitic Cursive Scripts of Iron II” (n 30): 219. RICHELLE: REVISITING THE AMMONITE OSTRACA 71 Mem A2.3.m1 A2.4.m1 A1.4.m1 A3.2.m1 A1.3.m1 Jal1.4.m1 Maz 3.1.m1 Jal1.4.m2 Maz3.3.m1 Maz3.2.m1 In contrast to the “zigzag” head of mem on the Tell Siran bottle and on seals, Cross pointed out in A3 a new form where the zigzag is simpliied and cut by a little vertical stroke (see A3.2.m1; A3.7.m1).87 Rollston, too, writes that the “basic four-stroke head” is present in A1.88 The new form is a very common shape in Aramaic semi-formal cursive from the sixth and ifth century,89 and, interestingly, different from the shape in the Nimrud ostracon. Yet in reality, there is such a median stroke in A1 (A1.3.m1; A1.4.m1). It would be tempting to compare some instances with the shape where the head consists in two rounded “U” as in the Tell Siran bottle, but this remains uncertain. 87 88 89 Cross, “Ammonite Ostraca from Tell Hisban” (n 6): 79. Rollston, “Northwest Semitic Cursive Scripts of Iron II” (n 30): 220. Naveh, Development of the Aramaic Script (n 71): ig 3–4. 72 MAARAV 22.1–2 (2018) Nun A2.1.n1 Jal1.2.n1 A1.2.n1 Maz3.3.n1 Um2.2.n2 A3.10.n1 In the cursive Ammonite script, nun retains a relatively traditional and stable shape. Samek A1.6.s1 As Cross noticed, samek here is different from what occurs in Aramaic semi-formal cursive script from the late seventh–sixth centuries b.c.e.90 In this shape a simpliication of the “Z”-shaped head seems to have occurred. ʿAyin A2.3.ayin1 A1.6.ayin1 Maz3.2.ayin1 A3.10.ayin1 90 Cross, “Ammonite Ostraca from Tell Hisban” (n 6): 79. RICHELLE: REVISITING THE AMMONITE OSTRACA Jal1.6.ayin1 73 Um2.2.ayin1 Maz3.3.ayin2 ʿAyin is closed in A1.6.ayin1 and then uniformly open in the other ostraca, which conirms a typological development between A1 and A2. On seals from the irst half of the seventh century b.c.e. (WSS 857–858) it is closed, as well as on the Baalis seal (WSS 860) from the early sixth century. However, it is open on two seal impressions from Tell el-ʿUmeyri.91 Pe A3.7.p1 Um2.3.p1 A1.6.p1 Maz3.4.p1 The most remarkable feature here is the fact that the head is rounded. Ṣade Jal1.5.ṣ1 A1.2.ṣ1 Um2.1.ṣ1 A1.7.ṣ1 Maz3.3.ṣ1 91 Herr, “Epigraphic Finds from Tell el-ʻUmeiri during the 1989 season,” in Madaba Plains Project 3 (n 13): 325–326. 74 MAARAV 22.1–2 (2018) The shape of ṣade here is globally in line with the shape attested on Aramaic papyri from the sixth–ifth centuries, but the shaft is generally short, with the exception of a previously unnoticed instance in Mazar 3 (Maz3.3.ṣ1), which displays such an elongation of the downstroke that it had been taken for a taw. Qop A1.3.q1 A3.5.q1 A3.6.q1 A3.12.q1 A3.9.q1 Most of the occurrences of qop occur in A3. In what seems to be the basic ductus (A3.6.q1, A3.12.q1), the letter is formed by two strokes: (1) the longest part comprises the left half of the head and the shaft, drawn continuously, while (2) the other part only is a small curve constituting the right part of the head. As a result, the shape of the letter is clearly asymmetrical. There is even a tendency (A3.5.q1, A3.9.q1) to disconnect the right stroke from the rest of the letter. Overall, the shape of qop in A3 could be compared with two others: (1) the open-top qop attested in the Siran lask (Siran.8.q1), similar to the Greek letter psi; (2) the S-shaped top qop attested in Aramaic from the seventh century b.c.e. on (a developed form of which will be adopted later by Ammonite scribes, cf. A6.1.q1). As Cross noticed,92 it is likely that the shape of qop in A3 derives from (1), rather than from (2). Indeed, a similar development seems to be evidenced on Aramaic Neo-Assyrian clay tablets from the seventh century, where one can see almost the same shape as in A3,93 but also the psi-shape,94 as well as intermediate forms.95 So what we have here might be an inner-Ammonite development. That said, an interesting occurrence of qop (A1.3.q1) was overlooked by Cross and Puech, but noticed by van der Kooij and conirmed by 92 Cross, “Ammonite Ostraca from Tell Hisban” (n 6): 79. F. M. Fales, Aramaic Epigraphs on Clay Tablets of the Neo-Assyrian Period (Studi Semitici 2; Roma: La Sapienza, 1986): pl. XIII n°14, 16, 23. 94 Ibid., pl. XIII, n. 11. 93 95 Ibid., pl. XIV, n. 33, 46–48, 56. RICHELLE: REVISITING THE AMMONITE OSTRACA 75 Weigl and the present author. Whereas van der Kooij’s drawing96 makes it resemble the “classic” asymmetrical form attested on A3, it turns out that its ductus is notably different. In A1.3.q1, the two components (left and right) of the letter are almost symmetrical; they are two curved halves, drawn in a similar movement which makes each stroke thick above and thin below. Consequently, there is no shaft. This shape seems to be intended to imitate the circular top of an archaic qop. Precisely, in the Siran lask we see a circular open-top qop (Sir.8.q1). On the Amman Citadel stela (Cit.4.q1), most of the head is erased, but a very long shaft is well preserved, with no trace of crossing stroke, so that there is little doubt that the top was not S-shaped but circular. In sum, we can perhaps follow the development of the Ammonite qop, from a probable archaic circle-on-shaft shape (Amman Citadel inscription, Tell Siran bottle) to a circular shape in A1, to an asymmetrical shape in A3, before the scribes returned to the Aramaic script in A6. Reš A2.3.r1 A1.5.r1 Um2.3.r1 A3.5.r1 A3.6.r1 Maz3.3.r1 Contrary to van der Kooij’s drawing,97 the top of reš in A1 is not open (see, e.g., A1.5.r1). Moreover, whereas Cross believes that reš is opening in A2.3.r1 (p. 80), this seems very uncertain. The tops of A1.5.r1 and maybe A2.3.r1 are simply not traced as circular strokes; they are only small segments. A development is manifest in A3 (see A3.5.r1 and A3.6.r1 ) and Mazar 3 (Mazar3.3.r1), because the tops are clearly open. In sum, the situation here is different from what we noticed concerning bet (the head of which is open in all the Ammonite ostraca). With regard to reš, there is a clear typological development from A1 and A2 to A3 and Mazar 3. The development occurred between A1 to A2 or between 96 Van der Kooij, “Identity of Trans-Jordanian Alphabetic Writing” (n 5): 110; idem, Early North-West Semitic Script Traditions (n 3): 299. 97 Van der Kooij, “Identity of Trans-Jordanian Alphabetic Writing” (n 5): 110. 76 MAARAV 22.1–2 (2018) A2 and A3. Notice that R is still closed in the Tell Siran bottle (Sir.4.r1, Sir 4.r2, Sir.7.r1, Sir.8.r1). Šin Jal1.2.š1 A1.9.š1 A3.7.š2 Maz3.2.š2 Maz3.3.š1 The general trend in the development of šin in the Aramaic inscriptions is characterized by two changes. On the one hand, šin originally consists in four separate stroke-segments, but a trident-shape form appears later. On the other hand, the locus of the junction between the two internal strokes descends through time. Rollston notes that the same general trend is attested in the Ammonite corpus, but that some archaic forms of šin persist on the Siran inscription.98 Such archaic forms, with four strokes, seems to be the norm since they appear on the other lapidary Ammonite inscriptions (Amman Citadel inscription, Amman Theatre Inscription [CAI 58], inscribed eyes on Amman double-faced heads [CAI 73]; see also the Umm Udheinah bowl [CAI 148]). A similar, albeit less marked, tendency to conservatism may be observed on the ostraca. Thus in some instances, šin is clearly made of four separate strokes (A1.9.š1, Maz3.2.š2, Maz3.3.š1), while on two others, it seems that it is made of two “V” or check marks (A3.7.š2, possibly Jal1.2.š1). The junction between the internal strokes is generally not high; in one case (Jal1.2.š1), it is very low. 98 Rollston, “Northwest Semitic Cursive Scripts of Iron II” (n 30): 227. RICHELLE: REVISITING THE AMMONITE OSTRACA 77 Taw A2.2.t1 Jal1.2.t1 Um2.2.t1 A3.14.t1 A1.5.t2 A1.7.t1 Maz3.3.t2 In this elongated shape, the most remarkable feature is the fact that the cross-bar has moved to the right.99 CONCLUSION In comparison to other sets of West Semitic inscriptions (e.g., the Palaeo-Hebrew texts), the small corpus of the Ammonite ostraca has not attracted much attention among scholars, who have often been content in following the editio princeps. In this article I have tried to point out the importance of critically revisiting the readings and the palaeography of these ostraca, especially those written with ink, thanks to direct examination when possible, and in any case thanks to the excellent digitized photographs available today. 99 Cross, “Ammonite Ostraca from Tell Hisban” (n 6): 79. 130 MAARAV 22.1–2 (2018) PLATE II Ostracon A1 from Heshbon (Photograph by Bruce and Kenneth Zuckerman, West Semitic Research. Courtesy Department of Antiquities, Jordan) RICHELLE: REVISITING THE AMMONITE OSTRACA PLATE III Tell el-Mazar Ostracon 3 Recto (Photograph by Bruce Zuckerman and Marilyn Lundberg, West Semitic Research. Courtesy Department of Antiquities, Jordan) 131 132 MAARAV 22.1–2 (2018) PLATE IV Tell el-Mazar Ostracon 3 Verso (Photograph by Bruce Zuckerman and Marilyn Lundberg, West Semitic Research. Courtesy Department of Antiquities, Jordan) RICHELLE: REVISITING THE AMMONITE OSTRACA 133 PLATE V Jalul Ostracon 1 (Photography by David Sherwin, with authorization by R. E. Gane and R. E. Younker, Courtesy The Institute of Archaeology of Andrews University)