2.3.1 Group C.1: Naples V F 34 (B) and BnF 5048 (Vi)

The text of ms B, like manuscripts C, La, and Vi, written in Beneventan script, represents the earliest form of the grC textual tradition.

Following Blatt (27–28), Flusser dates B (AJ 1–16 + DEH) to the end of the 10th century and evidently interprets Blatt’s citation of the book list from Benevento in Zazo, “L’inventario dei libri antichi,” to mean the manuscript originated there.⁴⁴ The manuscript, currently in the National Library of Naples, has a 15th century ex-libris indicating it was in the Benevento “Chapter Library” and it is listed in the 1447 library catalogue as “liber Iosephi continens expositionem ueteris testamenti qui alio nomine dicitur liber antiquitatum.”⁴⁵ It is uncertain whether the manuscript was written at Benevento⁴⁶ or at Naples.⁴⁷ While Blatt (27) dates the manuscript to the end of the 10th century, he allows that from f. 20 (= AJ 1.251), it is “perhaps later.” Lowe, Brown, and Newton⁴⁸ all date the manuscript to the 11th century. It should be noted that they also date ms C to the early 11th century, and since that manuscript appears to be derived from B, B also must be from the beginning of the century. The manuscript has a number of missing pages, including the last three folios of AJ 16 (= 16.379b to 16.394),⁴⁹ which have been cut out, and the end of DEH (from 5.40.1 [408r; Ussani, 383, 7]).

The precedence of B in the grC tradition is based on two primary pieces of evidence: (1) For seven variants, the uncorrected text of B has the earliest reading, which has been corrected to a unique (and secondary) grC reading; (2) the insertion of material from BJ 1.82–84, written in what appears to be a hand different from the surrounding text, appears to be added to B and then subsequently is found in all grC manuscripts written in the same hand as the rest of that section of those manuscripts.⁵⁰

2.3.1.1 Manuscript B as Representative of Earliest grC Textual Form: Evidence from the Uncorrected Text

There are six place where B has the earliest reading, which is corrected to a unique grC variant, which then becomes the reading in all subsequent grC manuscripts:⁵¹

269. *Grippi] grC Agrippa (small a added at the beginning and final i changed into a); 278. cogere] grC cogere cepit (cepit added in the margin); a] grC ut a (ut added above the line); recedere] grC recederet (abreviated t attached to final e in B); 295. contristabatur] grC contristabantur (n added above the line in B and La); 307. occiderent] grC non occiderent (non added above the line in B)

Blatt had already noted, on the basis of evidence from Antiquities 5, that ms B had frequent corrections from what he called the “Cassinesis-group (C, La, Pt, v;⁵² in fact, these corrections are found in all grC manuscripts and not just these).⁵³ Whether these corrections derive from another grC manuscript or are the emendations of a scribe followed by later grC manuscripts, it is clear that the uncorrected text of ms B preserves an earlier form of the grC tradition than any other extant manuscript from that group.⁵⁴

View Full Size

Synopsis of the material surrounding the lacuna in grG ms St and grC manuscript B

• Download Figure
• Download figure as PowerPoint slide

2.3.2 Group C.2: Monte Cassino 124 (C); BML Plut. 66.1 (La); Bas. s. Pietri A 37 (Pt); Vat. lat. 1998 (V¹ and V²)

Flusser follows Blatt in dating ms C (AJ 1–20) to the 10th CE⁶¹ and in ascribing an “unknown” origin to the eleventh-century La (AJ 1–16 + DEH).⁶² Lowe, followed by all recent scholarship I have seen, dates both C and La to the first half of the 11th century, when he says both were written at Monte Cassino.⁶³ An eleventh-century date eliminates Blatt’s suggestion⁶⁴ that Monte Cassino 124 (C) is probably the Josephus manuscript Duke John III of Naples (928–968) ordered to be made for the library he enhanced after the death of his learned wife Theodora.⁶⁵

For the present study, it is important to correct Flusser’s failure to recognize the close connection between manuscripts C and La, which are both part of Levenson-Martin grC.2, which, like Blatt’s family β, also includes the 12/13th century Bas. s. Pietri A37 (Pt) and the 15th century Vat. lat 1995 (v). Blatt’s stemma indicates that v is to be connected with manuscripts C and La, but in this passage it is, in fact, closer to ms Pt.⁶⁶ As can be seen in Appendix 2, grC.2 can be identified by 39 unique readings, a number of them quite striking (e.g. citius for Hyrcanus; ad deum si iram for ad mensuram; Startoris for Stratonis). Appendix 2 also provides the data for the relationship of the manuscripts to one another with five agreements of Pt/v vs C/La (e.g. C/La share the error regineam for reginam), three agreements of C/Pt/v vs La (e.g. C/Pt/v share the errors artea for argentea and principi for principis), and no agreements of Pt/v/La vs C. This suggests that both La and Pt depend on C. Both manuscripts, however, have used at least one source other than C, since in each case listed in Appendix 2, when all four do not agree, either Pt/v or C/La has a reading found elsewhere, in most cases the reading found in the vast majority of manuscripts, which makes it difficult to identify the source with a specific manuscript or group.

C’s dependence on B is suggested by the fact that all the corrections in B as well as the insertion of the material from BJ 1 at the beginning of the lacuna appear as part of the text in C written in the same hand as the surrounding text. The almost identical format of the lacuna in C and B, not found elsewhere (blank lines in the second part of one column and continuing for the entire next column) also suggests C’s dependence on the earlier manuscript B.

2.3.3 Group C.3: Vat. lat. 1998 (V¹ and V²) and Pisa 20 (Pi)

According to the most recent scholarship, Vat. lat. 1998 is of Roman origin and can be dated between the end of the 11th and beginning of the 12th century.⁶⁷ The most significant mistake Flusser makes in his description of the four manuscripts which he identifies with the tradition used by Yosippon is his inaccurate account of the format of this manuscript. Beginning with his review of Blatt, he consistently describes it as including AJ 1–16 and DEH and asserts that the information that it contains AJ 1–20 in Blatt’s catalogue must be a careless mistake.⁶⁸ In fact, ms V does include AJ 1–20. However, the story is more complicated, because the manuscript has two parts: (1) AJ 1–20 (here designated V¹); and (2) a fragment of a manuscript consisting of the last page of AJ 16 followed by the DEH (here designated V²).⁶⁹ This is clear from the fact that AJ 20 ends on 167r (Flavii Iosephi Antiquitatis Iudaicae Liber XX Explicit), 167v is blank, and the next folio page (168) has the end of AJ 16⁷⁰ on the recto side and the beginning of DEH on the verso side.⁷¹ Vat. lat. 1998, as it came to be constructed, then, has evidence for both a manuscript with AJ 1–20 and an additional manuscript with AJ 1–16 + DEH. Both LAJ manuscripts, however, are clearly from the same group. A comparison of the variants at the end of AJ 16 in the first part of the manuscript (V¹; 136v) with those on the recto side of the folio page containing the beginning of DEH on the verso side (V²; 168r–v) reveals that the last page of AJ 16 belonging with DEH is very close to the uncorrected text of the last page of LAJ 16 in the manuscript with all twenty books.⁷² Furthermore, the similar paleography and illuminations of the first letter in the AJ and DEH books provide clear evidence that both the complete and fragmentary manuscript were produced at approximately the same time and place.⁷³

The variants in the last page of LAJ 16 in both V¹ and V² are closely related to the variants in the 13th-century (2nd half) Pisa, Biblioteca Cathariniana ms 20 (Pi), which is one of the manuscripts with AJ 1–16 + DEH mentioned (but not seen) by Flusser.⁷⁴ This is not surprising, because the Antiquities text of Pi is exceptionally close to the text of V throughout the sections collated for this chapter (see list of common variants in Appendix 2). An important clue to the relationship of these two manuscripts can be found in the margin of Pi, where at 16.369 the reading “uel mutauer(u)nt” appears (in what seems to be the same or a contemporaneous hand)⁷⁵ as a comment on the reading “nuntiauer(u)nt” found in the text. Since, of the selection of manuscripts collated for V², the variant nuntiauerunt is found only in Pi and V² and mutauerunt is found only in the V¹ LAJ 16, it appears that Pi had access to both versions of the ending of LAJ 16 in Vat. lat. 1998. This strongly suggests that for the Antiquities, Pi depends primarily on the LAJ 1–16 + DEH manuscript of which only one page from LAJ remains (V²). This is supported by a comparison of the ending of LAJ 16 in both manuscripts found in Vat. 1998. Pi and the two manuscripts in V consistently agree against the rest of the manuscripts collated for this passage, and Pi almost always agrees with V² against the text in V¹. The close connection of Pi and the DEH text in Vat. 1998 (V²) is confirmed by the many distinctive readings in a short section of each collated in connection with this project.⁷⁶ What this means for the text of the Antiquities is that Pi, from the second half of the 13th century, is a witness to at least as early a version of the text as that in the 11th/12th-century Vat. lat. 1998. Further study is required to determine if the Pi DEH text is a copy of the DEH text in Vat. lat. 1998, but that would be the most likely model at this stage of research.

Group C.3 is clearly identified by the close relationship between V and Pi in all the passages we have collated. The data presented here correspond closely with Blatt’s stemma, which locates the two manuscripts together as one branch of family γ and not, as Flusser argues on the basis of the format alone (since he had not seen either manuscript), part of a subgroup with B and La. For the sections of LAJ discussed in this chapter, Appendix 2 lists the 26 unique variants shared by only these two manuscripts as well as the 41 unique variants shared by Groups C.3 and C.4. Evidence that the text of Pi was influenced both by a manuscript like V² and one like V¹, just as was the case in AJ 16, can also be found in a corrected reading. At 13.257, the grC.3–4 reading Mariso (for Marisso) is corrected to Matriso, the reading in V¹. As in AJ 16, this suggests that the uncorrected text of Pi depends on the lost LAJ text of V², and the correction on V¹. The common format of Pi and V² (AJ 1–16 + DEH) also supports the idea that V² was Pi’s primary source, which was supplemented or corrected by V¹.

The fact that Vat. lat. 1998 preserves evidence from two manuscripts does not present a problem for Flusser’s hypothesis. In fact, to some degree it strengthens it by eliminating the potential problem of a manuscript comprising AJ 1–20 + DEH.

The sources of grC.3 are difficult to establish, but the general trajectory from grC.1 through grC.3 and then into grC.4 is clear. Because grC.3 does not have the multiple distinctive variants of grC.2, one of its main sources must ultimately derive from grC.1 and not grC.2. There are, however, four variants that indicate some connection between grC.3 and manuscript La alone, one of which is a common omission of eight words by haplography.⁷⁷ While this might indicate a common source, it is also possible that La had a direct, though minor, influence on the grC.3 manuscript tradition. In any case, evidence from the form of the lacuna at 13.315–320 in Pi, V, La, and B clearly indicates that grC.3 does not depend on ms La for this section, because, like B and C, manuscripts Pi and V have large blank spaces indicating the lacuna, while La has a continuous text with no indication of a lacuna. Comparison of the DEH texts in Pt/V, B, and La might clarify the relationship of La to grC.3. An initial probe collating B, La, Pi, and V for DEH 1.1.8–1.9 (Ussani, 7–14), however, provided clear evidence of a close relationship between Pi and V, but no tendency of Pi/V to agree with B against La or La against B.

2.3.4 Unclassified Manuscript Harley 3691

In addition to manuscripts B, La, Pi, and V, Flusser mentions Harley 3691 (hr), a 15th-century Italian manuscript now in the British Library.⁷⁸ Based solely on Blatt’s description, he reports only that it was written in 1457 and includes AJ 1–16, BJ, and DEH.⁷⁹ While noting the format of the manuscript, thereby implying some connection to the manuscripts including AJ 1–16 + DEH, he does not comment on how hr might be related to B, La, Pi, and V. The place of Harley 3691 in the Latin Antiquities 13 manuscript tradition and possible connections to Yosippon will be discussed in some detail at various points in this chapter. Here it is important to provide an accurate description of the manuscript, since all descriptions of its unique format, including Blatt’s, are misleading or incomplete.⁸⁰ A colophon at the end of the War and before the DEH states that Julian of Viterbo wrote it for Guido de Gonzaga in 1457 (222v). According to Watson, on the basis of the “Mantuan” decoration and the fact that Guido de Gonzaga (d. 1483) was bishop of Mantua, the manuscript was probably written in that city.⁸¹ While beginning with AJ 1–16 and ending with DEH, the text of the War, in fact, includes only 1.552–2.373 and 5.366–7.455, with 5.366b seamlessly and without any notice following 373a.⁸²

There are, of course, parallels to an edition with AJ 1–16 (B, La, Pi, V),⁸³ as discussed in the previous section in this chapter, as well as to the format AJ + BJ + DEH (Clm 15841)⁸⁴ and AJ + large extracts from BJ (mss Ne, pa, Vt, Ptr),⁸⁵ but almost everything else about the manuscript is eccentric:

1. The War excerpts are divided into two unequal books, the first designated Book 17 with three chapters covering BJ 1.552–2.116, and the second called Book 18 with ten chapters covering BJ 2.117–373 + 5.366–7.455.

2. The excerpts from the War are introduced with the comment that no more from the “stilus Iosephi” is found after Book 16 and that what follows are diverse translations (referring to BJ and DEH) up through Book 22 (i.e. AJ 1–16 + Books 17–18 [= BJ extracts] + Egesippus 1–4 [= DEH 1–5]).

3. The colophon in all capital letters, which precedes the information mentioned above about the scribe and date in the same script as the text of the manuscript, says that it is the end of AJ Book 12 (sic)⁸⁶ “and no more is found” (222V).

4. DEH is divided into four instead of five books (Book 2 = DEH 2–3),⁸⁷ but the colophon (without a reference to the date or scribe)⁸⁸ runs: EGESIPPI HISTORIE LIBER QUINTUS EXPLICIT.⁸⁹

While there are a number of mistakes and puzzling secondary unique readings in Harley 3691, the Antiquities text in the passages we have collated also clearly reflects early manuscript traditions. For example, in passages from AJ 6 and AJ 9,⁹⁰ the text is very close to the 6th/7th century papyrus Cimelio 1, which was produced at most a century after the Cassiodoran translation was made. In the passages collated from AJ 13 for this chapter, the hr text has 53 variants found elsewhere only in grC: 20 with all grC mss; 16 with grC.3–4; 11 with grC.4; 3 with grC.3; 2 with grC.1–3; 1 with grC.2 + grC.4. However, there is a much larger number of readings in hr that are not shared with unique grC readings. These are distributed among multiple groups, with only one group (grN) having as many as four cases of a unique group reading shared with hr.⁹¹

Finally, although its closeness to grC and its AJ 1–16 format point to some connection with the manuscripts Flusser suggest belong to the same group as Yosippon’s AJ source, the fact that hr clearly belongs with grC.4 (with generally later variants) is puzzling, because the other manuscripts with AJ 1–16 and DEH are either with grC.1 (B); grC.2 (La), or grC.3 (Pi V). More significantly, the DEH text of hr, in all the places I have checked, clearly belongs to a textual tradition quite different from the Cassinese group. It does not have the large lacuna in DEH 1.41 and the transposition of 2.8–2.9 to 5.53, by far the most striking characteristics of all manuscripts in this group. This difference is also seen in the text of DEH 1.1.8–1.9, for which B, La, Pi, V, and hr were collated for this project. In addition, at 5.22.1, hr has cera rather than cythara, the sole textual criterion Flusser used to identify the DEH text in B and La with Yosippon’s source.⁹² There is, however, one passage where a possible connection of hr to the source of Yosippon must be considered. At 13.396, hr and Yosippon both omit the series of names Samaria, Mount Carmel, and Mount Tabor (Ithaburium).⁹³

2.3.5 Variants Found Only in Group C

The most obvious feature of Group C is the enormous number of readings which differ from all other manuscripts. Only Group P, which consistently replaces and expands individual words and phrases, and the related grL.2 manuscripts Pd (“Codex Gigas”) and Prag XXIII.D.121, which significantly modify the tradition by frequent omissions, paraphrasing, and simplification of the language, have more unique variants.

The following list of unique Group C variants found in all grC manuscripts provides the evidence for the distinctiveness of this manuscript tradition and a resource for evaluating its connection to Sefer Yosippon and to any other medieval literature making use of this section of the Latin Antiquities. Lists of all unique variants for grC.2, grC.3, grC.4a, and grC.3–4a can be found in Appendix 2. Here the focus is on the distinctive character of the earliest extant layer of the Group C manuscript tradition as it first appears in the uncorrected text of Naples V F 34. The list can also be a tool for reconstructing the prehistory of the Group C tradition, making it easier to imagine how it might have appeared several generations before Sefer Yosippon was composed.

In order to include all the data from Group C.1, variants that appear in Groups C.1–2 and C.1–3 are also included in addition to all variants shared by all Group C manuscripts. The list also includes variants from other grC subgroups when they are clear variations of the readings listed for the above groups. The lemmata give the reading in grG ms St. In a few cases, one or two manuscripts from another group are also cited. Variants from hr are cited when they agree with grC manuscripts.

2.3.6 Group C and the Identification of the Earliest Readings

Although the primary purpose of this chapter is not to determine the earliest variants for each reading, it is helpful for the purpose of analyzing the development of the Latin Antiquities manuscript tradition to get a general sense of how far each group departs from the earliest recoverable text. The determination of the earliest text starts with a comparison with the Greek, but in the many cases when the Greek text does not provide clear evidence, it is necessary to judge each variant individually. Given the limited aims of this project, I have relied primarily on the relationship to the Greek in trying to identify what constitutes the earliest text. On the basis of this criterion, it is clear that the variants found only in grC represent a text that is farther from the earliest text than any other group aside from grP. For these variants found only in grC, there are, based on the Greek Vorlage, only three probable cases of the preservation of the earliest reading, conabatur instead of minabatur at 13.315,⁹⁵ promisit instead of dimisit at 13.322,⁹⁶ and Azoton rather than Azotum at 13.395.⁹⁷ This situation changes dramatically, however, when turning to variants found in grC and one or two other groups, where there are a number of cases of grC readings preserving the earliest reading. This is especially true of variants found only in both Group C and Group G.⁹⁸

2.4.1 Readings only in Group G

2.5.1 Clearly the Earliest Reading

2.5.2 Earliest Reading Cannot Be Determined

2.5.3 Clearly Secondary Reading

2.6.1 Clearly the Earliest Reading

2.6.2 Earliest Reading Cannot Be Determined

2.6.3 Clearly Secondary Reading

2.7.1 Clearly Earliest Reading

242. dierum] ἡµερῶν. grC grG grP; diebus all other mss

2.7.2 Clearly Secondary Reading

3.1.1 Variant Patterns That Do Not Correspond to an Established Group

The data from these names are not sufficient to classify twelve manuscripts into a single established group. The relationship of these manuscripts to established groups will have to rely on further evidence, some of which can be provided by variants from other passages we have collated for AJ 13. For example, Cb, ld, d, and n agree fully with each other, having the same combinations of grJ and grH readings. Pa and h agree fully with each other, having the same combination of grJ and grH readings, but different from the combination of readings in Cb, ld, d, and n. Elsewhere in the passages collated for AJ 13, d, n, and Pa exhibit variants from both grJ and grH, but grJ readings clearly predominate in the very closely related manuscripts n and d and grH readings in Pa. For the purpose of the analysis here, I have, therefore, classified n and d with grJ (rather than create a grJ.2) and Pa with grH (rather than create a grH.2), noting when these manuscripts differ from their respective groups.

3.1.2 Relationships among Groups

Although a full consideration of the relationships among groups is beyond the scope of this study, clear connections can be observed in the case of a number of groups and manuscripts, e.g. grC and grG; Groups E, M, N (with connections between E and N, and E and M); Aus and grM and grE; pg and grG; hr and grN; Ba and grH and grP. The direction of the development of the first reading can be easily reconstructed: Grippi (grG) to either (1) Agrippa (grC);¹⁰⁸ or to (2) Crippi (grH grP mss Ba G w); Crippi to either (1) Erippi (grJ) or (2) Crispi (grE grN grL; Crispi (grE grN grL) to Erispi (grM).

3.1.3 Seleucid Ruler Names and Sefer Yosippon

Unfortunately, of these names, only Eucherus appears in the Sefer Yosippon textual tradition.¹⁰⁹ In many LAJ manuscripts, Eucherus appears in both 13.370 and 13.376. In others, the earliest reading Acerus is found in 13.370 instead of Euc(h)erus (grC grG grE grN). If it was clear that the name Eucerus in Yosippon was influenced by its appearance in 13.370, it would be possible to eliminate those manuscripts that have Acerus. In fact, however, the name appears in the context of 13.376, as can be seen by a comparison of the Hebrew and Latin texts:

And they rebelled against King Alexander, and they went to King Demetrius, King of Macedonia, who is called Eucerus, and they brought him against Alexander for war, and Demetrius came and encamped at Shechem, and Demetrius had an army of 40,000 Macedonian warriors and 3,000 horsemen.¹¹⁰

Yosippon (FLUSSER, SY 33, 26–29 [p. 1.134])

AJ 13.376–377 (parallel to BJ 1.92–93¹¹¹ and DEH 1.10 [Ussani, 14–15]) is clearly a source for Yosippon at this point:

Then they sent to Demetrius Eucerus to ask him to be an ally. When, with a very great army, he reached those who had summoned him, he encamped around the city of Shechem … He (Demetrius) had 3,000 horsemen and 40,000 foot soldiers.¹¹²

AJ 13.376–377

While clearly parallel to the passage in 13.376, the phrase “who was called Eucerus” is reminiscent of the previous mention of Demetrius in 370 (Demetrius qui Acerus [or Eucerus] dicebatur). This would raise the possibility that Yosippon was reading an LAJ ms with Eucerus in both places, thus eliminating Groups C, G, E, N and mss G, w, hr, and pg. However, even if it Yosippon took the phrase “who was called” from 13.370, it might still have read Acerus, but he chose to use the name as it appeared in 13.376. The question is further complicated by the fact that the reference to Demetrius’ nickname in BJ 1.92, as in Yosippon, occurs in the context of the appeal to Demetrius and not in the context of Demetrius being made king in Damascus as in AJ 13.370.¹¹³ This might then suggest that Yosippon was also influenced by the account in BJ.¹¹⁴ The use of at least the Antiquities, however, is clear from the word for encamped (castra metatus est/ויחן in LAJ vs uenisset in LBJ).¹¹⁵

While the use of Eucerus in Yosippon is striking and raises some interesting questions about Yosippon’s use of LBJ in addition to LAJ, unfortunately it cannot be relied upon to identify the use of a particular LAJ manuscript group.

AJ 13.395–397 and Sefer Yosippon

The following table presents a succinct overview of the relationship between the names in the Latin text and its Greek source and between the names in the Hebrew text and its Latin source. As an initial reference, the names are given as they appear in grG manuscript St and in Flusser’s edition (with the many emendations he proposes indicated by an asterisk). Variants for the names in each language are, of course, of the greatest importance for this study and will be listed and briefly discussed in the commentary following table 4, which focuses on the examples most relevant for understanding the LAJ textual tradition and its relationship to Sefer Yosippon. Since Flusser’s apparatus is not comprehensive, additional variants are listed when relevant.¹¹⁹

View Full Size

Greek, Latin, and Hebrew city names (* = emendation by Flusser)

• Download Figure
• Download figure as PowerPoint slide

3.2.2 Commentary

Apolloniam (Ἀπολλωνίαν; ואפולוניא)] all mss except Appollonium grC.3–4a; Antoniam grC.4b

Yosippon cannot depend on a manuscript in grC.3–4, which would include grC.3 mss V and Pi, which have the format AJ 1–16 + DEH.

Azotum (Ἄζωτον; ואשדוד)] Azoton C.1 C.3–4a (-O pat M l) Ptr; Azotan M l; Azaton pat; Azotam rg

While the grC.1 and grC.3–4a reading Azoton is probably the earliest reading because of its closeness to the Greek, Yosippon’s use of the biblical name makes it impossible to determine the reading in its Latin source.

Raphiam (Ῥάφειαν; ורפיס)] all mss except grJ, which has Raphia

Flusser prints his conjecture ורפיס ¹²⁰ but comments that “perhaps ורפים should be read,” which must be certainly correct, since the sameḥ here, as often elsewhere, is a misreading of final mem. Flusser’s conjecture can be supported by the appearance of the place name in SY 32, 73, where ורפיס is found in all SY mss I have seen. The proposed final mem representing the accusative form of the name would correspond to all LAJ mss except those in grJ.

Rinocora (Ῥινοκόρουρα; ורינוקורה)] grG (-Werd Best7010 GKS1571) grC.1–4a (-M l Vt) grL.1 pg Ly; Rinocoram C.4b M l Vt hr GKS1571; Rinocoro grH (Br Rinocero) grJ (- Ly) grN grP (Prs: Arynocoro) Ba G Werd Best7010 Aus Cp Pr; Rinocoron grE; Rincoro grM (-Aus); Rinocorura grL.2 (Pd: Rinocoruram; PragXXIII.D121: Rinocoruca) (cf. Ῥινοκόρουρα, the reading in almost all Grk mss); Ronocoruram w (transposed to after Cilicum)

Only grL.2 manuscripts have the reading Rinocorura, which is closest to the Greek (note also Ronocoruram in ms w, which shares a number of distinctive readings with grL.1 manuscripts). Of greatest significance for our analysis, Groups C, G, and L.1 have the same reading as Yosippon, which Flusser appropriately emends fromודינוקורה .¹²¹ This provides clear evidence for a connection of one of these groups to Yosippon’s LAJ source.¹²²

Omission of in mediterraneis uero per Idumeam

Both Sefer Yosippon and Groups E, L.2, N, and P and mss Ba, G, and w omit the geographical notice in mediterraneis uero per Idumeam. (The same groups and manuscripts also omit the words populares uero non eis obsequebantur in 13.298, but in that passage these words are reflected in Yosippon’s Hebrew text.) This common omission of what might be considered non-essential information might well be coincidental. However, the fact that the reading Baoronee in Groups N, L, P and mss Ba, G, and w best corresponds to Yosippon’s source (see below) is worth noting and suggests the value of exploring other connections between readings in Yosippon and variants in these groups.

Aboram (Ἄδωρα; ואת חברון)] Abora grC.2; Aboran Ly Mk U; omit w

In 13.257 Aboram (v.l. Adoreon; see p. 266) is transliterated as חברה, while here the biblical name חברון is used.

Marissam (Μάρισαν; ואת מרשה)] Marissimam C.1 C.3 C.4a; Marissima C.2; Marisam C.4b; Maresam grP; Marissa Alb; Marissan Ly Mk U; Marissiam Aus

In grC.1–4a, Marissima(m) appears in this passage instead of a form of Marissa(m) or Maresa. In 13.257 all grC mss have Marisso together with Groups G, J, and N. Yosippon has the biblical מרשה in both passages, making it impossible to determine which form of the Latin appeared in its source. For other cases of using biblical names, see Azotum, Gazam, Sebon, and Aboram in Table 4 above.

The reading מרשה אשר לאדום (“which belongs to Edom”), which Flusser rejects in favor of מרשה, is probably the original reading because (1) it corresponds to Aboram Marissam omnemque Idumaeum; (2) the same phrase is also found in the account of Hyrcanus’ capturing the city (SY 33); (3) it would make a reasonable four-part division of the cities corresponding to the introduction of the list in Yosippon: on the coast, followed by three areas each demarcated with ל אשר (“which belong to”): to Idumaea, to Syria, and to Moab; (4) it appears in two other mss, Vat. Ebr 52 and Vat. Borg. 1, in addition to Jerusalem oct. 41280, which is the only ms cited for this variant in Flusser’s apparatus; (5) it could have easily been omitted by a scribe who, seeing ואת מרשה אשר לאדום ואת שיטופולי אשר לארם, confused אדום with ארם and eliminated the former as a redundancy.

Omission of Samariam, Carmelum montem, et Itaburium montem and Ioppem, Iamniam

Yosippon shares with ms hr the omission of the sequence Samariam, Carmelum montem, et Itaburium montem, an omission found nowhere else in the LAJ manuscript tradition. This is particularly intriguing because hr, like mss B, La, V, and Pi, includes AJ 1–16 together with DEH. In addition, the grC tradition clearly provided a major source for the Antiquities in ms hr. However, ms hr has the closest connections to grC.4, making it a somewhat unlikely witness to the earliest layer of grC tradition. Of course, the common omission might be accidental, and Yosippon also omits Ioppe and Iamnia, which are found in all Latin manuscripts. Nevertheless, the case for a relationship between hr and Yosippon would be considerably strengthened if other distinctive connections emerge.

Gabala Moabitidem (Γάβαλα Μωαβίτιδας; גבלה אשר למואב)] Gabela Moabitidem Aus; Gabala Moabitiden Ly Mk U

Yosippon takes Moabitidem as an adjective modifying Gabala. Of the manuscripts that have punctuation marks separating the cities (almost all), grG ms St and grM mss Vat and rg are the only ones I have seen that do not have a punctuation mark separating the two words and thus possibly reading Moabitidem as an adjective describing Gabala.

Medaba (Μήδαβα; מידבא)] grG pg; Midabalam G re; Midabalem all other mss

Group G and the closely related ms pg read Medaba here and at 13.255, where several different variants appear (see p. 264). Here all groups except grG read Midabalem, which transfers the first syllable of the next word to the end of Midaba (medabalembaoronem). It should be noted that Midabalem is clearly intended to be read as one word, because virtually all the manuscripts have clear punctuation marks separating the words and not just spaces. Yosippon uses the biblical form מידבא, whose first syllable could be pronounced as ī or ē (as the Masoretic tradition vocalizes it).

Lembaoronem (Λεµβὰ Ορωναιµαγελ[εθων, Λεµβὰ Ορωναι µαιται λ[αιθωνα, Λεµβὰ Ορωναι αἱµαταιλ[αιθωνα]; בחורון] grG (Ml Lembada.Oronem); Lemboronee pg; Baoro C.1 C.2 (- Pt); Baora C.3–4 (V: Boara; Ptr O: Bocora) Pt; Baoronęę grL (Pd PragXIII.D121: Bagronee) grN grP Ba; Baorenee grE grH (Br: orenee) grJ (- Ly Mk U) grM (Aus: Baoreuce) G Pr; Baorene Ly Mk U; Borane hr; Barronee w

The Greek text for the next two (or three) words is uncertain. Niese lists a number of variants (three are cited here). For the Latin he gives oronemegaeton (for the word after Lemba and including Mega and et on[zora]). This reproduces the grC reading from Naples V F 34, but without the incorrect word division marked in all grC manuscripts. As we have seen, in all manuscripts outside of grG and the closely related pg, Lem is attached to the previous word, yielding for the next word the variants Baoro, Baora, Baoronee (often spelled Baoronȩȩ; cf. the Greek variant Ορωναι αἱµαταιλ- [αι = e]), Baorenee, Barrone, and Borane. All grC mss have the ending of the word (-ne) attached to the next word (= Nemega). Based on the Greek and Latin variants, Lembaoronee is probably the earliest reading.

In any case, Yosippon’s reading (בחורון) would correspond best to Baoronee (Groups N, L, P, and ms Ba). However, with the proper word division of the reading in the source of the grC archetype (Baorone|mega), grC might have preserved evidence of a reading not in the extant manuscript tradition but equally close to that in Yosippon.¹²³

Mega (µαγελ-; מיגן)] grG; Nemega grC; Maga grH grJ grL grM grP Ba G w Cp Pr; Magam hr; Magnam grE grN; pg mag(eton zora)

The uncertainty of the Greek text makes it impossible to determine if the next word was originally a separate word or part of the previous word in the Greek manuscript tradition. What is clear is that it is a separate word in Yosippon and the entire LAJ manuscript tradition (with the exception of ms pg). It is also clear that Yosippon depends on the form Mega and not Nemega (grC) or a form beginning with Ma-. It is possible, however, that Yosippon was using the source of the grC archetype, which, like grG, would have read Mega, with the ne belonging to the previous word. The origin of the final nun is not clear. Perhaps it was influenced by forms such as אולן and פילן, which follow it.¹²⁴

Aulonem (αὐλῶνα; אולן)] Oculonem grC (M l Aulonem; B? [hard to distinguish oc from a])

All groups except for group C have the reading corresponding to that in Yosippon. Groups C’s Oculonem (hr Occulonem) derives from a misreading of Beneventan a as oc. The reading in Naples V F 34 (B) could possibly (but not likely) be read as Aulonem. If so, this would be one more example of ms B agreeing with an earlier reading in that manuscript against all the other grC manuscripts.

Pellante (Πέλλαν; פילן)] grC.2 grH grJ grM grP B No Ba G; Pellantem grC.3 grC.4b grE grL Vi w pg; Pellente grG (- Lau Ml) grC.4 grN (-ve No); Pellentem Lau Ml hr ve

Yosippon depends on a Latin source with either Pellante or Pellantem.

Hanc (i.e. Pellante[m]) etiam destruxit, cum non promisissent habitantes in ea patrios Iudaeorum se mores suscipere. Alias quoque Syriae ciuitates euerterunt

ταύτην δὲ κατέσκαψεν οὐχ ὑποσχοµένων τῶν ἐνοικούντων ἐς τὰ πάτρια τῶν Ἰουδαίων ἔθη µεταβαλεῖσθαι, ἄλλας τε πόλεις πρωτευούσας τῆς Συρίας ἃ ἦσαν κατεστραµµέναι.

δὲ] om. P Niese. κατέσκαψεν] κατέσκαψαν F V W Naber Marcus. οὐχ] omitted by P Niese¹²⁵ ἐς τὰ] ἐς P Niese. µεταβαλεῖσθαι] P Niese Naber Marcus; µεταβαλέσθαι LAMVW Nodet. ἃ ἦσαν] ἦσαν P Niese. κατεστραµµέναι] κατεστραµµένοι P Niese

אלה הערים אשר לא הרס אלכסנדר כי באו בברית עמו וימולו את בשר ערלתם וישבו בעריהם, ויתר ערי ארם הרס המלך‪.‬

Both the Greek and Latin textual tradition are exceedingly complex for the final sentence of this passage and present a number of significant problems that cannot be addressed here. The following comments are focused on understanding the relationship of Groups G and C and on the text Yosippon might have been reading.

destruxit] Singular and plural forms appear here in both the Greek and the Latin textual traditions (κατέσκαψεν/κατέσκαψαν; destruxit/destruxerunt).¹²⁶ Yosippon is clearly reading a Latin text with the singular (הרס אלכסנדר).

cum non promisissent] This reading corresponds exactly to οὐχ ὑποσχοµένων, providing a close translation of the Greek (“He destroyed the city, because its inhabitants had not promissed that they would adopt the ancestral customs of the Jews”). Niese prints ὑποσχοµένων (i.e. without οὐχ) found only in P, but he is not followed by any subsequent editors. The grC reading cum promisisset is clearly secondary, since it does not correspond to the plural ὑποσχοµένων, and makes little sense in the context. Presumably it is concessive (“he destroyed the city, although he had promissed …”), but the rest of the sentence does not make clear what he promised. The reading cum promisissent (only in pg), however, does make sense (as does ὑποσχοµένων without οὐχ), even if it is probably not the earliest reading: “He destroyed this city, even though its inhabitants promissed to adopt the ancestral customs of the Jews.”

Of particular note is the fact that the uncorrected reading in St has cum promisisset, the same reading as in the unique grC reading, which is corrected to non promisissent, the reading of all the other grG manuscripts (as well as of a number of other manuscripts). Similarly, at 13.254, the uncorrected text of St omits eas as do all the grC mss, with eas added above the line. On the other hand, at 13.265 the uncorrected text of St has the uniquely grG reading uacuum habuerit with the unique grC.3–4 reading uacauerit above the line. It seems then that a grC manuscript was available to both the original scribe and to a corrector.¹²⁷ Whether this has any connection to the clear relationship between grG and grC, established above on the basis of multiple readings found only in these two groups, requires a full analysis of the textual affinities of the corrections in St.¹²⁸ At the very least it provides a case of grC and grG manuscripts in close proximity to one another.¹²⁹

Yosippon changes his source to produce a different but clear text: “These are the cities which Alexander did not destroy because they entered a covenant with him, and they circumcised the flesh of their foreskins, and they remained in their cities; and the rest of the cities of Syria the king destroyed.” In effect, it is as if Yosippon is reading non before destruxit instead of before promisissent. While not offering direct evidence, Yosippon’s text provides some support for the reading cum non promisissent,¹³⁰ since it nowhere suggests the idea that Alexander destroyed a city, even though (concessive cum) the inhabitants adopted Jewish customs.

se mores] All manuscripts except those in grG and grE and ms pg and grL ms t have the reading seniores, easily explained as reading m as ni. This produces a highly problematic text, with the relationship between patrios and seniores obscure.¹³¹ Here Yosippon is almost certainly reading a manuscript with se mores. In any case, there is no evidence for the reading seniores in Yosippon.¹³²

Hanc … euerterunt] This entire section is omitted by all grJ manuscripts. It is also omitted by the Lübeck edition, which is based on the grJ manuscript tradition, and therefore is also missing from the 1514 Paris and 1519 Paris editions, based on the Lübeck edition. This accounts for its absence in the 1524 Basel edition, Flusser’s primary source for the Latin Antiquities, which sometimes modified the 1524 Cologne edition using the text from one of these editions.¹³³

3.2.3 Summary and Conclusions for Case Study 2

1. The earliest text of this passage is found in grG, with the following exceptions:

   1. Rinocorura in grL.1 is probably the earliest reading because it reproduces the Greek;

   2. Mega in grG and in the source of grC variant Nemega is probably secondary because all Greek mss and all other Latin manuscript groups have a name beginning with the syllable ma;

   3. Pellante(m) rather than Pellente(m) in grG grN grC.4 hr corresponds better to the Greek Πέλλαν;

   4. Azoton in grC.1 and grC.3–4a is closer to the Greek Ἄζωτον than Azotum, found in almost all other manuscripts;

   5. promisisset in the uncorrected text of grG ms St in 13.397, which agrees with grC, is later than non promisissent in all other manuscripts.

2. Only group G has Lembaoronem (with the original first syllable), and only grG, grE, and ms t have preserved the original reading se mores.

3. Two lacunae provide important evidence for the establishment and connection of groups:

   1. The omission of in mediterraneis uero per Idumeam by all manuscripts from Groups E, L.2, N, and P and manuscripts Ba, G, and w indicates a clear connection, which is confirmed by the same manuscripts omitting populares uero non eis obsequebantur in 13.298;

   2. the omission of 13.397b (hanc … euerterunt) by all nine manuscripts in grJ is a significant marker for that group.

4. The obviously secondary readings in all almost all grC mss—Baora/Baoro, Nemega, Oculonem—clearly derive from earlier readings in the source of the grC archetype: Baorone, Mega, Aulonem, of which Baorone and Mega are connected to grG (grG Lembaoronem) becomes Baoronem after the first syllable is attached to Midaba and the last to Mega).

5. This passage provides potentially useful information about Yosippon’s Latin source for the following readings:

   1. רינוקורה Rinocora (grC grG grL.1) rather than Rinocoro (grH grJ grN grP), Rinocoron (grE), Rinocorura (grL.2), or Rincoro (grM).

   2. בחורון Baoronee (grL grN grP Ba) rather than Lembaoronem (grG), Baorene(e) (grE grH grJ grM G), Baoro (grC), Barronee (w), Barrone (hr), Lemboronee (pg).

   3. מיגן Mega (grG) rather than Nemega (grC), Maga (grH grJ grL grM grP Ba G w), Magnam (grE grN), Magam (hr).

   4. אולן Aulonem in all mss except grC (Oculonem) hr (Occulonem).

   5. פילן Pellante(m) (grC.1–3 C.4b grE grH grJ grL grM grP Ba G w pg) rather than Pellente(m) (grC.4 grG grN hr).

   6. Omission of in mediterraneis uero per Idumeam: SY and grE, grL.2, grN, grP, Ba G w.

   7. Omission of Samariam, Carmelum montem, et Ithaburium montem: SY and hr.

   8. כי באו בברית עמו וימולו את בשר ערלתם (“because they entered into a covenant with them and they circumcised the flesh of their foreskin”) corresponds to the reading se mores in grG, grE, ms pg, and grL ms t better than to the obscure reference to patrios … seniores in the other groups and manuscripts.

3.3.1 Commentary

254. eas (αὐτὰς)] omitted grC; added above St

One of two places the original reading in St agrees with a unique grC variant and is corrected to the reading found in all other manuscripts.¹³⁸ Based on the Greek text, the grC omission of eas is obviously secondary.

255. Medaba (Μήδαβαν; מידבתה)] grG (- Lau Tr Ml) grM grP; Medabam grL grE (- al); Minadabam grC.1 Ptr; Minadam grC.2; Nadabam grC.3–4; Midaba grH grN (L nādabam to nāmidabam) grJ Ba; Medebam al; Nabadam hr; Bedaba Lau Tr Ml

The place name Medaba is found in five places in LAJ and exhibits a variety of forms in the various groups and even within the same manuscript.¹³⁹ See p. 255 for discussion of the name in 13.397. Here Medaba is found in Groups G, M, and P, with the closely related forms Medabam and Medebam in Groups L and M. Bedaba is a clear indication of the distinct grG subgroup Lau Tr Ml. The grC subgroups present a typical case of grC.2 and grC.3 independently deriving from grC.1. Manuscript hr, as is often the case, has the reading in grC.3–4.

Yosippon’s reading here, as at 13.397, probably derives from the biblical form (מידבא) and not from the Latin variant Midaba or Medaba (the Masoretic tradition vocalizes the Hebrew with ē in the first syllable).

sexto mense (ἕκτῳ µηνὶ; ששה חודשים)] intra septem menses grC.1 (B intra s.l.) grC.3–4 hr; intra menses septem grC.2

All grC mss have intra septem menses. The uncorrected reading in ms B (septem menses) would mean “for seven months” rather than within seven months.

Yosippon was most probably reading a text with sexto mense rather than intra septem menses (or septem menses), since “six months” is closer to “in the sixth month” than “within seven months.” The uncorrected variant in B (“for seven months”) could clearly not be the reading in Yosippon’s source.

Samogan (Σαµόγαν; שמגן)] grG grC.1–2 V L; Samogam (often spelled Samogā) in all other manuscripts

The reading in Yosippon clearly corresponds to the earliest reading, found only in grG grC.1–2 grC.3 ms V and grN ms L. Flusser evidently did not know (or notice in B or La) this reading, since he prints Samogam in his commentary, which is found in the 1524 Basel edition as well as in the vast majority of mss.

256. ac Garizin gentemque (καὶ Γαριζεὶν τό τε [Κουθαίων] γένος; הר גריזים; cf. ויך את הכותים in the continuation of the story)] grG (- Ml) grJ (- li) w; nargariz ingentemque grC.1–3 (Vi nagariz in gentemque) Ptr; narzari ingentemque grC.4a (M narzani ingentemque); nazarinque ingente grC.4b; agarizin gentemque grN Ba G; garizim gentemque grL; et argarizim gentemque grE; ac Garizim gentemque grH grM li; et Garizim gentemque grP pg Ml; nagariz gentemque hr

The different variants correspond well to the established groups, with the reading in grG and grJ being closest to the Greek. Influence from the parallel passage in BJ 1.63 probably accounts for ac Garizin becoming Argarizin in grE (see below for another possible example of a grE reading being influenced by BJ 1.63).

Just as in 13.397, an incorrect word division in the grC archetype has given rise to names found only in all grC manuscripts: argarizin gentemque becomes nargariz ingentemque in grC.1–3, which then is corrupted to narzari ingentemque in grC.4 grC.4a. The initial n might derive from the last letter of the previous word Syciman, although the spelling with n rather than m or a line over the a is found only in Ba. Appearance in that 9th-century text does indicate that the variant with n is early. In any case, this is clear evidence that the source of the grC archetype had the reading Argarizin.

The reading הר גריזים in Yosippon almost certainly derives from the biblical form rather than from a secondary variant in either grE or in the source of grC.1–3. הר גריזים translating Garizin or Garizim is also found in in SY 10, 56, based on AJ 11.340.

(fratrem) Iaddi ((Ἰαδδοῦς; Ιαδδου΄Ιαδδουν); עידו] grG grH grJ grM; Iaddo grC grL w; Ieddo grE grN pg Ba G; Iadi hr; Iaddonis grP

In addition to this passage, the name of the high priest who met Alexander the Great occurs five times in AJ 11 (302, 306, 322, 326, and 347), where there are variants in both the Greek and Latin manuscript tradition, e.g. Ιαδδους, Ιωαδδους; Iaddus, Ioaddas, Ioiadas, Ioaddus, Iaadus). In most manuscripts the name is not spelled the same way in the different passages.¹⁴⁰ In the Greek Bible, this name, in its various spellings (Ιαδου, Iωαδα, and Ιωιαδα), corresponds to the High Priest ידוע in Neh 12:11 (II Esd 22:11), while Iωαδα and Ιωιαδα are also found as variants for Ιωδαε, which corresponds to יוידע, the grandfather of ידוע, who is also mentioned in Neh 12:11.

It is noteworthy that Yosippon names the High Priest עדו in the story of Alexander in SY 10 (AJ 11) and עידו in the reference to that story here in SY 29 (AJ 13).¹⁴¹ This corresponds to the name עִדּוֹא, a different priest, who is mentioned in the same list in Neh 12. Flusser suggests that Yosippon used this name because it is closer to Iaddo, a variant found in both manuscripts B and La, which, according to him, were in the same group as SY’s source. For the present project this would also seem to be evidence for the use of a grC ms, since the reading is found in all grC manuscripts. However, the same variant is also found in all grL mss and ms w.¹⁴² Furthermore, the Hebrew עידו could just as easily reflect the reading Ieddo in grN, grE, and mss Ba and G.¹⁴³

257. Abora cum Marisso (Ἄδωρα καὶ Μάρισαν; חברה מרשה)] abora cum] aboracum grL grH; adoreon grE grM; adoreon et grP; Marisso] grC.1–2 grG (- Ml) grJ grN; Marissam grL (b Marissum corr. to Marissam); maresan grE grM grP; Mariso grC.3–4 (V Matriso; Pi mariso corr. to matriso)¹⁴⁴ Ptr hr No; Mariso cepit Ml

Abora cum Marisso, the reading in all mss except Groups L, E, M, and P, is a possible, but odd translation of the Greek. The reading Aboracum Marissam/Marissum in grL (grH has Aboracum Marisso) could either be an attempt to correct the text or an earlier reading with aboracum later read as two words, which would require the ablative form for Marissum. The reading Adoreon in grE, grM, and grP is closer to the Greek, and grP’s adoreon et maresan corresponds exactly to the Greek. However, the specific spellings Adoreon and Maresan and the fact that grE elsewhere appears to be influenced by a BJ reading for this passage (see Argarizin above) make it somewhat more likely that this is a secondary correction, based on LBJ 1.63, of an awkward reading.

Yosippon is clearly reading a text with Abora and not Adora or Adoreon. Here חברה transliterates the Latin, while in 13.396 the biblical name חברון is used for Abora. The lack of the connective is probably accidental and not related to the same phenomenon in grE and grM. The spelling מרשה, as in 13.396, derives from the biblical spelling and not necessarily from a form without an i.¹⁴⁵

260. Fannius Marci filius consul omnem senatum octauo Idus Februarias in campo iussit conuenire praesente Lucio Manlio, Lucii Mentini filio, et Gaio Sempronio Falernae filio;

Φάννιος Μάρκου υἱὸς στρατηγὸς βουλὴν ἤγαγεν πρὸ ὀκτὼ εἰδῶν Φεβρουαρίων ἐν Κοµιτίῳ¹⁴⁶ παρόντος Λουκίου Μαννίου Λουκίου υἱοῦ Μεντίνα καὶ Γαΐου Σεµπρωνίου Πενναίου¹⁴⁷ υἱοῦ Φαλέρνα; פניאוס בן מרקוס ולוקיאוס ומליאוס בני מנטינוס גיוס סמפרוניאוס בן פלרנוס

omnem senatum] senatum omne grC.2; Februarias] grC.1–2 grG pg; Februarii grJ Mn H Sa hr ve No al; febr b Adm Bo Cl Nv grM; Feb Ba L El; Februariarum Ptr; Februarius grC.3–4; Februario mense grP; campo] grG (- Lau); campum all other mss except Lau (templo); Manlio] Mallio grC (- Ne^p.c. Plut18sin10); manlicio grP; Lucii] Luci grN pg Ba G w; omitted grC; Mentini] Mentinii grL; Lucii Mentini] lumentini grP

In both his 1959 review of Blatt and the introduction to his edition, Flusser utilizes the variant Mallio for Manlio to prove that Yosippon was reading a manuscript belonging to the group represented by the manuscripts with the format AJ 1–16 + DEH. This was based on Mallio in mss B and La (the only two manuscripts he had seen from the group) and Manlio in the 1524 Basel edition.¹⁴⁸ Here Flusser is clearly correct. The reading is in fact found in 20 grC manuscripts and nowhere else in the LAJ manuscript tradition. Furthermore, Flusser’s argument can be made even stronger by noting that Yosippon also follows grC by omitting Lucii, which is missing only in this group (SY: Lucius and Mallius, the sons of Mentinus; grC: Lucius Mallius, the son of Mentinus). Given the limited resources with which he was working, it is quite impressive that Flusser was able to identify the only unambiguous example of Yosippon using a manuscript related to grC that we have found in this part of the Antiquities.

3.3.2 Summary and Conclusions for Case Study 3

1. Five place names in this passage have multiple variants that make it possible to clearly identify specific manuscript groups and subgroups: Samogan (grC grG), Medaba (grC.1, grC.2, grC.3–4), ac Garizin gentemque (grC.1–3, grC.4, grE, grL, grN, grP), Abora (grL, grP), Marissa (grC.3, grL).

2. Because Yosippon uses the biblical names for Medaba, Garizin, and Marissa, it is not possible to determine which name would have been in its Latin source. Yosippon was clearly using a source with Abora and not Adoreon (grM, grN, grP). Samogan (vs Samogam) is found only in grC and grG manuscripts, indicating a connection between one of these groups and Yosippon, which has שמגן.

3. Yosippon’s ששה חודשים is more easily derived from the reading sexto mense than from intra septem menses or septem menses in grC.

4. Yosippon’s מליאוס, as Flusser noted, corresponds to the variant Mallius (vs Manlius). Flusser’s discovery of this reading in mss B and La and not in the 1524 Basel edition was one of the two textual bases for his argument that Yosippon’s Latin source belonged to the group of mss with the format AJ 1–16 + DEH. That the variant Mallius is a distinct marker for the group to which B, La, V, and Pi belong can now be confirmed, since it is found in 20 grC manuscripts and nowhere else. The omission of the next word Lucii in all grC manuscripts, Yosippon, and nowhere else provides futher evidence supporting this element of Flusser’s hypothesis. The appearance of Manlio Lucii in ms hr indicates that it is here using a source other than a grC manuscript.

5. Flusser suggested that the Biblical name עידו for the High Priest who met Alexander is chosen because it is close to the reading Iaddo in the group to which B and La belong. While it is true this is the universal reading in grC mss (vs Iaddi), Iaddo is also found in all grL manuscripts and ms w and is implied by Iaddonis in grP. The name עידו could also just as easily be derived from the variant Ieddo in Groups E and N and manuscripts B, G, and pg.

3.4.1 Unique Group C and Group G Variants

The most important variant in this passage found only in all 22 grC manuscripts collated is a large lacuna from 13.315c–320, which has already been discussed in connection with the precedence of ms B in the grC manuscript tradition. There it was shown that the insertion of material from the Latin War to fill in part of the lacuna that is found in all grC manuscripts probably derives from ms B, where the insertion appears to be an addition to the manuscript. In this case study, the other distinctive connections between grC and grG and their importance for reconstructing Yosippon’s LAJ source will be discussed with the help of English synopses comparing grG ms St and grC manuscript B and comparing the LAJ text with Yosippon. In the comparison below, aside from the lacuna, places where the grC and grG texts differ are indicated by italics. A passage where grG and grC agree against the rest of the tradition is indicated by bold italics, and readings only in grG are printed in bold. The commentary following the synopsis focuses on selected passages most relevant for understanding the grC and grG manuscript traditions and their relationship to the Latin text used by Sefer Yosippon. The complete Latin text of the passage in St with variants from B and El can be found in Appendix 3 (pp. 311–313).

View Full Size

View Full Size

View Full Size

• Download Figure
• Download Figure
• Download Figure
• Download figure as PowerPoint slide

3.4.2 Other Unique Group C Variants

13.313 (end)–314 (beg). Igitur uatem hoc perturbauit (314) Aristobolum autem (τὸν µὲν οὖν µάντιν τοῦτο διετάραξεν. (314) Ἀριστόβουλον δὲ). igitur] ita hr; uatem] autem grC.1 grC.2; per grC.3; autem per grC.4a Ptr hr; uatem hoc] Antigoni mors grC.4b (over erasure in Ne); perturbauit] turbauit grC.3–4 Ptr; autem] et eum grC hr

The text in this sentence is changed significantly by the grC archetype’s misreading of uatem as autem at the end of 13.313. This led to four distinct variations that correspond to grC subgroups:

314. Antigoni (Ἀντιγόνου)] Antigonus B C La Pt (corr. to Antigoni); Antigonis Pi O. in loco in quo maculae sanguinis adhuc Antigoni permanebant

εἰς τὸν αὐτὸν τόπον, οὗ σφαγέντος Ἀντιγόνου σπίλους ἔτι τοῦ αἵµατος ἐκείνου συνέβαινεν εἶναι

The clear error in B is reproduced in grC.2 manuscripts C, La, and Pt^a.c. (La and Pt both probably depending on C).¹⁵² The correction in grC.2 ms Pt might reflect an additional source in that manuscript, if it is not simply an obvious grammatical correction. The puzzling variant Antigonis in Pi is probably another attempt to correct the error in B.¹⁵³ The appearance of the same variant in O almost certainly depends on Ptr, in which the folio page where this passage occurs is replaced by a blank page. A connection between Ptr and grC.3 is found elsewhere, but, however that connection is to be explained (both Ptr and V, the earliest grC.3 manuscript, are dated to late 11th or early 12th century), it is an indication of a source, reflected in Ptr and grC.3–4, that introduces new elements into the grC manuscript tradition, sometimes (although not here) bringing in earlier readings to correct readings in Groups C.1 and C.2. This source might also account for the readings in grC.1 ms Vi (late 11th century) not found in B.

315. minabatur (ἐπετείνετο)] conabatur grC.1–2 grC.4b Vt l O hr; conabantur grC.3–4a (-Vt l O pat); conabobatur pat; mirabantur w

causamque requisisset tacentibus amplius minabatur/conabatur, discere uolens clamoris causam … Ut uero cogenti et interminanti; Ἀριστόβουλος τὴν αἰτίαν ἐπύθετο, καὶ µὴ λεγόντων ἔτι µᾶλλον ἐπετείνετο µαθεῖν… ὡς δὲ ἀπειλοῦντος καὶ βιαζοµένου).

While either minabatur or conabatur could be the earliest reading (conabantur is clearly secondary), the grC reading conabatur (exerted himself) seems a better fit for ἐπετείνετο (being intent on) than minabatur (threatened). With the reading conabatur, tacentibus would best be construed as an ablative absolute corresponding to the genitive absolute µὴ λεγόντων. Perhaps minabatur (found in all mss except w, which has mirabantur) is an attempt to simplify the text by taking tacentibus as the object of minabatur, influenced by the following interminanti.

321. ad faciem (εἰς ὄψιν)] ante faciem grC.3–4.

A clearly secondary reading in grC.3–4.

322. Antigonum (Ἀντίγονον)] Antiochum grC.1 (Vi corr. to Antigonum) grC.3 La; Antigonum over erasure C M Ne

With the readings in the erasures almost certainly correcting Antiochum, there remain only five grC manuscripts with the correct reading. Four of these are either direct copies or dependent on other mss, with the result that only grC.2 ms Pt would be a possible independent witness for the reading Antigonum, and this could easily be a scribe’s correction of an obvious error rather than evidence of Antigonum in the manuscript’s Vorlage.

322. dimisit (εἴασεν)] permisit grC.

genitum in Galilea nutriri demisit/permisit; γενόµενον εἴασεν ἐν τῇ Γαλιλαίᾳ τρέφεσθαι.

The grC reading permisit is closer to the Greek εἴασεν. The possibility that dimisit, found in all non-grC manuscripts except grG ms Werd (iussit), is the earliest reading cannot be excluded, since both dimitto and ἐάω can mean “let go.” See below for discussion of these readings in the context of Yosippon.

322. Hyrcano (Ὑρκανόν)] Hyrcani grC.

The grC reading Hyrcani is clearly secondary. All manuscript groups except for grC correspond to the Greek in having “he by no means lied to Hyrcanus” as the end of one sentence and “That one took up the rule after the death of Aristobulus” as the beginning of another: nequaquam mentitus est Hyrcano namque Hycani regnum post Aristobuli finem iste suscepit. (οὐ διεψεύσατο τὸν Ὑρκανόν. τὴν δὲ βασιλείαν µετὰ τὴν Ἀριστοβούλου τελευτὴν οὗτος παραλαβών). Group C manuscripts end one sentence with “he by no means lied” and begin another with “That one took up the rule of Hyrcanus after the death of Aristobulus.”

3.4.3 Unique Group G Variants

314. arbitror (οἶµαι)] grG; ut arbitror in all other manuscripts.

The earliest reading is unclear. For example, the LAJ text in grG ms St provides a number of different translations for parenthetical οἶµαι, for example, ut reor (6.59), reor (6.41), ut arbitror (6.63, 11.39), and arbitror (8.409).

317. meum [sanguinem] (τοὐµὸν [αἷµα])] grG pg; omitted in all other manuscripts (lacuna in grC).

grG clearly has the earliest reading, because “my blood” in grG corresponds to the Greek better than “the blood” in the rest of the manuscript tradition. It is impossible to know what reading would have been in the source of the grC archetype, which would not have had a lacuna.

fratribus et humiliorem multum (καὶ µετριότητα)] grG pg; omitted in all other groups (lacuna in grC).

grG clearly has the earliest reading, since et humiliorem corresponds to the Greek. There is no word corresponding to fratribus in the Greek, although it is clearly implied. It is impossible to know what reading would have been in the source of the grC archetype, which would not have had a lacuna.

3.4.4 Unique Variants Shared by Group C and Group G

314. sceleris (ὑπὸ τοῦ µύσους)] grG grC pg; celeri grE grH grJ grL grM grN Ba G w hr; omit grP

Groups C and G clearly have the earliest reading.

321. cum mox genitus fuisset odio patris despectus erat et usque ad mortem (συνέβη γεννηθέντι εὐθὺς µισηθῆναι ὑπὸ τοῦ πατρὸς καὶ µέχρι τῆς τελευτῆς αὐτοῦ)] grG grC pg; omit grE grH grJ grL grM grN grP Ba G w hr.

Here, most significantly, immediately after the lacuna, grC preserves an entire clause also found in grG, but nowhere else in the manuscript tradition. See below for the discussion of the connection of this phrase to the text in Yosippon.

322. suos filios] τῶν παίδων. grG (- Ml r) grC; filios suos grE grH grJ grL grN grM L Ba G w Ml r; filios grP pg

This transposition is significant only in so far as it provides another example of grC agreeing with grG against the rest of the manuscript tradition.

3.4.5 Groups C and G and Sefer Yosippon

3.4.5.1 The Lacuna in Group C and Sefer Yosippon

As can be seen from the synopsis below, Yosippon includes a number of details from the story that are missing in all grC manuscripts due to the lacuna. It is clear, therefore, that its sole LAJ source could not have been a grC manuscript.

View Full Size

View Full Size

• Download Figure
• Download Figure
• Download figure as PowerPoint slide

3.4.6 Summary and Conclusions for Case Study 4

AJ 13.313c–322 provides significant data both for the analysis of the grC and grG manuscript traditions and for the connection of those traditions to Yosippon.

1. A lacuna comprising five Niese sections (13.315c–320) with insertion of two texts from BJ 1.82–84, which replace the missing material, is a clear marker for identifying the 22 grC manuscripts collated.

2. Because Yosippon’s story includes a significant amount of material corresponding to the text missing in the lacuna, a grC manuscript cannot be its only source, although the possibility that Yosippon was reading the source of the grC archetype cannot be eliminated.

3. The unique grC reading permisit is much less likely to have been in Yosippon’s source than the reading dimisit, which is closer to the Hebrew וידיחהו.

4. Yosippon has Antigonus at 13.322 rather than Antiochus, found in almost all grC manuscripts, although this might have been an obvious correction made from the context.

5. This passage has three examples of readings shared only by all grG manuscripts: arbitror (vs ut arbitror), meum (not in any other manuscripts), and the phrase fratribus et humiliorem multum (not in any other manuscripts). The last two examples are clearly the earliest readings.

6. This passage has two examples of readings shared only by all grC and grG manuscripts: sceleris (vs celeris) and twelve consecutive words at the beginning of 13.320. Based on the Greek being translated, it is clear that both represent the earliest stage of the textual tradition.

7. Yosippon’s inclusion of some of the material in 13.320 found only in grC and grG indicates that at least one of Yosippon’s sources was a manuscript related to one of these two groups.