The Textual Nature of Shem-Tob's Hebrew Matthew

The Textual Nature of Shem-Tob's Hebrew MatthewAuthor(s): George HowardSource: Journal of Biblical Literature, Vol. 108, No. 2 (Summer, 1989), pp. 239-257Published by: The Society of Biblical LiteratureStable URL: http://www.jstor.org/stable/3267296 .

Accessed: 23/09/2013 06:54

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

The Society of Biblical Literature is collaborating with JSTOR to digitize, preserve and extend access toJournal of Biblical Literature.

http://www.jstor.org

This content downloaded from 150.108.161.71 on Mon, 23 Sep 2013 06:54:55 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=sbl

http://www.jstor.org/stable/3267296?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


JBL 108/2 (1989) 239-257

THE TEXTUAL NATURE OF SHEM-TOB'S HEBREW MATTHEW

GEORGE HOWARD University of Georgia, Athens, GA 30602

In 1986 I published an appraisal of the Hebrew Matthew issued by Jean du Tillet in 1555.1 Among other things I concluded the following: (1) the text of du Tillet (as well as the 1537 edition of a Hebrew Matthew published by Sebastian Miinster) is not an original Hebrew Matthew; (2) it is not a translation of the Latin Vulgate; and (3) it is not the same as the text of Matthew in Shem-Tob's Even Bohan. A relationship exists, however, between du Tillet's Hebrew Matthew and that of Shem-Tob. The closing words of the 1986

publication give the following description:

I now conclude with considerable finality that the Hebrew Matthew of du Tillet is a rather thorough revision of an earlier Hebrew Matthew reflected in a much less corrupted form in Shem-Tob. Du Tillet includes two basic types of revision: (1) stylistic modification and (2) revision designed to bring the Hebrew into closer harmony with the current Greek and Latin texts. Most stylistic modification consists of improvements in grammar and the substitution of synonymous words and phrases. Revisions designed to bring the Hebrew into closer harmony with the Greek and Latin were apparently for the purpose of establishing a common textual base for discussion and debate between Jews and Christians?

The present paper is about the Hebrew Matthew in Shem-Tob's Even Bohan. This treatise appeared in manuscript form in 1380 and was sub-

sequently expanded in one or more revised editions.4 The author, Shem-Tob

1 George Howard, "The Textual Nature of an Old Hebrew Version of Matthew,"JBL 105 (1986) 49-63.

2 In 1919 Alexander Marx had already determined the independence of the two texts. See Alexander Marx, "The Polemical Manuscripts in the Library of the Jewish Theological Seminary of America," Studies in Jewish Bibliography and Related Subjects in Memory of Abraham Solomon Freidus (1867-1923) (no ed.; New York: Alexander Kohut Memorial Foundation, 1929) 270-73. Cf. Pinchas Lapide, Hebrew in the Church (trans. E. F. Rhodes; Grand Rapids: Eerdmans, 1984) 55: "And yet with even the most superficial comparison of the two works the radical differences between their vocabulary, style, and diction would have demonstrated the impossibility of a common origin"

3 Howard, "Textual Nature" 63 n. 34. For a discussion of these later editions, see Marx, "Polemical Manuscripts," 247-78, esp.

265-70. See also W. Horbury, "The Revision of Shem Tob Ibn Shaprut's Eben Bohan," Sefarad 43 (1983) 221-37.

239



240 Journal of Biblical Literature

ben-Isaac ben-Shaprut (sometimes called Ibn Shaprut), was born in Tudela in Castile in the middle of the fourteenth century. His lengthy treatise contains the entire Gospel of Matthew in Hebrew with the text occasionally in-

terrupted by polemical comments, usually in the form of questions. My conclusion5 about this text is that it contains an old substratum that was

originally composed in Hebrew.6

The old substratum is not always present since the Gospel text apparently was subjected to repeated although sporadic revisions. As with the text of du Tillet, the revisions are basically of two kinds:

(1) stylistic modification and (2) revision designed to bring the Hebrew into closer harmony with the current Greek and Latin texts. Most stylistic modification consists of improvements in grammar and the substitution of synonymous words and phrases. Revisions designed to bring the Hebrew into closer

harmony with the Greek and Latin were apparently for the purpose of

establishing a common textual base for discussion and debate between Jews and Christians during the Middle Ages. An additional minor kind of alteration that appears is an occasional explanation of a name or place usually introduced by TY5(M), "in another language;" followed by a Hebrew trans- literation of a word or phrase in Greek, Latin, or some other language.

5 For my critical edition of the Hebrew text including an English translation and a detailed

analysis, see George Howard, The Gospel of Matthew According to a Primitive Hebrew Text

(Macon, GA: Mercer University Press, 1987). For preliminary analyses, see George Howard, "Shem-Tob's Hebrew Matthew,"' in Proceedings of the Ninth World Congress of Jewish Studies

(Jerusalem: World Union of Jewish Studies, 1986) 223-30; idem, "Was the Gospel of Matthew

Originally Written in Hebrew?" BibRev 2 (1986) 14-25. 6 A common misconception has been that the text is a translation of the Latin Vulgate. See,

e.g., Adolf Herbst, Des Schemtob ben Schaphrut hebraeische Obersetzung des Evangeliums Matthaei ... nach den Drucken des S. Miinster und J. du Tillet-Mercier (G6ttingen: Dieterichsche

Verlagsbuchhandlung, 1879) 16-29; Pinchas E. Lapide, "Der 'Priifstein' aus Spanien, Sefarad 34

(1974) 227-72. Among other things the comments of Shem-Tob show that he is a copyist, not

a translator of the Vulgate or any other text. For example, in his comment on Matt 21:5 (a quotation from Zech 9:9) he criticizes the use of 7111, "ass," rather than f111T1, "ass:' which appears in the MT. But since

'1•3M is as suitable a correspondent for the Greek 5vov as IfMn, why would

he not have rendered the Greek by the word llrnl?

If he is a translator, he has placed himself in the position of rendering the Greek by a word he does not like or think appropriate and then

of criticizing his own translation for not agreeing with the MT. Another example is Shem-Tob's comment on Matt 1:23 (a quotation from Isa 7:14). He has

no problem with the issue over "virgin" versus "young woman; since the Hebrew Matthew

already reads VM3?;1, "young woman." His criticism concerns the tenses of the verbs. The

Hebrew text reads: "Behold the young woman is conceiving (711•t)

and will bear ('Cfl) a son."

Shem-Tob comments: "It is not in Scripture, 'she is conceiving and will bear a son, expressing the future tense, but, 'she is conceiving and is bearing ( nf'f) a

son, expressing the present

tense"' The criticism is of the imperfect Ifl instead of the participle fl~t which appears in the

MT. In the Greek the verbs are both in the future, et and rgtierTt. If Shem-Tob is a translator, then, his comment places him in the awkward position of rendering one future with a participle, another future with a future, and then of criticizing himself for doing it. If he could have translated qet by

•n1, he could have translated -itertL by 'l15~'. The only reasonable conclu-

sion is that Shem-Tob is not translating at all but criticizing an already existing Hebrew text for

failing to correspond to the MT of Isa 7:14.



Howard: Shem-Tob's Hebrew Matthew 241

Since the present text is spotted by scribal modification, one must exca- vate it, so to speak, in order to retrieve the old substratum. One way of doing this is to compare manuscripts of Shem-Tob with the Greek and Latin texts of Matthew. Generally, those readings which are farthest from the Greek and Latin and are less polished in style are candidates for belonging to the oldest layer of the text. Those that are closest to the Greek and Latin and are more polished in style, especially when portraying a later Jewish scribal hand

(usually in terms of diction or grammar), should be considered as later revisions. Occasionally the comments of Shem-Tob help to establish an unrevised Gospel text. The scribes who altered the Gospel text either stylistically or substantively sometimes neglected to reproduce the same revision in a later quotation or allusion to the text in Shem-Tob's comments7 In some instances, therefore, his comments provide evidence for the readings of the oldest layer of the biblical text. Thus, the comments support the reading of "Father" in Matt 12:32, the absence of "Messiah" in 16:16, the reading of 10I in 21:5, and the reading of 1115M in 26:41.

There are some exceptions to the general rule that the readings that are closest to the Greek should be considered later revisions. The Hebrew Matthew appears to contain some Grecisms that were taken over from Greek sources used by the author in his composition of the Hebrew Gospel. These are to be distinguished, whenever possible, from later revisions toward the canonical Greek text. Some elaboration on this point is perhaps necessary.

For centuries scholars have argued for a Semitic background to Matthew as well as much of the Gospel tradition as a whole. Early in the second century Papias (ca. 60-130 CE), bishop of Hierapolis, wrote that "Matthew collected the oracles in the Hebrew language, and each interpreted them as best he could" (Eusebius, Hist. eccl. 3.39.16). Whether this statement refers to our Gospel of Matthew is hard to prove.8 Whatever the case, the early church writings after the time of Papias are replete with references to an original Hebrew Matthew. Among the many who refer to it are Irenaeus (Adv. haer. 3.1.1), Origen (quoted by Eusebius Hist. eccl. 6.25.4), Eusebius (Hist. eccl. 3.24.6), Epiphanius (Panarion 30.13.1-30, 22.4) and Jerome (Epist. 20.5).

Since the time of J. A. Widmanstadt in the mid-sixteenth century it has been commonplace to suppose that by "Hebrew" Papias meant that Matthew

7 It was not uncommon for medieval scribes to revise a lemma citation of a biblical reading in ancient documents without troubling themselves to revise, in a corresponding way, sub- sequent comments that repeated the quotation or, at least, alluded to it. A good example of this is to be found in Mss UF and sometimes L of Philo. See Peter Katz, Philo's Bible: The Aberrant Text of the Bible Quotations in Some Philonic Writings and Its Place in the Textual History of the Greek Bible (Cambridge: University Press, 1950). For some corrective to Katz, see George Howard, "The 'Aberrant' Text of Philo's Quotations Reconsidered," HUCA 44 (1973) 197-209.

8 W G. Kiimmel thinks it refers to our Matthew; he also thinks that Papias had never seen a Semitic Matthew and that his statement is false; see Introduction to the New Testament (17th ed.; Nashville: Abingdon, 1975) 120-21. Josef Kiirzinger's attempt to prove that Papias was




wrote in Aramaic? This supposition is due inter alia to the long-standing belief that Hebrew in the days of Jesus was no longer in use in Palestine as a vernacular but had been replaced by Aramaic. From the end of the nine- teenth through the twentieth century an Aramaic background to the Gospel tradition has been investigated and supported by many. Notable among the many are Wellhausen, Dalman, Burney, Torrey, Montgomery, Burrows, Taylor, Black, and Zimmermann.'o

Since the discovery of the Dead Sea Scrolls, many of which are Hebrew compositions, and the discovery of other documents from the Judean Desert, it is now known that Hebrew, and even Greek, was spoken along with

speaking only of Matthew's style of writing, not the language in which he wrote, is not totally convincing; see "Das Papiaszeugnis und die Erstgestalt des Matthiusevangeliums," BZ 4 (1960) 19-38; "Irenius und sein Zeugnis zur Sprache des Matthiiusevangeliums," NTS 10 (1963-64) 108-15; "Die Aussage des Papias von Hierapolis zur literarischen Form des

Markusevangeliums,' BZ 21 (1977) 245-64; "Papias von Hierapolis: Zu Titel und Art seines Werkes:' BZ 23 (1979) 172-86; Papias von Hierapolis und die Evangelien des Neuen Testaments (Regensburg: Pustet,

1983). Kiirzinger's argument, based on the assumption that Papias was using ancient rhetorical

terminology, may ascribe more erudition to Papias than he had. Eusebius said that it is obvious that Papias was a man of "very little intelligence" (Hist. eccl. 3.39.13). It is also possible that

Papias is quoting the "Presbyter" in regard to Matthew as he is in regard to Mark. The rhetorical abilities of the Presbyter are, of course, unknown to all.

9 Johann Albert Widmanstadt, Liber Sacrosancti Evangelii De Jesu Christo Domino & Deo Nostro ... characteribus & lingua Syra, Jesu Christo vernacula, Divino ipsius ore consecrata & a

Joh. Evangelista Hebraica dicta, Scriptorio Prelo diligenter Expressa (Vienna: M. Cymbermann, 1555). This reference was taken from Jean Carmignac, "Hebrew Translation of the Lord's Prayer: An Historical Survey" in Biblical and Near Eastern Studies: Essays in Honor of William Sanford LaSor (ed. Gary A. Tuttle; Grand Rapids: Eerdmans, 1978) 71 n. 5.

10 Julius Wellhausen, Einleitung in die drei ersten Evangelien (Berlin: Georg Reimer, 1905; 2d ed. used here, 1911).

Gustaf Dalman, The Words of lesus (trans. D. M. Kay; Edinburgh: T. & T. Clark, 1902) [I have had access to a 1909 edition].

C. F Burney, The Aramaic Origin of the Fourth Gospel (Oxford: Clarendon, 1922); see also

Burney, The Poetry of Our Lord (Oxford: Clarendon, 1925). C. C. Torrey, "The Translations Made from the Original Aramaic Gospels," in Studies in the

History of Religions Presented to Crawford Howell Toy (ed. D. G. Lyon and G. F Moore; New York: Macmillan, 1912) 269-317; The Composition and Date of the Acts (Cambridge, MA: Harvard

University Press, 1916); "Fact and Fancy in the Theories Concerning Acts," AJT 23 (1919) 61-86, 189-212; "The Aramaic Origin of the Gospel of John," HTR 16 (1923) 305-44; The Four Gospels: A New Translation (New York: Harper, 1933); Our Translated Gospels: Some of the Evidence (New York: Harper, 1936); "The Aramaic of the Gospels:' JBL 61 (1942) 71-85; Documents of the Primitive Church (New York: Harper, 1941).

James Montgomery, The Origin of the Gospel According to St. John (Philadelphia: Winston,

1923); "Torrey's Aramaic Gospels:' JBL 53 (1934) 79-99. Millar Burrows, "The Original Language of the Gospel of John," JBL 49 (1930) 95-139;

"Principles for Testing the Translation Hypothesis in the Gospels,' JBL 53 (1934) 13-30. W. R. Taylor, "Aramaic Gospels- Source and Form

Criticism,' ExpTim 49 (1934-38) 55-59. Matthew Black, An Aramaic Approach to the Gospels and Acts (3d ed.; Oxford: Clarendon,

1967). Frank Zimmermann, The Aramaic Origin of the Four Gospels (New York: Ktav, 1979).




Aramaic in first-century Palestine. I This is not to say that Hebrew and Greek

were spoken with the same widespread proficiency as Aramaic; nevertheless, they were known and spoken.2 A result of these discoveries is an increase in the belief that Jesus may have spoken Hebrew and Greek in addition to Aramaic and that Hebrew and Greek belong in the background of at least

parts of the Gospel tradition. Some who have supported a Hebrew back-

ground are Birkeland, Grintz, Emerton, Carmignac, Lapide, and Chomsky.'3 It should also be remembered that M. H. Segal argued long ago that Hebrew

along with Aramaic was spoken in the first century.14

The use of Greek in

first-century Palestine is now also well established.5 Given the multilinguistic milieu of first-century Palestine, it is reason-

able to assume that the Gospel tradition circulated originally partly in Aramaic, partly in Hebrew, and partly in Greek. No doubt one composing a Greek narrative about Jesus used Aramaic, Hebrew, and Greek sources; one

composing an Aramaic narrative used Aramaic, Hebrew, and Greek sources; and one who composed a Hebrew narrative used Aramaic, Hebrew, and Greek sources. The canonical Gospels present us with compositions on the Greek level only. The work of scholars has consequently produced numerous

examples of Aramaic and Hebrew sources reflected by the Greek texts. In a similar way, it is interesting to note that Shem-Tob's Hebrew

Matthew at times reflects Greek idiom, as the canonical Greek sometimes reflects Hebrew and Aramaic idiom. There is little doubt that many examples of Grecisms are the result of sporadic revision toward the canonical text which took place over the centuries through the process of scribal transmission. It is possible, however, that some Greek idioms in the Hebrew Matthew are reflections of primitive Greek sources used in the original composition of the Hebrew Gospel. Although it is not always easy to distinguish original

11 J. M. Sevenster, Do You Know Greek? (Leiden: Brill, 1968); G. Howard and J. C. Shelton, "The Bar-Kokhba Letters and Palestinian Greek,' IEJ 23 (1973) 101-2.

12 For an up-to-date discussion of the languages of Palestine, see especially Joseph A. Fitz-

myer, "The Languages of Palestine in the First Century A.D.,' in A Wandering Aramean:

Collected Aramaic Essays (Missoula, MT: Scholars Press, 1979) 29-56. 13 Harris Birkeland, The Language of Jesus (Oslo: I. Kommisjon Hos Jacob Dybwad, 1954);

J. M. Grintz, "Hebrew as the Spoken and Written Language in the Last Days of the Second

Temple" JBL 79 (1960) 32-47; J. A. Emerton, "Did Jesus Speak Hebrew?"JTS 12 (1961) 189-202; idem, "The Problem of Vernacular Hebrew in the First Century A.D. and the Language of

Jesus,"' JTS 24 (1973) 1-23; Jean Carmignac, "Studies in the Hebrew Background of the Synoptic Gospels:' ASTI 7 (1970) 64-93; Pinchas Lapide, "Insights from Qumran into the Language of

Jesus,"' RevQ 32 (1975) 483-501; W. Chomsky, "What Was the Jewish Vernacular During the Second Commonwealth?" JQR 42 (1951-52) 193-212. See further James Barr, "Which Language Did Jesus Speak?- Some Remarks of a Semitist,' BJRL 53 (1970) 9-29.

14 M. H. Segal, A Grammar of Mishnaic Hebrew (Oxford: Clarendon, 1927) 1-18. 15 Sevenster, Do You Know Greek?; Fitzmyer, "Languages of Palestine," 32-38. Cf. Morton

Smith's remarks: "Therefore the movement was not from Aramaic to Greek, but both languages were represented in both the primitive and the secondary stages of the religious developments"; see Smith, "Aramaic Studies and the Study of the New Testament,' JBL 26 (1958) 309.




Grecisms from later revisions toward the canonical Greek Matthew, the two should be viewed differently.

In spite of numerous revisions and some possible Grecisms in the Hebrew text, there is evidence that points to the antiquity of the old substratum of Shem-Tob's Matthew and that suggests that the old substratum is an original Hebrew composition. A brief summary of this evidence, as stated in the critical edition mentioned above (see n. 5), is placed here.

(1) The Shem-Tob text has an occasional unique or unusual linguistic or textual agreement with Aramaic and Hebrew quotations of or allusions to Matthew in medieval Jewish works.'6 These include the Talmud (b. Sabb. 116b, which in some way is related to Matt 5:17-18a),17 the Tol'doth Yeshu

(sixth to tenth centuries),'8 the Book of Nestor Hakomer (between the sixth and the ninth century),'9 the Milhamot HaShem by Jacob ben Reuben

(1170),20 Sepher Joseph Hamekane by Rabbi Joseph ben Nathan Official

(thirteenth century),21 and the Nizzahon Vetus (latter part of the thirteenth

century).22 The agreements between these early writings and Shem-Tob's Hebrew Matthew demonstrate that a Shem-Tob type Matthean text was in existence long before the time of Shem-Tob.

16 It is not the case that every Hebrew quotation of Matthew in medieval Jewish documents shows a relationship to the Shem-Tob Matthean tradition. In many instances the quotations appear to be ad hoc translations of the Greek or Latin text freshly prepared for the occasion. It is in fact the tendency to translate directly from the Greek or Latin that eventually con- taminated the primitive Hebrew Matthean tradition when it too was subjected to the influence of the canonical text through the process of revision.

17 The Gemara reading may not be a direct quotation of Matthew (it is said to come at the end of the book from which it is quoted). For the quotation and a discussion, see R. Travers

Herford, Christianity in Talmud and Midrash (Clifton: Reference Book Pub., 1966) 146-51. 18 Samuel Krauss dates it to ca. 500 CE; see Das Leben Jesu nachjiidischen Quellen (Berlin:

Calvary, 1902). Joseph Klausner dates it to the tenth century (Jesus of Nazareth [trans. H. Danby; New York: Macmillan, 1946] 53). For various proposals for the date, see Giinter Schlichting, Ein

jiidisches Leben Jesu (Tiibingen: Mohr-Siebeck, 1982) 2. For an assessment of the relationship between the Tol'doth Yeshu and Shem-Tob's Hebrew Matthew, see George Howard, "A Primitive Hebrew Gospel of Matthew and the Tol'doth Yeshu," NTS 34 (1988) 60-70.

19 This is according to Lapide, Hebrew in the Church, 23. The text may be found in J. D.

Eisenstein, D•'1n'1 I• (Jerusalem, 1969) 310-15. The editor there dates it in the ninth century (p. 310).

20 Judah Rosenthal, UD71 T1"'1 .=1N0 7 Mp3' (Jerusalem: Mosad Kok, 1963) viii; see also

Rosenthal, "•

7= = -1V5 ' V-51Y 71=t7l 5V • • Y1r," Tarbiz 32 (1962) 48-66.

21 Judah Rosenthal, R~3n 1 ' V iE)DD (Jerusalem, 1970) 17. Ms Or. #53 of the Biblioteca Nationale Centrale in Rome includes material quite close to the Paris manuscript of Sepher Joseph Hamekane. See E. E. Urbach, "Itudes sur la litterature polkmique au moyenage, Revue des 6tudes juives 100 (1935) 49-77. Judah Rosenthal published the material on the Gospels in

MS Or. Rome #53 in ""'1 Tn D 7 0 t071 n• n1 1 ' 5Vt, "1 l'1p," in Studies in Jewish

Bibliography: History and Literature in Honor of I. Edward Kiev (ed. Charles Berlin; New York:

Ktav, 1971) 123-39. 22 David Berger, The Jewish-Christian Debate in the High Middle Ages (Philadelphia: Jewish

Publication Society of America, 1979) 33.




(2) Although the manuscript tradition of Shem-Tob's Matthew shows many revisions, its old substratum displays numerous literary and composi- tional elements such as puns, word connections, and alliteration. Often its variations are unlike any other known readings and at times help form theological patterns running throughout the Gospel. At other times it has affinitives with known textual traditions.

In the remainder of this paper a brief profile of the textual affinitives of Shem-Tob's text with known textual traditions will be given. These affinitives go far in proving the antiquity of the Hebrew text since it is difficult to believe that all the text-types reflected in Shem-Tob were readily available to Shem- Tob in the fourteenth century, if indeed he prepared the Hebrew text afresh. On the contrary, the following profile demonstrates that the Hebrew Matthew of Shem-Tob sometimes overlaps with readings and documents that vanished in ancient times only to be rediscovered in recent years. The profile thus suggests that this Hebrew Matthew is a primitive text that predates Shem-Tob by many centuries.23

23 The following editions and apparatuses have been used in the profile: Constantinus Tischendorf, Novum Testamentum Graece (8th ed.; Leipzig: Giesecke & Devrient, 1869) vol. 1; Kurt Aland, Synopsis Quattuor Evangeliorum (9th ed.; Stuttgart: Wiirttembergische Bibel- anstalt, 1976); Albert Huck and Heinrich Greeven, Synopsis of the First Three Gospels (13th ed.;

Ttibingen: Mohr-Siebeck, 1981); S. C. E. Legg, Novum Testamentum Graece: Evangelium Secundum Marcum, and Evangelium Secundum Matthaeum (Oxford: Clarendon, 1935, 1940); The New Testament in Greek. The Gospel According to St. Luke. Edited by the American and British Committees of the International Greek New Testament Project. Part One. Chapters 1-12 (Oxford: Clarendon, 1984). For the Old Syriac I have used F. C. Burkitt, Evangelion da- Mepharreshe (Cambridge: University Press, 1904) and Agnes Smith Lewis, The Old Syriac Gospels or Evangelion da-Mepharreshe (London, 1910). For Tatianic readings I have used, in addition to the critical apparatuses listed above, Louis Leloir, Saint Ephrem: Commentaire de l'Evangile Concordant: Texte Syriaque (Manuscript Chester Beatty 709) (Dublin: Hodges Figgis, 1963); Saint Ephrem, Commentaire de l'Evangile Concordant, Version Armnnienne (CSCO 137, Scriptores Armeniaci I; Louvain: Louvain University Press, 1953; Latin trans. 145, Scriptores Armeniaci 2; Louvain: Louvain University Press, 1964); Hamlyn Hill, A Dissertation on the Gospel Commentary of S. Ephraem the Syrian (Edinburgh: T. & T Clark, 1896). For the Arabic Diatessaron I have used the translation by Hope W Hogg in The Ante-Nicene Fathers (ed. Allan Manzies; reprint, Grand Rapids: Eerdmans, 1965) vol. 10. I have cautiously consulted the vast collection of readings in Ignatius Ortiz de Urbina, Vetus Evangelium Syrorum, et exinde excerptum Diatessaron Tatiani (Biblia Polyglotta Matritensia. Series VI; Madrid, 1967). For the Old Latin I have used Adolf Jiilicher, Itala: Das Neue Testament in Altlateinischer Uberlieferung (revised by Walter Matzkow and Kurt Aland; Berlin and New York: de Gruyter) vol. 1, Matthiius-Evangelium (1972); vol. 2, Marcus-Evangelium (1970); vol. 3, Lucas-Evangelium (1976). For the Pseudo- Clementine literature I have used the following: Bernhard Rehm, Die Pseudoklementinen I Homilien (Berlin: Akademie-Verlag, 1969); Die Pseudoklementinen II Rekognitionen in Rufins Obersetzung (Berlin: Akademie-Verlag, 1965); Wilhelm Frankenberg, Die Syrischen Clementinen mit Griechischem Paralleltext (Leipzig: Hinrichs, 1937). For all other texts I have relied on the standard translations and critical apparatuses to the Greek New Testament.




I. Shem-Tob, The Old Syriac, and The Old Latin Traditions

By far the most prominent agreements of Shem-Tob's text are with the Old Syriac, the Old Latin, and combinations of them, sometimes accompanied by other witnesses that are usually considered to belong to the "Western" tradition.

Shem-Tob and the Old Syriac

The following is a sampling of instances where Shem-Tob's Hebrew Matthew agrees with the Old Syriac, either Sys and/or Syc (occasionally accompanied by SyP or other Syriac witnesses) against all other Matthean witnesses.

2:19 + "king" Shem-Tob Sys.c.p 3:11 + "behold" Shem-Tob SyC 4:18 + "shore" Shem-Tob Sys 5:12 "who were before you" omit: Shem-Tob Sys 8:9 "I have authority" Shem-Tob Sys.c 10:6 "strayed" Shem-Tob Sys.p 12:5 "on the sabbath" omit: Shem-Tob Syc.p 12:31 "sons of men" Shem-Tob Sys.c.p 14:26 "demon" Shem-Tob Sys.c 15:2 "ordinances" Shem-Tob Sys.c 15:3 "ordinances" Shem-Tob Sys.c 16:17 "and having answered" omit: Shem-Tob SyC 22:35 "scribe" Shem-Tob SyC 26:10 "good and wonderful deed" Shem-Tob Sys 27:51 "immediately" Shem-Tob Sys.p.h

There are many other instances not listed here where Shem-Tob agrees with

Sys/c accompanied by a few other witnesses such as the Aeth, Arm, or some other text or version.

Shem-Tob and The Old Latin

The following is a sampling of readings where Shem-Tob's Hebrew Matthew agrees with Old Latin (= OL) witnesses occasionally accompanied by the Vulgate (= vg) against all other Matthean witnesses.

1:12 "and after the Babylonian Exile" omit: Shem-Tob k 8:30 "near by" Shem-Tob; "not far" a b c ff' g' h 1 vg 9:11 "sit" Shem-Tob k 9:35 "towers" Shem-Tob OL vg




10:39 "love" Shem-Tob fPl 15:36 "people" Shem-Tob a aur b c fl g' 1 q vg 16:21 "and scribes" omit: Shem-Tob a r2 16:25 "for the life of the world to come" Shem-Tob; "for eternal

life" g2 17:21 + "demon" Shem-Tob [[a]] b c n

Shem-Tob, The Old Syriac, and The Old Latin

In addition to the above readings the Hebrew Matthew of Shem-Tob sometimes agrees with a combination of witnesses to the Old Syriac and the Old Latin, sometimes accompanied by Codex Bezae and an assortment of other witnesses. The following is a brief sampling.

4:22 "nets" Shem-Tob 126 SyC OL(mss) vg Copsa 5:47 whole vs omit: Shem-Tob Sys k 8:2 "certain (leper)" Shem-Tob Syc.p OL(mss) vg(4mss) Hil 9:22 "Jesus" omit: Shem-Tob R* D 1555 OL(mss) Sys 11:2 "Jesus" Shem-Tob D d min SyC Aeth 18:2 "one" Shem-Tob D d e Sys.C Arm 21:33 "in its midst" Shem-Tob vg(1 ms); "in it" SyS.c.p c h Copbo

The following three readings deserve attention in the same context

although they fail to agree with the Old Syriac. (In 25:28 neither Sys or Syc is extant.)

16:4 "and adulterous" omit: Shem-Tob D OL(mss) 18:9 "of fire" omit: Shem-Tob D d 1675 25:28 "five" Shem-Tob D d

Shem-Tob and The Coptic Gospel of Thomas

In a tabular study of 1960 Tjitze Baarda demonstrated a connection between Thomas and the Syro-Latin tradition, including especially the Diatessaron of Tatian.04 He listed numerous agreements with Codex Bezae, Marcion, the Old Latin manuscripts of the Gospels (especially a, b, and e) and the Old Syriac tradition. He located more than seventy agreements between Thomas and Sys with scarcely fewer than this with SyC and SyP. He also located some 130 instances where Thomas has a variant in agreement

24 Tjitze Baarda, "Thomas en Tatianus," in R. Schippers and T Baarda, Het Evangelie van Thomas (Kampen: Kok, 1960) 135-55. I have made use of the English version published in T. Baarda, Early Transmission of Words of Jesus: Thomas, Tatian and the Text of the New Testament

(ed. J. Helderman and S. J. Noorda; Amsterdam: VU Boekhandel, 1983) 37-49.




with one or more recensions of the Diatessaron. A conclusion which Baarda drew from his tabulation is: "All indications are that the Gospel of Thomas must be closely connected to a certain type of Western text in Syria.25

In a final column of the above tabulation, called "Varia,'

Baarda listed an occasional agreement between Thomas and other textual traditions including five agreements with Mt(H), which he called a "well-known Hebrew text of

Matthew"'26 This "Mt(H)" is in fact the Hebrew Matthew published in 1555

by du Tillet and discussed in my previous article." When Shem-Tob's Hebrew

Matthew is inserted into Baarda's chart as a replacement for the text of du Tillet, the agreements between Thomas and the Hebrew Matthew (usually accompanied by various witnesses to the Diatessaron) rise significantly. In- stead of five agreements listed with du Tillet's text, Thomas agrees with Shem-Tob twenty-two times.28 This is considerably higher than Thomas's

agreements with Codex Bezae (15) and Marcion (17), as noted by Baarda, and is close to half as many agreements as Thomas has with the Old Latin (48). The agreements between Thomas and Shem-Tob's Hebrew Matthew are the

following (the order of the list is that of Baarda):

19:30 "and the last first" omit: Shem-Tob Thomas 4 13:47-48 "fisherman/fishermen" Shem-Tob Thomas 8 13:48 "chose" Shem-Tob Thomas 8 7:3 "see2" Shem-Tob Thomas 26 5:14 "being built on" Shem-Tob Thomas 32 5:15 "in a hidden place" Shem-Tob Thomas 33 5:15 "he puts/they put" Shem-Tob Thomas 33 23:13 "wish" Shem-Tob Thomas 39 12:32 "the Father" Shem-Tob (comment only following 12:37)

Thomas 44 12:35 "which is in his heart" Shem-Tob Thomas 45 6:24 "honor" Shem-Tob Thomas 47 9:17 "lest" Shem-Tob Thomas 47 9:16 (explicit contrast between "old" and "new") Shem-Tob

Thomas 47 17:20 "from here" omit: Shem-Tob Thomas 48 17:20 "to there" omit: Shem-Tob Thomas 48 13:24 "is like" Shem-Tob Thomas 57 13:29 "to them" Shem-Tob Thomas 57

25 Ibid., 48. 26 Ibid., 39. 27 See n. 1 above. 28 Within the limitations of Baarda's chart, Shem-Tob agrees with some form of the

Diatessaron twenty-one times; Matt 9:38 =Thomas 73, omit: "therefore"' fails to agree with Tatian.




9:38 "therefore" omit: Shem-Tob Thomas 73 6:20 "devour" Shem-Tob Thomas 76 8:20 "of the air" omit: Shem-Tob Thomas 86 8:20 "have•" Shem-Tob Thomas 86 8:20 "his (head)" Shem-Tob Thomas 86

II. Shem-Tob and "Harmonistic" Readings

One of the most characteristic features of Shem-Tob's Hebrew Matthew is its many "harmonistic" readings. "Harmonistic" is placed within quotation marks because there is some uncertainty whether all the readings that belong to this category are truly harmonistic. By "harmonistic" one usually means a variant reading in one Gospel that corresponds partially or wholly to the text of another Gospel because of scribal alteration. It was common for scribes to obliterate differences in Gospel parallels in order to make them read more alike. Hort, in his description of the "Western" text, described it as characterized by harmonization.

But its most dangerous work is "harmonistic" corruption, that is, the partial or total obliteration of differences in passages otherwise more or less resembling each other. Sometimes the assimilation is between single sentences that happen to have some matter in common, more usually however between parallel passages of greater length, such especially as have in some sense a common origin.9

An interesting point with this description is that Hort also described readings from the "Western" text as the earliest that can be fixed chronologically?3

It is well known that scribes had a tendency to harmonize parallel passages. It is also a fact that early literati had a tendency to make the Gospel tradition diverse. The four canonical Gospels testify to this fact. In dealing with ancient textual traditions one cannot always tell whether diverse parallels are more primitive than close parallels. In some instances close parallels may represent earlier forms of the tradition than diverse parallels. Redaction criticism, for example, operates within the presupposition that the early Gospel tradition diversified as it was crystallized into written form. The fact that readings from Hort's "Western" text are the earliest he could fix chronologically suggests the possibility that some of his "Western harmonizations" may be original and the more divergent readings the less primitive.

It has been argued that a great supplier of harmonistic readings was Tatian's Diatessaron, which dates to the latter half of the second century.

29 B. F Westcott and F. J. A. Hort, The New Testament in the Original Greek: Introduction and Appendix (New York: Harper & Bros., 1882) 124-25. H. J. Vogels investigated the harmonistic element in Codex Bezae and found it so harmonistic that he concluded a harmony stood behind it (Die Harmonistik vom Evangelientext des Codex Cantabrigiensis [TU 36; Leipzig: Hinrichs, 1910]).

30 Westcott and Hort, The New Testament in the Original Greek, 120.




Tatian created this Gospel harmony by weaving together primarily the text of the four canonical Gospels. The language in which it was composed was

Syriac, Greek, or Latin, although the former is probably to be preferred?•

Whatever the case, Tatian's harmonized Gospel apparently had an effect on the transmission of the canonical Gospels in both their Greek and versional forms. It is doubtful, however, that all early nondivergent readings are due to the influence of Tatian. Several studies in recent times, notably those by Helmut K6ster and A. J. Bellinzoni, have shown that a "harmonistic" ten-

dency existed long before Tatian wrote his Diatessaron?2 In 1980 I pointed out a number of "harmonistic" readings in the Old Syriac Gospels which are

unsupported by the Diatessaron and agree with pre-Tatianic authors?3 From these studies one may conclude that early in the second century there existed nondivergent Gospel parallels that predate all known existing witnesses to the same parallels in a divergent form. In my judgment the evidence for the originality of some nondivergent parallels (= "harmoniza-

tion") is too early and weighty to be ruled out altogether. If this is true, it is

possible that some "harmonies," even in Tatian's Diatessaron, may have come from the earliest layer of tradition and do not represent the work of Tatian himself.

It is within this context that readings in Shem-Tob's Matthew which

agree with the wording of the other Gospels should be considered. They should not be considered a priori the result of scribal harmonization.

Shem-Tob and the Gospel of John

In the following list of readings Shem-Tob's Matthew agrees with the Gospel of John. Unless otherwise indicated, each reading is in disagreement with the

remaining Matthean tradition including the Greek Sys.c.p OL vg the synoptic parallels and available evidence for the Syriac Diatessaron.

8:6 (pars Luke 7:2 John 4:46). Greek = "servant"; Shem-Tob

John 4:46 ="son."

31 On Syriac, see A. Baumstark, "Das griechische Diatessaronfragment von Dura Europos,' OrChr 32 (1935) 250; Arthur Voobus, Studies in the History of the Gospel Text in Syriac (Louvain:

Durbecq, 1951) 12. On Greek, see Carl H. Kraeling, A Greek Fragment of Tatian's Diatessaron

from Dura (London: Christophers, 1935) 15-18; Adolf von Harnack, "Tatian's Diatessaron und

Marcion's Commentar zum Evangelium bei Ephraem Syrus,' ZKG 4 (1981) 494-95. On Latin,

see F C. Burkitt, "The Dura Fragment of Tatian;' JTS 36 (1935) 257-58.

32 See Helmut K6ster, Synoptische Oberlieferung bei den apostolischen Viitern (Berlin: Akademie-Verlag, 1957); A. J. Bellinzoni, The Sayings of Jesus in the Writings of Justin Martyr

(Leiden: Brill, 1967). Cf. Leslie L. Kline, The Sayings of Jesus in the Pseudo-Clementine Homilies

(Missoula, MT: Scholars Press, 1975). 33 George Howard, "Harmonistic Readings in the Old Syriac Gospels," HTR 73 (1980) 463-71.




9:6 (pars Mark 2:11 Luke 5:24 John 5:8). Greek = "and go into

your house"; Shem-Tob John 5:8 = "and walk:' [Mark 2:11 Lect. 184 reads: "and walk" instead of "and go into your house:' Luke 5:24 MS 903 reads: "and walk" after "arise.']

10:39 (pars Luke 17:33 John 12:25). Greek = "He who finds"; Shem-Tob OL (fP) John 12:25 = "He who loves."

13:55 (pars Mark 6:3 Luke 4:22 John 6:42). Greek = "Is not his mother called Mary?" Shem-Tob "Do you not know all these; his mother Mary?" John 6:42 "Do we not know his father and mother."

14:21 (pars Mark 6:44 Luke 9:14 John 6:10). Greek = "And those

eating"; Shem-Tob OL (abcfflg' hlq aur) vg "And the number of those eating"; John 6:10 "The men, therefore, reclined, in

number." 16:25 (pars Mark 8:35 Luke 9:24 John 12:25). Greek = "whoever

loses his life for my sake will find it"; Shem-Tob "he who loses his life in this world for my sake will save his soul

for the life of the world to come"; John 12:25 "he who hates his life in this world will keep it for eternal life." [Matt OL (g2) and Luke 9:24 OL (b) add at the end: "for eternal life']

21:12 (pars Mark 11:15 Luke 19:45 John 2:14). Greek = "and cast out all those who sold and bought"; Shem-Tob "and he found there those who buy and sell"; John 12:25 "and he

found in the temple those who sold"34

Shem-Tob, Mark, and Luke

There are many more readings in Shem-Tob's Matthew that agree with Mark and Luke than with John. This is to be expected since Matthew parallels Mark and Luke more often than John. Some of Matthew's agreements with Mark and Luke are interesting. In order to describe Shem- Tob's text vis-A-vis Mark and Luke I have compiled the following statistics. They consider only plus and minus readings.

Plus Readings I have isolated 159 plus readings in Shem-Tob's Matthew. These are

readings in the Hebrew text that are longer than the Greek text of Matthew including all its manuscripts and versions. Of the 159 readings 10 percent agree exclusively with one or more witnesses to Luke; 3.1 percent agree

34 The Arabic Diatessaron reads both "found" and "drove out,"' a combination of the two forms. The difference perhaps goes back to the similarly looking 0N221, "and he found,' and R'I (hiphil), "and he cast out:'




exclusively with one or more witnesses to Mark. Thus, in regard to plus readings Shem-Tob agrees exclusively more often with Luke than with Mark. An interesting point is that 81 percent (13 out of a total of 16 instances) of Shem-Tob's plus agreements with Luke occur in the triple tradition and, therefore, agree with Luke against Mark. (Only 3 occur in the Q sections.)35 This raises somewhat the total number of so-called "minor agreements" between Matthew and Luke against Mark06 These statistics are perhaps tempered, however, by the fact that all but 4 (namely, Matt 9:24; 13:19; 13:21; 17:1) agree with the Diatessaron. Since these may be significant readings, I list here the 13 instances of Shem-Tob = Luke # Mark.

8:32 + "from the men/man" Shem-Tob Luke 8:33 9:24 + "stop weeping" Shem-Tob Luke 8:52 10:1 + "every (unclean spirit)" Shem-Tob Luke 9:1 13:19 + "The sower is the Son of Man" Shem-Tob Luke 8:11

Lvt (ri) 13:21 + "little (trouble)" Shem-Tob Luke 8:13 Ms 1654 17:1 + "to pray" Shem-Tob Luke 9:28 17:3 + "and they told Jesus all which would happen to him in

Jerusalem. Peter and his companions were fully asleep, neither drowsing or fully awake; [then] they saw his body and the two men with him." Shem-Tob -+ Luke 9:31-32

17:4 + "When they went away" Shem-Tob Luke 9:33 17:5 + "they were greatly alarmed; while they were under the

cloud" Shem-Tob -+ Luke 9:34 17:7 + "voice" Shem-Tob Luke 9:36 20:30 + "they asked what this might be. It was said to them"

Shem-Tob Luke 18:36 20:34 + "and they praised God and followed him and all the

people praised God because of this" Shem-Tob Luke 10:43 26:70 +"woman" Shem-Tob Luke 22:57

Minus Readings

I have isolated 51 minus readings in Shem-Tob's Matthew. These are readings in the Hebrew text that are shorter than the Greek text of

35 For "Q" I have used the list of passages compiled by Joseph Fitzmyer, The Gospel According to Luke (I-IX) (AB 28; Garden City, NY: Doubleday, 1981) 77-79. The statistics are virtually the same if the outline of "Q" by Richard A. Edwards is used (A Theology of Q [Philadelphia: Fortress, 1976] XI-XIII). The list of J. S. Kloppenborg has only minor variations from that of

Fitzmyer; see Kloppenborg, The Formation of Q (Philadelphia: Fortress, 1987) 74-76. This major work on Q came to me only after completion of my manuscript.

36 For a definitive treatment of the minor agreements, see Frans Neirynck, The Minor Agree- ments of Matthew and Luke Against Mark with a Cumulative List (Leuven: Leuven University Press, 1974).




Matthew including all its manuscripts and versions. Of these 51 readings 23.5 percent agree exclusively with one or more witnesses to Luke; 5.8 percent agree exclusively with one or more witnesses to Mark. Again it is the case that Shem-Tob agrees more often with Luke than with Mark. In the non-Q sections the disparity between the Marcan and Lucan agreements is un- remarkable, there being 11.1 percent agreement with Luke and 8.3 percent agreement with Mark. In the Q sections, however, the agreement with Luke is much more pronounced. In these sections Shem-Tob agrees with Luke 53.3 percent of the time. Although Mark sometimes overlaps with Q, in the minus readings Shem-Tob never agrees with Mark. The following is a list of the agreements of Shem-Tob with Luke in the Q sections. There are 8 agreements with Luke out of a total of 15 minus readings.

4:7 "it is said/written" omit: Shem-Tob Luke 4:12 Sys 4:8 "again" omit: Shem-Tob Luke 4:5 6:9 "who art in heaven" omit: Shem-Tob Luke 11:2 P75 B etc.

Lvt (aur) vg Sys Marcion Origen 8:21 "first" omit: Shem-Tob Luke 9:59 W 69 1187 1352 Sj(1 ms) 11:22 "in the day of judgment" omit: Shem-Tob Luke 10:14 P45

D 472 1009 1241 Lvt (e d 1) Gg (I) 11:11 "yet he who is least in the kingdom of heaven is greater

than he" omit: Shem-Tob Luke 7:28 5 475* 1080* [Clem Rec 1.60.1-2]

11:27 "wishes" omit: Shem-Tob (Mss EF only) Luke 10:22 Mar- cion [Athpt Clem Cyr Just]

12:22 "the dumb spoke" omit Shem-Tob Luke 11:14 Lvt (a2 b)

III. Shem-Tob and Q

There appears to be a relationship between Shem-Tob's Hebrew Matthew and Q, as the statistics above regarding minus readings indicate. This becomes even clearer in regard to the Sermon on the Mount. The Greek text of Matthew 5-7 is laid out in the form of one long sermon beginning with the beatitudes and ending with the parable of the house built on a rock. There are no breaks in the text or indications that these chapters are formed from an anthology of sermons or sources. By contrast, many of these same sayings presented in Matthew's sermon are scattered throughout the various chapters of the Gospel of Luke and are listed by scholars as belonging to Q.

In the Hebrew Matthew the matter is different. The sayings are basically in the same order as the Greek Matthew, but they are frequently interrupted by the words "Jesus said to his disciples" or the like. There are no fewer than sixteen of these interruptions in the Matthean text of the Sermon, which leaves the distinct impression that various sources are being pieced together. As it stands, the Hebrew text of the Sermon is choppy and stylistically rough,




as though an editor, for whatever reason, failed to remove the introductory words of his various sources. The Greek appears to be a more polished edition of these same sayings with the disruptive introductory words removed. It thus gives the appearance of being one long sermon.

The Hebrew version of the Sermon is reminiscent of the Coptic Gospel of Thomas, which also appears to be a piecing together of various, sometimes unrelated Jesus sayings. The sayings in Thomas are constantly introduced by the words "Jesus said" or "Jesus said to his disciples" or the like. The similar-

ity between Shem-Tob and Thomas in this respect may suggest that Shem- Tob's form of the Sermon is primitive and unrevised.

Even more important is the relationship between the Q sayings in Luke and the introductory formulas in Shem-Tob's form of the Sermon. When the

sayings in Luke are placed alongside their parallels in the Hebrew text of Matthew 5-7, every time the Hebrew is interrupted by the words "Jesus said to his disciples,' without exception Luke jumps to a different place in his

Gospel or has a void. This may suggest that a common source or common sources stand behind this collection of Q sayings in Matthew and Luke and that the two evangelists have chosen a different arrangement for the source(s). The superfluous introductory formulas appearing in the Hebrew text may supply telltale marks of the parameters of the source or sources commonly used by Matthew and Luke. In order to see this point more clearly, Matthew's form of the Sermon with the introductory formulas found in the Hebrew text is placed here in outline form with Luke's parallels set beside them.

Matthew 5-7 (Introductory Formulas Supplied from Shem-Tob's Hebrew Matthew)

Matthew Luke 5:2-12 6:20, void, 6:21, void, 6:22-23 5:13-16 "At that time Jesus said 14:34-35, 8:16, 11:33

to his disciples" 5:17-19 "At that time Jesus void

said to his disciples" 5:20-24 "At that time Jesus void

said to his disciples" 5:25-26 "Then Jesus said 12:57-59

to his disciples 5:27-30 "Again he said to them" void 5:31-42 "Again Jesus said void, 16:18, void, 6:29-30

to his disciples" 5:43-6:1 "Again Jesus said 6:27-28, 32-36, void

to his disciples" 6:2-4 "Again Jesus said to void

them




6:5-15 "At that time Jesus 11:1-4 said to his disciples

6:16-18 "Again he said to them" void 6:19-23 "Again he said to them" 12:33-34, 11:34-36 6:24-7:5 "At that time Jesus said 16:13, 12:22-32, 6:37-42

to his disciples 7:6-12 "Again he said to them" void, 11:9-13, 6:31 7:13-14 "At that time Jesus said 13:23-24

to his disciples" 7:15-23 "Again he said to them" 6:43-44, 6:46, 13:27 7:24-29 "Again he said to them" 6:47-49, 4:32

One can see that the only possible exception to the rule stated above is the sequence of the Lucan parallels to 5:31-42 and 5:43-6:1. The Lucan

parallel to 5:31-42 ends with Luke 6:30; the parallel to 5:43-6:1 begins with Luke 6:27. Even here, however, the sequence is broken. In every other instance where the Hebrew Matthew has an introductory formula, the Lucan

parallel has a void or jumps to a different chapter. The phenomenon demonstrated by this outline seems to go beyond the point of mere coincidence and

may suggest that the introductory formulas in the Hebrew text mark different

segments of one source or mark different sources altogether which stand in common behind Matthew and Luke. It is significant that with the exceptions of the Lucan voids (some of which are unlisted for the sake of simplicity) all the parallels between Matthew and Luke belong to the hypothetical Q source. Other passages in the Hebrew Matthew which begin with the same formula (i.e., 10:23, 34; 11:11, 16; 13:44; 18:23; 21:28, 33; 23:34; 24:37, 42; 25:1, 14, 31) should be investigated in this light.

The disruptive Hebrew formula in Shem-Tob's form of the Sermon and its relationship to Q may have serious implications in regard to relationships between the Synoptic Gospels. Wellhausen, Torrey, Zimmermann and others traced some differences between the Synoptic Gospels back to Semitic roots and phrases. Others have found difficulty with this theory since it tends to place a stumbling block in the way of the "orthodox" explanation for the synoptic problem, namely, the two-document hypothesis. H. F D. Sparks, for example, wrote in 1951 the following: "it is impossible to maintain at one and the same time both that St. Matthew and St. Luke were dependent on St. Mark in Greek and also that the differences between them and St. Mark are to be explained as different translations of the same Aramaic original."37

Since 1951, however, considerable activity has taken place in regard to the synoptic problem. Butler, Farmer, and many others have challenged the

37 H. F D. Sparks, "Some Observations on the Semitic Background of the New Testament," Bulletin of the SNTS 2 (1951) 39.




"orthodox view" of synoptic origins38 Moreover, the Hebrew Matthew in Shem-Tob might be considered as adding fuel to the fire in disproving the two-document hypothesis. A scenario could be developed that makes the Greek Matthew a version based on the Hebrew Matthew and that argues that Mark and Luke used the latter, in one order or another, in their compositions.

But the evidence from the Hebrew Matthew does not in fact support this scenario. The relationship between the minus readings in Shem-Tob's Mat- thew and their parallels to the Q sections of Luke and between Shem-Tob's disruptive formulas in the Sermon on the Mount and the Q parallels in Luke militate against the above scenario and actually point to a common source or common sources that stand behind Matthew and Luke. The evidence suggests that the area of synoptic relationships vis-a-vis the Semitic background to the Gospels is a relatively unexplored frontier in synoptic studies that should not be avoided for the sake of an "orthodox" position, or, for that matter, for the sake of any previously expounded position.

IV. Shem-Tob and the Pseudo-Clementine Literature

There are several readings in Shem-Tob that agree exclusively or almost exclusively with either the Clementine Recognitions or the Clementine Homilies. They are the following:

8:26 + "immediately" = Shem-Tob Clem Homrn 19.14.3

1l:llb "yet he who is least in the kingdom of heaven is greater than he" omit: Shem-Tob Clem Rec 1.60.1-2 Luke 7:28 (Mss 5 475* 1080*)

13:21 "forget"= Shem-Tob; allusion in Clem Rec 1.26.1-2 (cf. Syriac)

13:22 "forget"= Shem-Tob; allusion in Clem Rec 1.26.1-2 (cf. Syriac)

16:16 "Christ" omit: Shem-Tob (comment only, after 16:20) Clem Hom 17.18.2, allusion in Clem Homrn 16.15.2 ("Christ" alluded to in Clem Hom 17.19.6). (Cf. John 6:69 b SyC "You are the Son of God:')

23:26 "cup" omit: Shem-Tob Clem Rec 6.11.3-4

In addition the two following should be noted:

5:14 "built" Shem-Tob Sys.c.p Clem Homrn 3.67.1 Thomas 32 POxy. 1 no. 7 Dta 8.41

10:5 "cities" (pl) Shem-Tob [a] b aur f h 1; vg Clem Rec 1.57.3 (sg = Clem Rec 1.63.1)

38 B. C. Butler, The Originality of St. Matthew (Cambridge: University Press, 1951); W R. Farmer, The Synoptic Problem: A Critical Analysis (New York: Macmillan, 1964).




V. Summary and Conclusion

The present paper investigates the textual affinitives of the Hebrew Mat- thew found in the fourteenth-century Jewish polemical treatise Even Bohan, by Shem-Tob ben-Shaprut. Shem-Tob's text lies closest to the traditions reflected by the Old Syriac, the Old Latin, and combinations of them. It also appears to have close ties with the Coptic Gospel of Thomas and the Pseudo- Clementine literature. Perhaps most interesting of all is its relationship to Luke's version of Q, especially in its minus readings and in the Lucan parallels to Matthew's Sermon on the Mount as introduced by the superfluous disruptive introductory formulas in the Hebrew text.

A conclusion of this study is that the textual affinitives of Shem-Tob's Hebrew Matthew support the theory that this Gospel text, in its earliest form, is not a product of fourteenth-century Spain. It is replete with readings reflected in diverse ancient traditions, some of which were lost in antiquity and were rediscovered only in modern times. It is highly unlikely that all of these readings were available to Shem-Tob on an ad hoc basis in the fourteenth century. It is more likely that Shem-Tob inherited, by whatever means, a Hebrew Matthew that already contained them. Whatever the case, the above analysis suggests that an old substratum exists in Shem-Tob's Hebrew Matthew which deserves further investigation.

