First Published in New Testament Studies 47 (2001): 118-25.
By Stephen C. Carlson. © 2001 Cambridge University Press, all rights reserved.
[¶1] Proponents of the Griesbach hypothesis have often appealed to a tradition handed down by Clement of Alexandria in support of their position that Mark used Matthew and Luke.1 Written early in his career in the now lost Hypotyposeis, Clement's information has been preserved for us by Eusebius of Caesarea as follows:
[¶2] But again in those very books Clement presented a tradition of the original elders (paradwsin twn anekaqen presbuterwn) about the "order" of the gospels (peri thV tacewV twn euaggeliwn) in this manner: He said that those of the gospels comprising the genealogies were "written before" (progegrafqai elegen twn euaggeliwn ta perieconta taV genealogiaV), but (de) that Mark had this "disposition" (tauthn eschkenai thn oikonomian): that when Peter was in Rome preaching the word openly (dhmosia) and proclaiming (exeipontoV) the gospel by the spirit, those present, who were many, entreated Mark, as one who followed him for a long time and remembered what was said, to record what was spoken; but that after he composed the gospel, he shared it (metadounai) with those who wanted it; that, when Peter found out about it, he did not actively discourage or encourage it; but that John, last, aware that the physical facts were disclosed (sunidonta oti ta swmatika en toiV euaggelioiV dedhlwtai) in the gospels, urged by friends, and inspired by the spirit, composed a spiritual gospel. So much for Clement. (Eus., Hist. eccl. 6.14.5-7)2
[¶3] Although leading source critics have disputed the value of this information,3 Clement's statement, progegrafqai elegen twn euaggeliwn ta perieconta taV genealogiaV, is widely understood to mean that Matthew and Luke, which include genealogies, were written first, i.e., before Mark and John.4 Under this interpretation, Clement's statement raises perplexing questions that have not been satisfactorily resolved.5
[¶4] For example, Clement's order stands isolated in apparent contradiction with the chronological order Matthew-Mark-Luke-John given by virtually every other patristic witness.6 In fact, Clement's order was unsupported until the ninth century.7 While variations and inconsistencies in tradition are to be expected in the diversity of early Christianity, Origen's conflicting statement on the order of the gospels is particularly troublesome due to Origen's direct relationship to Clement:
[¶5] As learned by tradition (wV en paradosei maqwn) about the four gospels, which alone are undisputed in the church of God under heaven, that first written was (prwton men gegraptai) Matthew, once publican but later apostle of Jesus Christ, who published it for the believers from Judaism (ekdedwkota auto toiV apo Ioudaismou pisteusasin), composed in Hebrew letters; but second (deuteron de), Mark, who composed as Peter led him, whom he avowed as a son in the catholic epistle, saying as follows: "She who is in Babylon, chosen together, sends you greetings and so does my son Mark"; and third (kai triton), Luke, who has composed for those from the gentiles (toiV apo twn eqnwn pepoihkota) the gospel praised by Paul; after all of them (epi pasin), John. (Eus., Hist. eccl.6.25.4-6)
[¶6] Though one of Clement's pupils as well as his immediate successor at the catechetical school of Alexandria,8 Origen explicitly ordered the gospels as Matthew first, Mark second, Luke third, and John after all of them, an order that places Mark before the genealogy-bearing Luke and contradicts Clement's order in terms of chronology.9 This contradiction is all the more striking because Origen claimed that he learned this from tradition (wV en paradosei maqwn), a tradition irreconcilable with that of Clement and his sources.10
[¶7] The difficulties posed by Clement's statement have prompted commentators to search for other interpretations, with increasing skepticism about the integrity of its context. For example, Theodor Zahn conjectured that Clement's comparison was meant to be pair-wise, Matthew before Mark and Luke before John.11 George Kennedy offered the possibility that "Eusebius has muddied the text of Clement by converting it into indirect discourse," which may have originally compared Matthew, Luke, and Mark with John.12 Helmut Merkel surmised that the ensuing discussion of Mark was a secondary insertion by Clement; thus, the original comparison in Clement's tradition was between John and the gospels with the genealogies.13
[¶8] In the most recent treatment of this passage, Denis Farkasfalvy argued that Clement's statement fits neither its literary context as found in Eusebius nor a second century historical context. Observing that Clement's discussion of Mark lacked "any reference to the idea of sequential ordering" and was only interested in how the gospel came about and its connection with Peter, Farkasfalvy ventured that Clement's testimony did not originally form a continuous text.14 Farkasfalvy further argued that Clement's statement in its present context could not have belonged to either Clement or his source, because the gospels that have the genealogies are implicitly compared in chronological sequence to those others that do not. According to Farkasfalvy, this comparison presupposes the closure of the four-fold gospel canon, an anachronism for Clement's "original elders" and in conflict with Clement's openness to other gospels.15 Farkasfalvy concluded that the present context of Clement's statement was a creation of Eusebius and that the original comparisons were with derivative, non-canonical editions of Matthew and Luke that lacked genealogies, i.e., between Matthew and the Gospel of the Ebionites and between Luke and Marcion's gospel.16
[¶9] These proposals are theoretically possible, but only at the expense of taking Clement's statement out of its context and into the realm of speculation. What is needed is an interpretation of Clement's statement that overcomes the difficulties but is still faithful to its textual and historical contexts.
[¶10] All of these interpretations, including the standard, construe progegrafqai in a temporal sense, "to have been written before, earlier, or first (in time)." Just as the preposition pro has a locative sense in addition to a temporal sense, so too does prografein have another sense, "to write before the public," i.e., "to set forth publicly" or "proclaim in public." 17 In particular, J. B. Lightfoot, in connection with Gal 3.1, commented that this verb "is capable of two meanings; (1) 'To write beforehand,' [and] (2) 'To write up in public, to placard.' It is the common word to describe all public notices or proclamations."18
[¶11] In classical usage, prografein was used for the written notice of official acts and decrees of the state, including meetings (e.g., Aeschines, Fals. leg. 60), military duty (e.g., Aristotle, Ath. pol. 53.7), and trials (e.g., Demosthenes, [Euerg.] 42). Aristophanes described these notices as written on tablets (prografwmen en toiV pinakioiV, Av. 450). The Hellenistic Jewish historian Josephus used this word to refer to an official decree published by Ptolemy Philadelphus (A. J. 12.33).19 In patristic times, Justin Martyr employed this verb to request the Roman Senate for a public sanction to his apology: "And if you will proclaim this in public, we will make it clear for everyone" (ean de umeiV touto prograyhte, hmeiV toiV pasi faneron poihsomen, 2 Apol. 15).
[¶12] The only other occurrence of prografein in Clement's writings is also preserved by Eusebius.20 According to Eusebius a little earlier at Hist. eccl. 6.14.2, Clement explained that Luke, who allegedly translated Paul's epistle to the Hebrews into Greek, did not prefix (mh progegrafqai) an identification of the apostle Paul to avoid prejudice among its intended readers. Undoubtedly, this word takes on the locative meaning, "to write in front of," because Paul's name is written at the beginning of every other letter in the Pauline Corpus. The public meaning also fits the context because Luke did not want Paul's authorship to be openly known to the recipients. The chronological meaning, though, is excluded.
[¶13] If progegrafqai is taken in the sense of writing publicly, Clement's statement would mean: "He said that those gospels having the genealogies were published openly," with an implication that their publication was official. In contrast with the standard chronological interpretation, this sense provides a better fit with its literary context and poses no difficulty for Origen's ordering of the gospels. Finally, Clement's statement under this proposal suits its context in Eusebius.
[¶14] Clement's discussion of Mark becomes coherent when Clement's statement about Matthew and Luke refers to the public nature of their promulgation. Immediately after stating that the gospels with the genealogies were published openly, Clement focused on the manner, not the time, in which Mark was written and disseminated.21 Peter was preaching openly and proclaiming the gospel. Some of Peter's audience wanted a record of the preaching and went to his associate Mark for it. Mark complied and circulated his gospel among them. When Peter found out about it, Peter did not stop or promote it. Thus, Clement explained that Mark's gospel was initially distributed to a limited number of people without the awareness or endorsement of Peter.22 The adversative conjunction de implies that the passage about Mark contrasts in some way with the statement on Matthew and Luke.23 If progegrafqai refers to the public (and sanctioned) nature of their publication, the comments on Mark begin to make sense as an apology for Mark's publication. While Mark could be faulted for its limited and haphazard circulation, it nonetheless consisted of the open, public preaching of Peter.
[¶15] Even Clement's following statement about John suits the proposed interpretation. Clement furthered the apology for Mark by explaining that John was aware of what was disclosed in the earlier gospels, i.e., although Mark's gospel had a limited publication, John nonetheless knew about it. Certainly, the temporal interpretation of Clement's statement on Matthew and Luke is consistent with the designation of John as last, but such chronological concerns are not present in the extended discussion of Mark's origin. As critics such as Merkel and Farkasfalvy recognized, interpreting Clement's statement on Matthew and Luke chronologically renders the whole of Clement's testimony on the gospels rather disjointed.24 Their objections to the passage's coherence, however, lose their force in the proposed interpretation, because Clement's testimonies were not concerned with chronology.
[¶16] The proposed interpretation heals the rift between Clement's statement on gospel origins with that of his successor Origen. Origen's explicit chronology is no longer in conflict, because Clement merely discussed the public nature, not the chronology, of the gospels. In fact, Origen expanded upon Clement's remarks on the publication of the gospels. He explained that Matthew was composed for the Jewish believers and Luke for the Gentiles. These are public compositions. On the other hand, Origen related only that Mark wrote as Peter led him, tying Mark to Peter with no implication about a public dissemination. Therefore, Origen's information, which he claimed to have learned from tradition, is concordant with Clement's testimony.
[¶17] The non-chronological interpretation also obviates Farkasfalvy's historical objection to the standard interpretation as anachronistic. If Clement's original elders had discussed the public and official nature of the publication of the gospels,25 then their testimony would comfortably fit within the growing orthodoxy's critique that the Gnostics rely on secret, non-public teaching (e.g., Irenaeus, Adv. haer. 3.1-3.5).
[¶18] The strongest objection to understanding progegrafqai non-chronologically does not stem from its context in Clement but from its context in Eusebius. In that context, Eusebius introduced Clement's tradition as "concerning the order (taxewV) of the gospels," which has been argued to mean a chronological order of composition.26 As a preliminary matter, if there is any discrepancy in the meaning of a word or phrase between its original context and its embedded context, the meaning in the original context should prevail. Thus, Eusebius's characterization presents a mild threat at best.
[¶19] In this case, there is no need to resort to this expedient, because the meaning of taxiV is much more general than that of a strict chronological order, referring to any kind of order, arrangement, or disposition.27 TaxiV is used in the title for Eus., Hist. eccl. 3.24, "Peri thV taxewV twn euaggeliwn," a chapter that recounts the origins of each gospel in canonical order, though with some chronological markers.28 Clement's own introduction to the publication of Mark employed the synonymous oikonomia, which can mean "arrangement" or "dispensation" like taxiV.29 Thus, taxiV would be an appropriate word for Eusebius to use in introduction Clement's account of the origins of the gospel, regardless of any possible chronological connotation.
[¶20] Clement's statement has puzzled commentators over the years because it has been uniformly interpreted in chronological terms, but its difficulties evaporate when progegrafqai is understood as a reference to the open publication of Matthew and Luke. In terms of its textual and historical contexts, therefore, the best interpretation of Clement's statement is that the gospels with the genealogies were written before the public, as gospels for all Christians.30 This interpretation makes better sense of the etiology of Mark that immediately follows and explains the patristic unanimity of the chronological order for the gospels outside of Clement. For the cause of synoptic source criticism, however, Clement's testimony can no longer be relied upon as evidence for the relative order of the gospels.31
[n1] 1. For example, in the 1984 Jerusalem Symposium on the Interrelations of the Gospels, the position paper for the modern form of the Griesbach hypothesis argued: "This one direct statement by Clement, the most highly respected Christian scholar of his day, is of more historical value than all the theories that have ever been propounded about the sequence in which the Gospels were written, including the theories of Augustine." Farmer 1990: 145. Other examples include Farmer 1976: 1, 225-7, 282; Farmer 1983: 1-15; Orchard & Riley 1987: 163-8; Orchard 1983: 603; and Dungan 1992: 6:677.
[n2] 2. My translation, rendering the words in quotation marks as generally as possible to avoid prejudicing the interpretative issues. Other translations, however, are more specific. For example, J. E. L. Oulton translated Clement's statement for the Loeb Classical Library as follows: "He said that those Gospels were first written which include the genealogies." Oulton 1932: 2:47, 49.
[n3] 3. For example, Tuckett 1983: 60 ("Clement's statement is very much out on a limb as far as patristic evidence is concerned").
[n4] 4. Farkasfalvy 1992: 261 ("Current scholarship routinely uses this text for claiming that Clement of Alexandria . . . held that Matthew and Luke were composed before Mark and John."). Some of these modern scholars include Hengel 1985: 4; Metzger 1987: 132; Davies & Allison 1988: 129 (only with respect to Matthew); Farmer 1990: 145; Brown 1993: 28 n9; and Schnelle 1998: 176.
[n5] 5. Some scholars have speculated on how Clement could have been wrong. For example, Theodor Zahn supposed that Clement derived his statement not from tradition but from the observation of Irenaeus that Mark was written after the death of Peter and an inference that Luke should predate the close of Acts. Zahn 1909: 2:295. B. H. Streeter suggested that Clement was influenced by the Western order of the gospels. Streeter 1924: 561 n1.
[n6] 6. For example, Irenaeus, Adv. haer. 3.1.1, in Greek at Eus., Hist. eccl. 5.8.2; Origen at Eus., Hist. eccl. 6.25.3-6; and Augustine, De cons. evang. 3.1.2. Efforts to demonstrate that Irenaeus and Origen did not intend to present a chronological order but a "theological" order (e.g., Farmer 1983: 9-14) are undermined by the explicit chronological indicators in the texts and uncorroborated by contemporary evidence of that theological order and its dominance. For example, Bruce M. Metzger noted that the Matthew, Mark, Luke, John order was "made popular by Eusebius and Jerome," but that as many as eight other orders have been attested among the manuscripts. Metzger 1987: 296.
[n7] 7. In the ninth century, the Irish monk Sedulius Scottus adopted the standard interpretation of this passage: "Matthaeo et Luca, qui secundam quosdam, ut ecclesiastica dicit historia, ante Marcum primo sua conscripsere Evangelia," quoted by Gamba 1983: 23. Although pseudo-Augustine (likely Ambrosiaster) gave the order Matthew, Luke, Mark, John, the author ruled out a chronological understanding: "Evangelium ordinatione collegitur magis quam tempore." Gamba 1983: 25-26. It is possible that the Monarchian Prologue to Mark could support Mark's posteriority to Matthew and Luke, but that text is "notoriously obscure" and requires resolution of a crucial textual variant. Tuckett 1990: 49-50. Clement failed to corroborate this order in the rest of his writings. Farkasfalvy 1992: 265 n10.
[n8] 8. See Eusebius, Hist. eccl. 6.3.3 and 6.6.1.
[n9] 9. Farmer characterized Origen's statement as "remarkable" in light of Clement's statement but supposed that "[t]here is no reason to think that Origen had evidence to support the order" and that "it is possible that . . . he does himself think of this order as the historical order in which these gospels were composed." Farmer 1983: 13-14. The very weakness of Farmer's rebuttals underscores the tension between Clement and Origen.
[n10] 10. On this point, I am indebted to Mark S. Goodacre, who spotted the importance of Origen's explicit reference to tradition in connection with my argument.
[n12] 12. Kennedy 1978: 150. Kennedy argued, "The Greek particle de in 6.14.16 can perfectly well associate Mark with the other two gospels rather than with John."
[n13] 13. Specifically, Merkel gave three reasons that the passage on Mark was an interpolation: (1) the basic idea is a comparison of the earlier-written gospels with John, but Mark neither has a genealogy nor is "pneumatic" like John; (2) only the passage of Mark is interested in biographical details; and (3) no explicit chronological signals are found in the Mark passage. Merkel 1990: 579-580. Tuckett (1990: 48) endorsed an earlier, similar argument by Merkel 1984: 382-385.
[n14] 14. Farkasfalvy 1992: 262. Farkasfalvy also thought the phrase tauthn eschkenai thn oikonmian signaled "an abrupt change in topic" (262) and noted an inconsistency in the usage of euaggelion. (262-63)
[n15] 15. Farkasfalvy 1992: 263-65.
[n16] 16. Farkasfalvy 1992: 268.
[n17] 17. LSJ 1996: 1473 s.v. prografw ("set forth as a public notice," "was proclaimed or set forth publicly," "= Lat. proscribere"); BADG 1979: 704 s.v. prografw ("show forth or "portray publicly, proclaim or placard in public," "placard publicly, set forth in a public proclamation so that all may read"); Lampe 1968: 1142 s.v. prografw ("set forth, show manifestly," all entries pertaining to Gal 3.1); Schrenk 1964: 1:771 s.v. prografw ("public promulgation"); and Sophocles 1914: 1936 s.v. prografw ("proscribo"). Cf. Lewis & Short 1879: 1474 s.v. proscribo ("To publish any thing in writing").
[n18] 18. Lightfoot 1865: 134. The latter meaning of prografein in Gal 3.1 is still maintained today. See generally, Davis 1999: 194-212.
[n19] 19. In the two other instances of prografein in Josephus (A. J. 11.283, 12.30), the word is used to refer back what had previously been written in the decree, which is compatible with both possibilities.
[n20] 20. Stählin 1936: 4:672 s.v. prografw.
[n21] 21. The reference to Peter is to clarify Peter's role in the publication of Mark; as a time indicator, it is merely incidental.
[n22] 22. Richard Bauckham contested the understanding of Clement to mean that Mark was written for a limited audience, because "Mark would have expected further copies to be passed on to other churches, in the normal way in which literature circulated in the early Christian movement." Bauckham 1998b: 14 n7. Regardless of Mark's motivations, admittedly speculative, Clement's discussion on Mark is concerned with its initial, limited public release.
[n23] 23. Pace Kennedy 1978: 150, whose idea that de could be conjunctive builds on his assumption that "Eusebius has muddied the text."
[n24] 24. Merkel 1990: 579-580; Farkasfalvy 1992: 262-263.
[n25] 25. Clement's statement is presented in indirect discourse (progegrafqai elegen), raising the question of who is speaking, Clement, as restated by Eusebius, or some earlier elder, as paraphrased by Clement. While Farkasfalvy supported the former possibility (1992: 260 n2), there are several reasons to adopt the latter. First, Eusebius had already introduced the quotation (touton ecousan ton tropon). The imperfect elegen suits an oral source, where another tense (e.g. perfect) would be more appropriate for a statement from a book. Eusebius just one paragraph earlier at Hist. eccl. 6.14.4 had previously quoted Clement's quoting a "blessed elder" (wV o makarioV elegen presbuteroV) in similar language.
[n26] 26. Zahn 1909: 400 n9 ("Clement means an historical account of the composition of the writings, observing the chronological order.") and Farmer 1990: 125 ("The use of taxiV in this context is decisive."). Bernard Orchard is more circumspect, deriving the chronological meaning of taxiV from the standard interpretation of progegrafqai. Orchard & Riley 1987: 164.
[n27] 27. For example, LSJ 1996: 1756 s.v. taxiV ("drawing up in rank and file, order or disposition of an army," and "generally, arrangement, order, . . . disposition").
[n28] 28. I am grateful to Peter M. Head for this point. The usage of taxiV is varied in Eusebius, Hist. eccl., with the most common meaning being a social or organizational "rank" (e.g. 1.3.17 (bis), 3.37.1, 5.19.2, 7.15.2, and 9.9.2). This noun was used once to refer to the arrangement of the books of the OT (4.26.13, opoia thn taxin), but there are no clear examples of the meaning of chronological order, with the possible exception of 5.6.5 th auth taxei, "in the same order" i.e. in the apostolic succession.
[n29] 29. LSJ 1996: 1294 s.v. oikonomia; Lampe 1968: 940-942 s.v. oikonomia ("disposition" and "dispensation"). See also BAGD 1979: 559-560 s.v. oikonomia ("in the pap. of the arrangements and directions of the authorities").
[n30] 30. Accordingly, the proposed interpretation provides additional patristic support for the thesis provocatively introduced by Bauckham 1998a.
[n31] 31. The author would like to thank Mark S. Goodacre, Peter M. Head, Leonard Maluf, and other members of the Synoptic-L mailing list for their comments, criticism, and encouragement of this idea.