The Letter to Theodore: Stephen Carlson’s Case against Clement’s Authorship

The Letter to Theodore: Stephen Carlson’s Case against Clement’s Authorship

SCOTT G. BROWN

Chapter four of Stephen Carlson’s The Gospel Hoax identifies certain features in the Letter to Theodore as anomalous within a private letter of Clement of Alexandria but explicable if the letter is a modern forgery. This paper exam-ines the logic of those arguments and demonstrates that each feature of the letter that Carlson deems suspicious is fully consistent with Clement’s modus operandi as revealed in his undisputed writings; the concordance is so striking as to add new support for the letter’s authenticity.

November of 2005 saw the release of Stephen C. Carlson’s The Gospel Hoax, a book alleging that Morton Smith forged his 1958 discovery of a letter of Clement of Alexandria “to Theodore” describing “a more spiritual gospel” that the evangelist Mark created in Alexandria for believers who were advancing in knowledge.1 Several scholars of Christian origins have already commended Carlson’s arguments, some going so far as to hail the demise of this controversial gospel.2 The author of the book’s foreword,

My sincere thanks are due to Jeff Jay and Allan Pantuck for their helpful critiques.

1. Stephen C. Carlson, The Gospel Hoax: Morton Smith’s Invention of Secret Mark (Waco, TX: Baylor University Press, 2005).

2. In addition to Mark Goodacre’s backcover endorsement (“Fascinating, com-pelling, and utterly convincing”), see Michael J. Kruger, review of Gospel Hoax, Journal of the Evangelical Theological Society 49 (2006): 422–24 (“a compelling, if not devastating, case against the authenticity of this apocryphal story,” p. 423); Craig A. Evans, Fabricating Jesus: How Modern Scholars Distort the Gospels (Down-ers Grove, IL: InterVarsity Press, 2006), 94–97 (“The evidence is compelling and conclusive: Smith wrote the text.” “No research into the Gospels and the historical

Journal of Early Christian Studies 16:4, 535–572 © 2008 The Johns Hopkins University Press

536 JOURNAL OF EARLY CHRISTIAN STUDIES

Larry W. Hurtado, more cautiously acknowledged that these arguments still need to be scrutinized, but added, “I think I can safely estimate where fair-minded scholarship will come down in the end, recognizing Carlson’s diligent and cogent exposure of Smith’s letter of Clement and its refer-ences to a secret Mark as an impressive fake.” I concur with Hurtado’s view that one cannot fairly assess Carlson’s arguments without meticu-lously reviewing his research, which is why I have not elected to bypass the process by forecasting the results.3 To this point I have examined the issue of motive, cross-examined Carlson’s evidence connecting Smith to the document, and, with Allan J. Pantuck, demonstrated that Carlson mis-identified an uneducated monk’s notation in another manuscript as both the same handwriting as the Letter to Theodore and a forgery by Smith.4 I turn now to Carlson’s case against Clement’s authorship of the letter, which comprises the fourth chapter of The Gospel Hoax.

I have already addressed the arguments at the beginning and end of this chapter, namely, Andrew Criddle’s statistical study, which Carlson summa-rizes, and Carlson’s exposition of an anachronistic allusion to the Morton Salt company in the letter’s metaphor “even the salt loses its savor.”5 Most

Jesus should take Smith’s document seriously,” pp. 95, 97); and Craig L. Blomberg, review of Gospel Hoax, Review of Biblical Literature (http://www.bookreviews.org) (2007) (“‘Secret Mark’ is now best laid to rest once and for all in the quest for the historical Jesus!”).

3. Larry W. Hurtado, foreword to Gospel Hoax, xii.4. Scott G. Brown, “The Question of Motive in the Case against Morton Smith,”

JBL 125 (2006): 351–83; “Reply to Stephen Carlson,” The Expository Times 117 (2006): 144–49; “Factualizing the Folklore: Stephen Carlson’s Case against Morton Smith,” HTR 99 (2006): 291–327; Allan J. Pantuck and Scott G. Brown, “Morton Smith as M. Madiotes: Stephen Carlson’s Attribution of Secret Mark to a Bald Swin-dler,” Journal for the Study of the Historical Jesus 6 (2008): 106–25. For my reply to a similar thesis published in 2006 by a musicologist at Princeton, see Scott G. Brown, an essay review of Peter Jeffery, The Secret Gospel of Mark Unveiled: Imagined Ritu-als of Sex, Death, and Madness in a Biblical Forgery, Review of Biblical Literature (http://www.bookreviews.org) (2007) (47 pages).

5. Carlson, Gospel Hoax, 50–54, 60–63; Andrew H. Criddle, “On the Mar Saba Letter Attributed to Clement of Alexandria,” JECS 3 (1995): 215–20. Criddle’s meth-odology is critiqued in Scott G. Brown, Mark’s Other Gospel: Rethinking Morton Smith’s Controversial Discovery (Waterloo, ON: Wilfrid Laurier University Press, 2005), 54–57; Allan Pantuck has since informed me that Criddle’s statistical method, which examines the letter’s ratio of words never before used by Clement to words used only once before, was tested on Shakespeare’s writings and shown to be unre-liable in determining authorship when only these two categories of words are con-sidered. Specifically, this method correctly identified the writer of only three out of seven poems tested, a success rate of forty-three percent, which is about as reliable

BROWN/LETTER TO THEODORE 537

of the remaining arguments stem from the premise that modern forger-ies of ancient letters convey authenticating information to their intended modern readers that would be unnecessary to convey to the stated (fic-tional) addressee(s). In this case, the letter’s opening remarks about the Carpocratians constitute a sphragis or literary seal of authenticity which would be unnecessary in an authentic letter of a private nature. The letter’s specification that Mark bequeathed his longer gospel (the term I prefer to “secret gospel”6) to the church “in Alexandria” (2.1) is likewise unneces-sary for its named addressee, who, having written to Clement, would know where Clement is residing, but is necessary for a modern reader. Further-more, the real Clement would have no need to reproduce two excerpts from the longer gospel in order to answer Theodore’s questions about it, nor would Clement need to describe precisely where these passages occur within Mark’s narrative. Clement would simply have asserted that Mark’s longer gospel did not contain the disturbing phrases that Theodore inquired about (3.13, 17). The distinctly Markan character of the gospel excerpts is also problematic because it implies that Clement quoted this gospel pericope more accurately than normal. And the letter’s tradition about the evangelist Mark inexplicably shares the modern gospel schol-ar’s preoccupation with literary composition, but differs from undisputed early traditions about Mark, “which are tantalizingly brief and usually stress the role of Mark’s memory, . . . not his note-taking ability.”7 The final argument is somewhat different. Carlson claims that the letter uses the imagery of salt losing its savor in a manner that differs from Clement’s interpretation of this metaphor.

These arguments sound plausible. But how well do they stand up to scrutiny?

as a coin toss. See Ronald Thisted and Bradley Efron, “Did Shakespeare Write a Newly-Discovered Poem?” Biometrika 74 (1987): 445–55. Pantuck further noted to me that if Criddle’s method had been applied to their data, it would have excluded at least two of the four undisputed poems of Shakespeare that Thisted and Efron used as controls. These statisticians concluded that authorship tests based on words not previously used and words previously used once were unreliable and that “there is no consistent trend toward an excess or deficiency of new words” (451). This fact by itself negates the conclusions about the Letter to Theodore that Carlson and oth-ers have based on Criddle’s study, leaving no basis for the common refrain that the letter is more like Clement than Clement normally is. Carlson’s claim to reveal an anachronism in the letter’s salt metaphor is thoroughly disproved in Brown, “Factu-alizing the Folklore,” 306–13.

6. “Secret gospel” is a mistranslation; see Brown, Mark’s Other Gospel, 121–43.7. Carlson, Gospel Hoax, 58.


A PRIVATE LETTER OF CLEMENT WOULD NOT INCLUDE A SPHRAGIS OR THE WORDS “IN ALEXANDRIA”

Argument: Following Charles E. Murgia, Carlson asserts that the letter contains a “sphragis immediately after the first line by alluding to Clement’s condemnation of the Carpocratians already known to modern critics in his Stromateis.”8 The sentences in question read as follows:

From the letters of the most holy Clement, the author of the Stromateis. To Theodore. You did well in silencing the unspeakable teachings of the Carpocratians. For these are the “wandering stars” referred to in the prophecy, who wander from the narrow road of the commandments into a boundless abyss of the carnal and bodily sins. For, priding themselves in knowledge, as they say, “of the deep things of Satan,” they do not know that they are casting themselves away into “the nether world of the darkness” of falsity, and, boasting that they are free, they have become slaves of servile desires. (1.1–7)9

These remarks, according to Carlson, “allud[e] to Clement’s condemna-tion of the Carpocratians” in the Stromateis (3.2; GCS 52[15]:197–200). The real Clement would not have needed to guarantee his authorship of his letter to Theodore by alluding to his published remarks because this is a “personal letter” of a “private nature” which would have been sent “by a trusted courier . . . and, thus, its receiver would have been able to trust that Clement was the author without the sphragis.”10 The sphragis would make sense only if Clement intended to publish this letter at a later point. Similarly, adopting an argument made by Attila Jakab,11 Carlson points to the letter’s statement in 1.28–2.1 that Mark bequeathed his gospel to the church “in Alexandria” as unnecessary clarification for Theodore, who supposedly wrote to Clement with questions about the Carpocratians (2.19) and therefore knows where Clement is situated, but might appear in a fictional letter (ancient or modern) or an authentic literary letter that

8. Carlson, Gospel Hoax, 54. Charles E. Murgia, “Secret Mark: Real or Fake?” in Longer Mark: Forgery, Interpolation, or Old Tradition? ed. Wilhelm H. Wuellner, Protocol of the Eighteenth Colloquy: 7 December 1975 (Berkeley, CA: Center for Hermeneutical Studies in Hellenistic and Modern Culture, 1976), 35–40.

9. Quotations of the letter are based on the translation in Morton Smith, Clement of Alexandria and a Secret Gospel of Mark (Cambridge, MA: Harvard University Press, 1973), 446–47, as modified in Brown, Mark’s Other Gospel, xvii–xxii.

10. Carlson, Gospel Hoax, 54, 55, 56.11. Attila Jakab, “Une lettre ‘perdue’ de Clément d’Alexandrie? Morton Smith et

‘l’Évangile secret’ de Marc,” Apocrypha 10 (1999): 7–15, at 13.


was written for a wider audience. If Clement was writing to Theodore from Alexandria, the word “‘here’ would have sufficed.” If, on the other hand, Clement was living elsewhere at this time, then the detail “in Alexan-dria” is reasonable, but Theodore’s letter to Clement is not, for Theodore would instead have written to someone who is in Alexandria and therefore had “access to the text.” That option is also excluded by the accuracy of Clement’s quotation, for he could not have quoted so carefully unless he took the longer gospel with him, whereas the letter indicates that this text was to this day carefully guarded in Alexandria (2.1).12

Assessment: The premises that a real personal letter of Clement would not have included any information that certified his authorship or any clarifying details that would be unnecessary for the original recipient are based on a false dichotomy, namely, that this letter, if authentic, must be either a private letter intended for no one but Theodore or a literary let-ter intended for publication. Other alternatives are possible, including the likelihood that Clement envisioned Theodore sharing this letter with other church officials. That possibility is certainly not excluded by the fact that this letter is directed to a particular individual. As M. Luther Stirewalt notes, ancient Greek letters “are often intended for a wider audience than the people formally addressed. Even letters bound to a familiar or limited official [i.e., administrative] context are not always confined to an exclu-sive party-to-party communication.” “Even in normative settings a writer may assume or intend that his message be shared with a larger audience than those people addressed. Thus an official letter may be publicized and permanently displayed; the writer of a personal letter may expect the letter to be passed among others not named by him, or its reception may be the occasion for a social gathering.”13 Lars Hartman notes that a letter-writer’s intention to address unknown readers who may find his or her advice useful is evident in the general way in which this advice is formulated. Texts occasioned by a particular situation “very often” treat that situation “as a type of problem, assessing it on a general basis” and supplying advice that can be applied in analogous circumstances. Hart-man argues that Paul’s letters fall into this category. Although they include some of the conventions of private letters, such as “introductory thanks-giving” and expressions of concern for the recipients’ well being, they are

12. Carlson, Gospel Hoax, 56–57.13. M. Luther Stirewalt Jr., Studies in Ancient Greek Epistolography, SBL Resources

for Biblical Study 27 (Atlanta, GA: Scholars Press, 1993), 15, 2.


also like “treatise[s]” in the sense that they are much longer than private letters and offer general rationales for Paul’s specific directives.14

In keeping with Hartman’s observation about letters that are written for a wider audience, the directives contained in the Letter to Theodore for dealing with the Carpocratians are formulated as general principles:

Such men [Carpocratians] are to be opposed in all ways and altogether. For, even if they should say something true, one who loves the truth should not, even so, agree with them. For not all true things are the truth, nor should that truth which merely seems true according to human opinions be preferred to the true truth, that according to the faith. (1.7–11)

To them, therefore, as I said above, one must never give way; nor, when they put forward their falsifications, should one concede that it is Mark’s mystic Gospel, but should even deny it on oath. For, “Not all true things are to be said to all men.” For this reason the Wisdom of God, through Solomon, advises, “Answer the fool from his folly,” teaching that the light of the truth should be hidden from those who are mentally blind. Again it says, “From him who has not shall be taken away,” and, “Let the fool walk in darkness.” (2.10–16)

Notice that the author carefully avoids the second person singular in the body of the letter, directing his words more generally to “one who loves the truth.” Theodore is directly addressed only in 1.2 (“You did well”), 2.19 (“To you, therefore”), and 3.13, 17 (“about which you wrote”). In its avoidance of the second person singular, the Letter to Theodore resem-bles letter-essays, which, although addressed to specific individuals, were intended for any academic who is interested in the topic.15 Moreover, as Jeff Jay demonstrated in his article in this volume, the Letter to Theodore fits an identifiable genre of letters that were devised to delegitimize unap-proved editions of a book.16 That function entails some dissemination of this letter’s contents. Hence we have good reason to suppose that this letter was intended for more than one pair of eyes. Since this is an administra-

14. Lars Hartman, “On Reading Others’ Letters,” HTR 79 (1986): 137–46, cit-ing 142, 138.

15. See Martin Luther Stirewalt Jr., “The Form and Function of the Greek Letter-Essay,” in The Romans Debate, ed. Karl P. Donfried, revised and expanded edition (Peabody, MA: Hendrickson, 1991), 147–71, at 163. Although the Letter to Theo-dore is not a letter-essay, it shares other characteristics of this genre, including the function of supplementing a literary work (i.e., Mark’s longer gospel) and limited use of epistolary conventions.

16. Jeff Jay, “A New Look at the Epistolary Framework of the Secret Gospel of Mark,” JECS 16 (2008): 573–97.


tive letter discussing the proper strategy for dealing with Carpocratians, the implied audience would not include ordinary Christians and persons outside the church. Nor would the audience include Carpocratians, who would learn from the letter that the orthodox church in Alexandria knows that their claim to possess a work by Mark that supports their beliefs is partly true. Rather, the implied extended audience consists of other church officials like Theodore who are in a position to “silenc[e] the unspeakable teachings of the Carpocratians” within their churches (1.2). They can carry-out its strategy of delegitimizing the falsified version of this text by avouch-ing that the Alexandrian church disconfirms Mark’s authorship.17

The two specific arguments constructed on the premise that the letter was written only for Theodore are problematic in other respects. To begin with, the identification of 1.2–7 as a sphragis assumes two things: that the Stromateis was written first and that the letter’s comments about the Carpocratians presuppose the reader’s familiarity with the things Clement wrote about Carpocratians in the Stromateis. It is not evident that we can assume these things. The mere fact that both of these writings condemn the Carpocratians does not demonstrate that the Stromateis was written first or that 1.2–7 deliberately alludes to it.18 The two descriptions of the Carpocratians are not close literarily, even in the reference to Jude at Str. 3.2.11.2 (GCS 52:200): “I fancy Jude was speaking prophetically of these and similar sects in his letter when he wrote: ‘So too with these people caught up in their dreams’ who do not set upon the truth with their eyes fully open, down to ‘pompous phrases pour from their mouth.’”19 The assumption that the letter alludes to the Stromateis would be less tenden-tious if the former’s remarks about the Carpocratians were inexplicable except as assurance that the writer is Clement, but that is not the case. From

17. The prescript of the Letter to Theodore (“From the letters of . . .”) implies that this letter was eventually published but cannot tell us whether this was Clement’s intention. Since, as E. Randolph Richards has pointed out (“The Codex and the Early Collection of Paul’s Letters,” Bulletin for Biblical Research 8 [1998]: 151–66), it was a standard practice for writers to keep copies of the letters they wrote, one might conjecture that the person who inherited Clement’s notebooks not only fashioned a collection of letters from Clement’s own copies but also published his Excerpta ex Theodoto and Eclogae Propheticae, mistakenly assuming that the latter were intended for publication.

18. The reference to the Stromateis in the opening sentence presupposes the existence of a collection of letters of Clement of which this one is a part and is not, therefore, something that the implied author wrote to Theodore.

19. John Ferguson, trans., Clement of Alexandria: Stromateis, Books One to Three (Washington, D.C.: The Catholic University of America Press, 1991), 263.


the perspective of the implied reader(s), these comments are informative. They give Theodore and other Christian officials basic information about these doctrinal opponents, which would be unnecessary if the implied author is presupposing knowledge of the Stromateis. More importantly, these comments flow naturally from the chain of events that the letter presupposes. One can deduce from the letter that Theodore encountered Carpocratians who were justifying their heretical positions with reference to a different gospel that Mark wrote in Alexandria. Theodore censured them but wrote to Clement to find out whether Mark really did write a special Alexandrian gospel that contains phrases such as “naked man with naked man.” Clement wrote back to assure Theodore that he was right to censure the Carpocratians and to set him straight about Mark’s lon-ger gospel. Accordingly, this letter begins, “You did well in silencing the unspeakable teachings of the Carpocratians.” I am hard put to imagine how Clement could have answered an inquiry about a different Gospel of Mark used by the Carpocratians without mentioning the Carpocrati-ans at the outset of his reply. Furthermore, as I will demonstrate later, the letter’s claim that Jude prophesized the Carpocratian heresy accords with Clement’s strategy for countering heretics who misrepresent scripture, which was to show, where applicable, that scripture predicted their heresy and to refute their heretical statements by giving the true interpretation of their proof texts (see Str. 3.6.45.1–3; GCS 52:216–17). Since these com-ments about the Carpocratians are integral to the letter and accord with Clement’s approach to polemics, the notion that they also function as a sphragis to assure Theodore of Clement’s authorship is superfluous.20

The argument concerning the words “in Alexandria” is problematic as well. Let us take a look at the sentence in question: “Thus, in sum, [Mark] prepared matters, neither grudgingly nor incautiously, in my opinion, and, dying, he left his composition to the church in Alexandria, where it even yet is preserved with utmost discretion, being read only to those who are being initiated into the great mysteries” (1.26–2.2).21 The claim that the words “in Alexandria” are unnecessary for Theodore but somehow nec-essary to a modern reader is puzzling, because the modern reader already learned at 1.18–21 that the longer text was created in Alexandria: “But when Peter died a martyr, Mark came over to Alexandria, bringing both

20. For a critique of Murgia’s arguments to this effect, see Brown, Mark’s Other Gospel, 28–34.

21. On the translation of ésfal«w eÔ mãla thre›tai, see Jay, “Epistolary Framework,” 591.


his own notes and those of Peter, from which he transferred to his former book the things suitable to those studies which make for progress toward knowledge.” Had the letter stated that Mark, “dying, . . . left his com-position to the church here” rather than “in Alexandria,” the modern reader would naturally infer from 1.18–21 that “here” means Alexandria and not some other city. Even if the letter had used “here” in both places, that conclusion would be hard to avoid, for the letter gives no indication that Mark relocated before he died, and tradition places Mark’s death in Alexandria.22

The repeated reference to Alexandria therefore conveys no new informa-tion, but it does have a rhetorical purpose. It underscores the point that Mark bequeathed his composition to this specific church. The implied author is not suggesting that the church in Alexandria happens to have this gospel because Mark happened to die there. Rather, he is indicating that Mark’s more spiritual gospel is the exclusive property and responsibil-ity of the church in Alexandria and was not supposed to circulate outside this church in any form. The circle of hearers has always been restricted to “those who are being initiated into the great mysteries,” which is a distinc-tive aspect of Alexandria’s program of theological instruction.23 Given the claim of proprietorship that the implied author is making here on behalf of the Alexandrian church, the reference to Alexandria in 2.1 is natural.

CLEMENT WOULD NOT NEED TO QUOTE THE GOSPEL EXCERPTS

Argument: The next argument is the most cogent argument in this chap-ter, and perhaps in the whole book. Carlson asks why Clement would need to quote the story of the raising and instruction of the young man in order to answer Theodore’s questions about it. Why indeed would he tell Theodore where these verses occur in relation to the canonical gospel?

22. E.g., Jerome, Famous Lives 8 (Aldo Ceresa-Gastaldo, ed., Gerolamo: Gli Uomini Illustri, Biblioteca Patristica 12 [Florence: Nardini Editore, 1988], 90]); The Acts of Mark; Palladius, Lausiac History 45.7 (Adelheid Wellhausen, Die lateinische Übersetzung der Historia Lausiaca des Palladius: Textausgabe mit Einleitung, Patris-tische Texte und Studien 51 [Berlin; New York: Walter de Gruyter, 2003], 642). For discussion of these sources, see Birger A. Pearson, “Earliest Christianity in Egypt: Some Observations,” in The Roots of Egyptian Christianity, ed. Birger A. Pearson and James E. Goehring, SAC (Philadelphia: Fortress, 1986), 140–44.

23. The great mysteries are theological mysteries that are revealed through anagogical exegesis of scriptures. See Brown, review of Secret Gospel of Mark Unveiled, 3–14.


All Clement needed to do was tell Theodore that the real longer gospel did not contain the disturbing phrases about which he enquired. On the other hand, this information is “essential for modern scholars—to analyze its text” and to conceptualize it as part of “a source document underly-ing Secret Mark and John, a source Smith did not fail to hypothesize (e.g. Clement 194).”24

Assessment: Carlson’s point that Clement would not need to quote the longer text in order to answer Theodore’s inquiry is correct. The problem with this argument is not its logic but its neglect of the evidence. Certainly it is possible to imagine a church father flatly asserting that the Carpocra-tian text does not include the things that Theodore asked about, for that is what many of us would do in his place, perhaps in a hurried e-mail. But we live in a very different world from that of Clement. In his world, it was not unusual for Christian letter-writers to use lengthy scriptural excerpts to prove their points, as 1 Clement and the Epistle of Barnabas—two favor-ites of Clement—amply attest. The question we should be asking is not, What would I do in this situation? but, What would Clement of Alexan-dria do? As his undisputed writings abundantly demonstrate, Clement put much stock in quotations. He quoted the Bible thousands of times, along with thirty-two other early Christian writings and over 350 non-Christian authors.25 It should not surprise us therefore to discover that when coun-tering heretical interpretations of scripture, Clement quoted the passages that his doctrinal opponents misappropriated.

The most apropos quotations occur in Stromateis 3. When Clement addressed Christian groups that he opposed, his practice was either to discredit them as Christians by quoting passages from their own writings that highlight their insincerity, immorality, or heresy, or to refute their doctrines by quoting their scriptural proof texts and then giving a truer exegesis. Thus, he quoted extensively from the writings of the early fol-lowers of Basilides in order to prove that those among Basilides’ later fol-lowers who thought that they were free to sin due to their election were not adhering to the high morality of their predecessors (3.1.1.1–3.4; GCS

24. Carlson, Gospel Hoax, 57–58.25. Andrew C. Dinan, “Fragments in Context: Clement of Alexandria’s Use of Quo-

tations from Heraclitus” (PhD diss., The Catholic University of America, 2004), 2–3. The exact numbers for Clement’s biblical quotations are difficult to specify. Counting only direct quotations, Annewies van den Hoek (“Techniques of Quotation in Clem-ent of Alexandria: A View of Ancient Literary Working Methods,” VC 50 [1996]: 223–43, at 241 n. 55) finds roughly 1600 OT references and 2100 NT references.


52:195–97). And Clement quoted several pages from the writings of Epi-phanes in order to prove that he and his father Carpocrates held particular immoral beliefs (3.2.5.1–8.3, 9.2–3; GCS 52:197–200).

The tactic that is most relevant for our purposes is that of countering false doctrines through proper exegesis of the proof text. Clement did this numerous times in Stromateis 3. For instance, where certain heretics (apparently Prodicus’s school) quoted Mal 3.15 as evidence that salvation comes from challenging the “shameless god” of this world, Clement quoted the verse himself (twice), isolated “the shameless god” as an interpolation, and then offered his own exegesis, describing what the verse meant in its larger, literary context:

They say that Scripture has this: “They opposed God and found salvation.” But they add, “the shameless God.” They accept this saying as advice extended to them. They think that it is salvation to oppose the creator. Scripture does not say, “the shameless god.” And even if it did, you idiots, it would be talking of the one we call the devil as shameless, whether as the maligner of humanity, or as the prosecutor of sinners, or as an apostate. At any rate, the people referred to in the passage objected to being disciplined for their sins; they protested and murmured at the passage quoted because the other nations were not being punished for their offenses while they alone were put down for every single offense. Even Jeremiah was led to say, “Why is the path of the wicked easy?” The passage from Malachi already quoted is to the same effect: “They opposed God and found salvation.” The prophets in their oracular utterances do not merely say that they have heard certain messages from God; they demonstrably report the popular conversations, replying to objections voiced, as if they were officially recording questions from human sources. The saying before us is an example of this.26

The same approach in the same order is taken in the Letter to Theodore, where the author quotes the disputed passages (2.23–3.11, 14–16), notes that the heretics’ exegetical interpolations do not appear there (3.13–14, 17), and then gives the proper exegesis (beginning at 3.18, where the manuscript ends).

Clement explicitly described his tactic of exposing false exegesis as the basis of false doctrine when demonstrating how certain heretics use scrip-ture (in this case the Gospel of the Egyptians) to justify their anti-worldli-ness and rejection of procreation:

26. Str. 3.4.38.2–5 (GCS 52:213; Ferguson, Clement, 279).


This is what we must say to them; first, in the words of the apostle John: “Now many antichrists have come, from which we know that it is the last hour. They went out from us, but were not of our company: if they had been, they would have stayed with us.” Next we must [over]turn their [heretical] statements on the grounds that they destroy the sense of their citations. Here is an example: When Salome asked, “How long will death maintain its power?” the Lord said, “As long as you women bear children.” He is not speaking of life as evil and the creation as rotten. He is giving instruction about the normal course of nature. Death is always following on the heels of birth.27

These people do anything rather than walk by the canon of the gospel in conformity with truth. Why do they omit what follows in the words spoken to Salome? She said, “Then would I have done better if I never had a child?” suggesting that childbearing was not a necessary obligation. The Lord replied in the words, “Eat every plant but do not eat a plant whose content is bitter.” By these words he is indicating that the choice of celibacy or wedlock is in our power and not a matter of the absolute constraint of a commandment. He is also clarifying the point that marriage is cooperation with the work of creation.28

This, in a nutshell, is Clement’s strategy for answering heretics who distort the gospels in order to justify impious practices: show (wherever feasible) that scripture predicted their heresy and refute their heretical statements by quoting their proof texts and giving the true interpretation. This two-pronged attack is taken in the Letter to Theodore, which first informs Theodore that the Carpocratians are “the wandering stars” predicted in Jude 13 (1.2–7) and then deconstructs their interpretation of their proof texts.29

On matters of scriptural exegesis, Clement’s main complaint against the heretics was that they quote their passages selectively and out of context (Str. 7.16.96.2–3; GCS 17:68). Here is a typical complaint:

Those who drag in a doctrine of moral indifference do violence to some few passages of Scripture, thinking that they support their own love of pleasure; in particular, the passage “Sin shall have no authority over you; for you are not subject to sin but to grace” [Rom 6.14]. But there are other such passages. . . . Let me quickly cut through their attempt. The admirable

27. Str. 3.6.45.1–3 (GCS 52:216–17; Ferguson, Clement, 284).28. Str. 3.9.66.1–3 (GCS 52:226; Ferguson, Clement, 297); cf. 63.1–64.3 (GCS

52:225).29. Cf. Str. 3.2.11.2 (GCS 52:200), where Clement applied Jude 8–16 to the Car-

pocratians; likewise, 3.1.3.4 (GCS 52:196–97), where Clement applied 2 Cor 11.13–15 to followers of Basilides.


Apostle in person will refute their charge in the words with which he continues the previous quotation: “Well then! Shall we sin because we are no longer under Law but under grace? God forbid!” With these inspired prophetic words, at a single stroke he undoes the sophistical skill at the service of pleasure.30

As the three preceding examples illustrate, Clement responded to these challenges by quoting not only the abused scripture but also elements of its literary context that confound heretical interpretations, often quoting what came immediately before or immediately after the passage in ques-tion and working that information into his refutation. We see this tactic again when he countered the anti-marriage interpretation of Jesus’ say-ing about men who have “made themselves eunuchs for the kingdom of heaven” (Matt 19.11–12):

What about these words: “Not everyone can take this saying. There are some eunuchs born as eunuchs, and some who were made eunuchs by human action, and some who have made themselves eunuchs for the kingdom of heaven. Let anyone who can accept the words accept them.” They [the Christians opposed to marriage] do not recognize that it was after his words about divorce [i.e., Matt 19.3–9] that some of them asked whether, if that is the position with regard to the wife, it is not better to refrain from marriage [19.10], and it was then that the Lord said, “Not everyone can take this saying, only those who have a gift.” Those asking the question wanted to find out whether, when a wife had been condemned for sexual misconduct and removed, there was any advantage in marrying another.31

By quoting the saying and describing its place within its literary context, Clement was able to establish that the saying about eunuchs is not anti-marriage and not directed at everyone. He did the same thing when refut-ing Carpocrates’ use of Matt 5.42 to justify sexual communism:

This is the way to undermine the “righteousness” of Carpocrates and those who match him in sharing in a fellowship of immorality. In the moment of saying, “Give to anyone who asks,” Scripture goes on, “and do not turn away anyone who wants a loan.” This is the sort of fellowship Scripture teaches, not fellowship in lust. . . . What does the Lord say? “I was hungry and you gave me food. I was thirsty and you gave me drink. I was a stranger and you took me into your home. I was naked and you gave me clothes to wear.” Then he adds, “Insofar as you have done so to one of the humblest of these, you have done so to me” [Matt 25.35–36, 40]. The same

30. Str. 3.8.61.1–2 (GCS 52:224; Ferguson, Clement, 294).31. Str. 3.6.50.1–3 (GCS 52:219; Ferguson, Clement, 287).


law is established in the Old Testament in the words “Anyone who gives to a beggar is making a loan to God” and “Do not evade doing good to one in need” [Prov 19.17; 3.27].32

As this example shows, Clement sometimes expanded his exegesis to include relevant statements from elsewhere in the same text and from other parts of scripture, an exegetical technique he described in Str. 7.16.96.4 (GCS 17:68). Unlike Clement’s didactic exegeses, which often disregard the context and the plain sense in favor of subjective allegorical meanings, his polemical exegeses are properly exegetical inasmuch as they interpret the disputed passages in their immediate literary contexts. That is what makes them compelling.

Clement’s attention to the immediate literary context when refuting heretical misuses of scripture adequately accounts for why the author of the Letter to Theodore cites the sentences that occur before and after the longer Gospel of Mark’s story of the raising and initiation of a young man in Bethany (hereafter LGM 1) and before its account of Jesus’ refusal to receive the young man’s sister, his mother, and Salome in Jericho (LGM 2) in the Gospel of Mark (Letter to Theodore 2.21–22; 3.11–14). If the author is Clement, then his lost “true explanation and that which accords with the true philosophy” (3.18) would probably have interpreted this pericope within its literary context. The letter’s abbreviated quotations of the verses that surround LGM 1 (Mark 10.32–34 and 35–45) therefore anticipate his exegesis by establishing the literary context. There is one other, equally sufficient reason for describing the literary context. The addition constituting LGM 2 conveys little meaning on its own: “And there were there the sister of the young man whom Jesus loved and his mother and Salome, and Jesus did not receive them.” This incident cannot be imagined, let alone interpreted, without the information that it occurs after the words “And he comes into Jericho” (Mark 10.46). Since it was necessary to state where LGM 2 occurs in relation to the canonical gos-pel, the author had additional reason to state where LGM 1 occurs, for it would seem odd to his reader to learn the position within Mark of the second half of this story but not the position of the first half.

The quotation of LGM 1 and 2, the information about their literary contexts, the isolation of interpolations, and the true explanation are all consistent with Clement’s approach to refuting heretical misappropria-tions of scripture, as is the letter’s opening quotation from Jude, which

32. Str. 3.6.54.1–4 (GCS 52:221; Ferguson, Clement, 289–90).


Carlson labeled an unnecessary sphragis. Clement wrote at a time when heresies threatened to discredit the church and lead its members astray. Accordingly, he felt obligated to convince his Christian readers of “the true truth” with intelligent refutations of heretical interpretations.33 To this end, simple assertions that the heretics have misrepresented the scriptures would not do. As Clement explained in the latter part of chapter 15 and much of chapter 16 of Stromateis 7, such matters cannot be settled by conflicting assertions but must be established by the only true demonstra-tion, which is the scriptures themselves (esp. 7.15.92.1–3; 16.95.3–96.1; GCS 17:65, 67–68).

CLEMENT WOULD NOT QUOTE THE GOSPEL EXCERPTS SO ACCURATELY

Argument: Carlson’s next argument concerns the accuracy of the letter’s gospel quotations:

Clement’s quotations of gospel materials are usually so free that it is sometimes difficult to tell whether Clement used a different form of the text than what we now possess, or whether Clement was in the habit of quoting loosely and harmonizing to the other gospels and even to extra-canonical material. By contrast, the Markan features in the Secret Mark passages are so salient that Clement must have quoted from it more meticulously than usual. Since Clement’s decision to quote Secret Mark in the first place is unnecessary, Clement’s unusual care in doing so is inexplicable. . . . On the other hand, Theodore’s care in quoting the exact text of Secret Mark, even to the point of preserving its Semitisms (Clement 133–34), is vital for modern form critics to date the content to be earlier than the gospel of John.34

Assessment: We have no manuscripts or other patristic citations from which to construct a “critical text” of the longer Gospel of Mark’s addi-tional verses, so this argument’s basic assumption that the letter’s gospel quotations are not sufficiently free or corrupt to be by Clement is impos-sible to verify. One may certainly concur that LGM 1 and 2 are typically Markan, for they contain the same theological emphases and employ

33. Citing Letter to Theodore 1.10–11. Ironically, Clement himself acknowledged that his refutation of the heretics in Stromateis 3 was excessive: “But our critique of the hypocritical pretenders to knowledge, however essential, has gone beyond what is necessary and stretched out our discourse to a considerable length” (18.110.3; GCS 52:247; Ferguson, Clement, 326).

34. Carlson, Gospel Hoax, 57.


the same literary techniques as the canonical gospel.35 Yet most of these emphases and techniques would be retained in a loose or corrupt quota-tion, provided it preserves the basic sentences of the original, which is true of Clement’s loose quotation of Mark 10.17–31 in Quis dives salvetur 4. At issue here are the small-scale Markan features of vocabulary and syn-tax. Are these features so thoroughly and uniquely Markan that they rule out an inaccurate quotation? The many scholars who presume that LGM 1 and 2 are a pastiche of phrases and story elements derived from all four of the canonical gospels would be restrained by their own logic from fol-lowing Carlson here.36 I judge their scenarios of literary borrowings highly implausible, but the fact that Pierson Parker and Helmut Merkel could make a case for this position by discussing the relative frequencies of longer Mark’s phrases within the four canonical gospels shows that the language of LGM 1 and 2, while typical for Mark, is not as distinctively Markan as Carlson’s argument requires.37 The Markan character of the wording and syntax of the longer text therefore is not something one could quantify in a way that could establish even a base level for the accuracy of its repro-duction, such as that it must be at least eighty percent accurate. Carlson himself offers only a relative statement: “Clement must have quoted from it more meticulously than usual.”

Although Carlson has not established his premise that the quotations are more accurate than usual (or that Clement’s text of longer Mark must be less corrupt than his text of canonical Mark), I am willing to grant that claim for the sake of argument because my critical instincts tell me that it is correct, and the question of whether Clement could have transmit-ted an accurate quotation of LGM 1 and 2 is important to anyone who treats these excerpts as authentic traditions. I propose therefore to exam-ine two related issues: would Clement have quoted a pericope accurately under the conditions represented in the letter? And could a reasonably accurate copy of the longer gospel have existed in Alexandria by the end of the second century?

There is an obvious problem with the notion that the real Clement of Alexandria would have provided a less accurate quotation of this pericope: the implied author has just condemned Carpocrates for changing scrip-

35. See part 2 of Brown, Mark’s Other Gospel.36. For discussion of their position, see Brown, Mark’s Other Gospel, 9–11,

92–101.37. Pierson Parker, “On Professor Morton Smith’s Find at Mar-Saba,” ATR 56

(1974): 53–57; Helmut Merkel, “Auf den Spuren des Urmarkus? Ein neuer Fund und seine Beurteilung,” ZTK 71 (1974): 123–44.


ture and is now demonstrating what the text really says. In this context, an erroneous quotation of the contested passages would both undermine the author’s criticism of Carpocrates and invalidate his demonstration that the Carpocratian longer gospel is corrupt. Hence the letter-writer has good reason to avoid the careless interpolations, omissions, paraphrases, and harmonizations to other gospels that we see, for example, in Clem-ent’s quotation of Mark 10.17–31 and instead to quote “word for word” (katå l°jin) “the very words of the Gospel” (2.22; 2.20). Carlson evades this conclusion with the following deduction: “Since Clement’s decision to quote Secret Mark in the first place is unnecessary, Clement’s unusual care in doing so is inexplicable.” Although this claim has the form of a logical statement, the premise is faulty (this unnecessary decision is nev-ertheless the one Clement would make) and the deduction is a non sequi-tur (anyone who made this decision would be constrained to quote the passage accurately).

What, then, should we make of the fact that Clement’s renditions of Jesus’ sayings are often rather loose? The main reason for this is straight-forward: in didactic contexts, Clement was more concerned with the meaning of a gospel saying or passage than with its precise wording. He therefore felt free to quote the gospels loosely, occasionally adjusting the wording to fit his syntax or better convey his point, as preachers typically do in sermons.38 Although the freedom in these quotations may at times be intentional,39 I agree with Carl P. Cosaert that the looseness usually results from faulty memory. Cosaert’s recent book on Clement’s text of the gospels identifies various indicators of reliance on memory in Clement’s gospel quotations, including conflation of similar passages, misattribution, vague attribution (“it says somewhere in Scripture”), and unique read-ings (the variants have no parallels in extant manuscripts or other church fathers).40 There is, however, no simple way to characterize Clement’s habits of quotation, as Cosaert’s conclusions reveal:

38. See, e.g., H. B. Swete, Introduction to the Old Testament in Greek, 2nd ed. (Cambridge: Cambridge University Press, 1914), 426; Leon E. Wright, “Contextual Adaptation in Incomplete Quotation,” JBL 67 (1948): 347–51; Joshua Bloch, “Was There a Greek Version of the Apocalypse of Ezra?” The Jewish Quarterly Review, n.s., 46 (1956): 309–20, at 312 (citing Swete).

39. See John Whittaker, “The Value of Indirect Tradition in the Establishment of Greek Philosophical Texts or the Art of Misquotation,” in Editing Greek and Latin Texts: Papers Given at the Twenty-Third Annual Conference on Editorial Problems, University of Toronto, 6–7 November 1987, ed. John N. Grant (New York: AMS, 1989), 63–95.

40. Carl P. Cosaert, The Text of the Gospels in Clement of Alexandria, SBLNTGF 9 (Atlanta, GA: Society of Biblical Literature, 2008), 24–32, at 25–26.


Clement’s citation habits generally depend on the source of his quotation, or the nature of his polemic. If he is referring to a passage from Paul, Clement generally cites the passage with a high degree of accuracy, especially if some form of introductory formulae precedes it. The high degree of accuracy in such cases strongly suggests that these quotations derive from a New Testament manuscript before him. The length and consistently high level of textual exactitude of his Pauline quotations makes any other conclusion unlikely. The situation, however, is not so simple when it comes to Clement’s use of the Gospels.

Clement’s Gospel citations focus almost exclusively [on] the words of Jesus. His tendency for the majority of these quotations is to cite the passage from memory. This can often result in a conflation of similar passages, or even a very loose citation. At other times, his memory is more accurate, but even in these cases he usually modifies the text to fit his context, if ever so slightly. In a more limited number of places, Clement’s quotations are more exact, and this appears to be a result of his dependence upon a manuscript for the citation. This typically occurs when Clement introduces his quotation with either the introductory formula eÈaggel¤ƒ or fhs‹n ı kÊriow, though even here the text may be slightly modified to fit his context. Gospel citations introduced by other more common introductory formulae used by Clement are not as accurate, the only exception being those isolated places where a chain of quotations occurs.41

The fact that the sayings of Jesus are generally easier to remember, in their essence, than the verbose and theologically dense discursive writing of Paul helps explain why Clement more often consulted manuscripts of Paul’s letters. Clement, moreover, viewed the words of Jesus as “a living voice . . . [that] continues to speak catechetically to all those who are will-ing to hear.” He therefore felt freer “to allow that voice to address people living in his own day” by liberating “the sense/meaning of that voice” from “the actual ‘words’ of the text.”42 In this connection it is worth not-ing that the letter’s excerpts from longer Mark contain no words of Jesus, and the formula katå l°jin in the Letter to Theodore 2.22 also introduces one of the two chains of quotations that Cosaert associates with greater accuracy (the other is introduced with diarrÆdhn).43

So the issue is complex. The general observation that Clement felt free to quote loosely when the exact wording did not matter does not tell us how accurately he would have quoted a pericope when his intention was

41. Cosaert, Text of the Gospels, 30–31.42. Cosaert, Text of the Gospels, 32.43. Cosaert, Text of the Gospels, 29 n. 33.


to demonstrate that it does not contain particular phrases. A sensible way to determine this is to examine the places in Clement’s undisputed writ-ings where he claimed to be quoting a text katå l°jin, especially those places where the precise wording of that text mattered to his argument, as it does in the letter.

Clement used katå l°jin twenty-four times in his undisputed writings, all but one of them in the Stromateis. The formula appears in a variety of circumstances. In two instances there is no evident reason for him to note that his quotation is verbatim, specifically, his quotation of Clement of Rome (Str. 1.7.38.8; GCS 52:25) and his string of four quotations from the gospels (4.6.34.3–6; GCS 52:263). The vast majority of instances, however, more closely resemble the situation in the letter insofar as Clement’s claim that he is quoting verbatim bolsters his argument. When citing passages with which he agreed, Clement normally used this expression in the sense “(authority x) says this explicitly” or “in these exact words.” Most of his “word for word” citations of non-Christian writers fall into this group.44 Those that do not, occur in connection with a charge of plagiarism. He assured his reader that Homer’s line “As a man trains a luxuriant shoot of olive” and what follows occur katå l°jin in Orpheus’s Disappearance of Dionysus and that Hesiod took some verses katå l°jin from the poet Musaeus (Str. 6.2.26.1–3; GCS 52:442). More often, though, katå l°jin introduces passages that offend Clement. Indeed, half of the twenty-four instances of katå l°jin occur in connection with Christian writers whom Clement deemed heretical, as Annewies van den Hoek noted.45 In all but two instances (where Clement approved of things said by Valentinus and Isidore),46 the quotations introduced by this formula are offered to dis-credit their authors by proving that they hold the shameful beliefs that he attributes to them. In these cases, he always noted the book from which he was quoting (even though he rarely did this elsewhere47) and frequently added biographical information about the author, such as his school and relation to other heretics.

In fifteen of the twenty-four occurrences of katå l°jin, Clement’s excerpt is our only source for that passage (excluding authors dependent

44. Aristobulus in Str. 1.22.150.1 (GCS 52:92–93); Dionysius Thrax in 5.8.45.4 (GCS 52:356); Plato in 5.14.98.5–8 (GCS 52:391) and 5.14.102.3–4 (GCS 52:395); Xenophon in 5.14.108.5 (GCS 52:399); Timaeus the Locrian in 5.14.115.4 (GCS 52:404); and Solomon in Fragment 36 (GCS 17:218–19).

45. Van den Hoek, “Techniques of Quotation,” 233.46. Str. 6.6.52.3–4 and 6.6.53.2–3 (GCS 52:458).47. Van den Hoek, “Techniques of Quotation,” 233.


on Clement).48 This leaves nine quotations that we can compare with other manuscripts or with modern critical editions as a way of gauging the gen-eral accuracy of Clement’s “verbatim” quotations. We cannot hope for precision here, for his manuscripts of those works would not have been identical to the ones on which our critical editions are based, and our criti-cal editions of Clement’s Stromateis are themselves based mainly on one highly corrupt eleventh-century manuscript.49 At best such comparisons reveal a minimal level of accuracy: his quotations might well have been more accurate than the comparisons suggest, but are unlikely to have been less accurate.

The details of these comparisons may be found in the appendix. They demonstrate that when Clement said he was quoting katå l°jin, he was reproducing, with a high degree of accuracy, the wording of a text. In one of the two instances in which the precise wording of his quotation was not important to his point (Str. 4.6.34.3–6), the agreement between the four sayings introduced by katå l°jin and the fourth edition of the United Bible Societies’ Greek New Testament (UBS4) averaged eighty-nine percent; the last of these four quotations was interrupted by an exegetical expansion. In the second instance where the precise wording has no evi-dent significance (Str. 1.7.38.8), Clement’s quotation of Clement of Rome is completely faithful to his manuscript. In the remaining seven instances, where the actual words of the passage in question mattered in some way to Clement’s argument, Clement’s accuracy ranged between ninety-four and 100 percent, with the exception of a quotation of Plato, the accuracy of which was eighty-two percent. The accuracy over all nine quotations,

48. No other source: Tatian in Str. 1.21.102.2 (GCS 52:65) and 3.12.81.1–2 (GCS 52:232); Aristobulus in 1.22.150.1–3 (GCS 52:92–93); Basilides’ son Isidore in 2.20.113.3–114.1 (GCS 52:174), 3.1.2.1–2 (GCS 52:195–96), and 6.6.53.2–3 (GCS 52:458); Carpocrates’ son Epiphanes in 3.2.8.1–3 and 3.2.9.2–3 (GCS 52:199–200); the Docetist Julius Cassian in 3.13.91.1–2 (GCS 52:238); the Valentinian Heracleon in 4.9.71.1–72.4 (GCS 52:280–81); Valentinus in 4.13.89.1–3 (GCS 52:287) and 6.6.52.3–4 (GCS 52:458); the grammarian Dionysius Thrax in 5.8.45.4 (GCS 52:356); and Hesiod in 6.2.26.3 (GCS 52:442). No other independent source: Timaeus the Locrian in 5.14.115.4 (GCS 52:404). The same quotation occurs in both Eusebius, Praepa-ratio evangelica 13.13.42 (GCS 43.2:216) and Theodoretus, Graecarum affectionum curatio 2.108 (Theodoreti Graecarum affectionum curatio, ad codices optimos denuo collatos, recensuit Iohannes Raeder, Bibliotheca scriptorum Graecarum et Romanorum Teubneriana [Leipzig: 1904; Stuttgart: B. G. Teubner, 1969], 65), but the former is explicitly dependent upon Clement; the latter, on either Clement or Eusebius.

49. For discussion of the manuscripts of Clement’s writings, see Cosaert, Text of the Gospels, 13–14.


totaling 355 words, is ninety-three percent. Each one of the quotations introduced with katå l°jin is markedly more accurate than the didactic quotation chosen for comparison, Clement’s rendering of Mark 10.17–31 (265 words), whose agreement with the UBS4 is seventy percent. We may conclude, then, that if faced with the task of proving that an inspired writ-ing did not contain discreditable statements, such as “naked man with naked man,” Clement would have proved their absence by consulting a manuscript and reproducing it very accurately.50 Accuracy of, say, ninety-five percent would result in roughly nine of the 181 words of LGM 1 and 2 differing from the exemplar; the differences would probably consist mainly of substitutions of synonyms, missing or added articles or prefixes, single-letter variants in spelling, and transpositions in word order, which are the sorts of things we see in his avowedly verbatim quotations. Such differ-ences would not perceptibly alter the character of the passage.

We now need to consider the second issue, which is whether Clement could have had access to a manuscript of the longer Gospel of Mark that was accurate enough to allow him to reproduce a pericope that sounds as typically Markan as the one in the Letter to Theodore. Carlson is right to raise the possibility that Clement’s manuscripts of the gospels were poor in comparison with our critical editions. Where the canonical Gospel of Mark is concerned, however, it is impossible to determine the overall quality of Clement’s manuscript(s) on the basis of the two quotations that can be ascribed with certainty to Mark.51 As Barbara Aland observed, the variants in Clement’s quotation of Mark 10.17–31 consist mostly of “banalities” that are unique to him and therefore unlikely to represent readings in his

50. It is conceivable that Clement copied the unique passages from the longer gos-pel into one of his own notebooks for his personal reference. If so, then the letter’s quotations from the immediate context in Mark might be the notations he made to tell himself where these passages are situated.

51. Apart from his long quotation from Mark 10, Clement quoted Mark 8.38 in Str. 4.9.70.2 (GCS 52:280) and four words from Mark 9.29 in Ecl. 15.1 (GCS 17:141). He might have quoted Mark elsewhere, but these are the only passages that are definitely not taken from their parallels in Matthew or Luke. Cosaert (Text of the Gospels, 262) rightly emphasizes that his own conclusions about the text-type of Clement’s manuscript of Mark cannot be generalized beyond Mark 10. Cosaert does not address the possibility that Clement’s text of Mark 10 appears “mixed” because Clement knew more than one manuscript of Mark (e.g., one he owned and one he heard read in church) or encountered verses of Mark 10 in other patristic writings. Clement might have been familiar with manuscripts of distinct text-types whose dis-tinctive readings are lumped together in his recollection.


exemplars. Such things are more readily attributed to faulty memory.52 The remaining textual variants that do have parallels in extant manuscripts provide very little from which to generalize about the quality of Clement’s manuscript(s) of Mark. Fortunately, we do not need to resolve this issue, because an expanded Gospel of Mark would have been a distinct gospel with a separate transmission history. The Letter to Theodore states that Mark bequeathed his more spiritual gospel “to the church in Alexandria, where it even yet is preserved with utmost discretion, being read only to those who are being initiated into the great mysteries” (1.28–2.2). The only way the elders in this church could have restricted its hearing to theo-logically advanced Christians is by making very few copies and ensuring that those copies do not circulate. If those conditions prevailed for most of the second century, then the Alexandrian manuscript(s) of the longer gospel should have been less corrupt than those of the widely distributed and more frequently copied canonical gospel. The fact that scribes were less apt to treat the gospels as sacrosanct during this period would exag-gerate the differences between the canonical and longer texts of Mark. Although it might seem logical for the Alexandrian church to correct its manuscripts of canonical Mark using the better-preserved longer edi-tion, the reality is that the Alexandrian church rarely corrected its gospel manuscripts using more accurate exemplars prior to the fourth century.53

52. Quoting Barbara Aland, “The Significance of the Chester Beatty Papyri in Early Church History,” in The Earliest Gospels: The Origins and Transmission of the Earli-est Christian Gospels—The Contribution of the Chester Beatty Gospel Codex P45, ed. Charles Horton, JSNTSup 258 (London and New York: T&T Clark International, A Continuum imprint, 2004), 120. The relevance of Aland’s discussion was noted in Cosaert, Text of the Gospels, 26 n. 21; his reasons for concluding that Clement cited Mark 10.17–31 from memory are presented on pp. 26 n. 21; 120–21 n. 63; 235.

53. Maurice A. Robinson, “In Search of the Alexandrian Archetype: Observations from a Byzantine-Priority Perspective,” in The New Testament Text in Early Christi-anity: Proceedings of the Lille Colloquium, July 2000, ed. Christian-B. Amphoux and J. Keith Elliott, Histoire du texte biblique 6 (Lausanne: Zèbre, 2003), 47–52. On pp. 51–52, Robinson wrote: “The present writer’s collation research within the pericope adulterae (where the potential for cross-comparison and correction was extremely high due to the wide amount of textual variation) indicates that cross- comparison and correction did not occur in a quantity sufficient to alter the otherwise relatively independent streams of textual descent. It thus would be no surprise to find that the early papyrus documents used as the basis of the Alexandrian archetype for each NT book were for the most part left uncorrected in regard to transcriptional error, particularly in cases where the resultant reading was sensible.” See also, Aland, “Significance,” 118: “the concern for an absolute degree of textual precision did not appear before the fourth/fifth century, and then it was chiefly among copyists of the Byzantine text.”


Consequently, it is entirely plausible that a relatively incorrupt longer text of Mark could have existed in Alexandria alongside less accurate copies of the canonical gospel.

THE LETTER’S TRADITION ABOUT MARK IS UNLIKE AUTHENTIC PATRISTIC TRADITIONS

Argument: This brings us to Carlson’s objections about the nature of the tradition about Mark the evangelist in the letter:

Clement inexplicably shares the same sensibilities and interests of a twentieth-century source critic: Theodore discloses far more details about the literary origin of the gospel of Mark than is typical for Clement’s time. . . . This kind of information comfortably fits into modernity’s proclivity for dissecting texts into their sources but is very different from the usual, early Christian etiologies for Mark, which are tantalizingly brief and usually stress the role of Mark’s memory. In fact, it even differs from Clement’s own explanation for the origin of Mark in his Hypotyposeis, “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,” an explanation that also emphasizes Mark’s memory, not his note-taking ability.54

Assessment: We need to address several questions. Is the letter’s inter-est in literary origins anachronistic? Would Clement have said this much about Mark’s compositional activity or imagined him using notes? Should an authentic tradition about Mark be brief and focused on his memory? We can best answer these questions by letting the sources speak for them-selves. Ancient writers in fact said a great deal about how they composed their works, and scholars are now beginning to synthesize these widely scattered comments and write monographs on the subject.55 One of the more detailed discussions of an ancient author’s compositional technique comes from Clement himself. Let us compare the procedure he used in composing his Stromateis with the one ascribed to Mark in the Letter to Theodore:

54. Carlson, Gospel Hoax, 58.55. Jocelyn Penny Small, Wax Tablets of the Mind: Cognitive Studies of Memory

and Literacy in Classical Antiquity (London and New York: Routledge, 1997); David S. Potter, Literary Texts and the Roman Historian (London and New York: Routledge, 1999); Tiziano Dorandi, Le stylet et la tablette: dans le secret des auteurs antiques (Paris: Les Belles Lettres, 2000).


This work is not a writing rhetorically shaped for exhibition. It is a collection of memoranda [ÍpomnÆmata], a treasure for my old age, a remedy against forgetfulness, a mere reflection, a sketch of vividly alive originals, words I was thought worthy to hear, and blessed and genuinely memorable men from whom I heard them. . . . There is a promise, not to give a full interpretation of the secrets—far from it—but simply to offer a reminder, either when we forget, or to prevent us from forgetting in the first place. I am very well aware that many things have passed away from us into oblivion in a long lapse of time through not being written down. That is why I have tried to reduce the effect of my weak memory, by providing myself with a systematic exposition in chapters as a salutary aide-mémoire [mnÆmhw ÍpÒmnhma]; it has necessarily taken this sketchy form. There are things which I have not even recorded—these blessed men were endowed with great power. There are others which remained with me with no need of noting, but with the passage of time have now escaped me. Others were growing faint to the point of extinction in my mind, since service of this kind is not easy for those who are not qualified experts. These I took good care to rekindle by making notes [ÍpomnÆmasi]. Some I am deliberately putting to one side, making my selection [§kl°gvn] scientifically out of fear of writing what I have refrained from speaking—not in a spirit of grudging [oÎ t¤ pou fyon«n] (that would be wrong), but in the fear that my companions might misunderstand them and go astray and that I might be found offering a dagger to a child (as those who write proverbs put it).56

As for Mark, then, during Peter’s stay in Rome he wrote an account of the Lord’s doings, not, however, declaring all of them, nor yet hinting at the mystic ones, but selecting [§klegÒmenow] what he thought most useful for increasing the faith of those who were being instructed. But when Peter died a martyr, Mark came over to Alexandria, bringing both his own notes [ÍpomnÆmata] and those of Peter, from which he transferred to his former book the things suitable to those studies which make for progress toward knowledge. Thus he composed a more spiritual gospel for the use of those who were being perfected. Nevertheless, he yet did not divulge the things not to be uttered, nor did he write down the hierophantic teaching of the Lord, but to the stories already written he added yet others and, moreover, brought in certain traditions of which he knew the interpretation would, as a mystagogue, lead the hearers into the innermost sanctuary of that truth hidden by seven veils. Thus, in sum, he prepared matters, neither grudgingly [oÈ fyoner«w] nor incautiously, in my opinion, and, dying . . .57

Both accounts depict a former devotee (one a student of Christian teach-ers, the other a follower of an apostle) composing a theological work

56. Str. 1.1.11.1, 14.2–3 (GCS 52:8, 10–11; Ferguson, Clement, 30, 32–33).57. Letter to Theodore 1.15–28.


for advanced Christians by selecting appropriate materials from his (and his teacher’s) memoranda, prudently—but not grudgingly—setting aside teachings that should not be put in a book. In other words, the Letter to Theodore envisions Mark following the same basic procedure that Clem-ent followed in producing a similar kind of book.

As Clement’s comments aptly illustrate, “notes” were not substitutes for memory but aids to memory—the means by which authors kept their memories alive—as the word ÍpomnÆmata itself implies (from ÍpÒmnhsiw, the act of recalling something to mind). The letter says nothing about Mark’s “note-taking ability.” It indicates that Mark preserved his recol-lections using notes, which was the standard procedure in antiquity. The creation and utilization of notes based on memory, interviews, books, inscriptions, and public records was common among ancient historians and biographers.58 Students of philosophers kept notes of their teachers’ lectures, as did Clement.59 Authors regularly made extracts of the parts of books they might later wish to quote. Some even published their favorite extracts in anthologies (florilegia). The presumably posthumous eighth book of Clement’s Stromateis shows us what his own notes looked like before he reworked them for publication, as do his Excerpta ex Theodoto and Eclogae propheticae.

Although the notion of Mark using notebooks might seem novel to some, there are good reasons to suppose that Christians were using them from the very beginning. “The diversity of text-types that appear in Paul’s quotations” of scripture leads Christopher D. Stanley to postulate, for instance, that Paul compiled a “growing collection of biblical excerpts” in a notebook “from a variety of manuscripts housed at sites all around the eastern Mediterranean world, where he was a constant traveler.”60 This conclusion is quite reasonable, for, as Graham Stanton points out, it

58. On the whole process by which ancient historians produced their works, see Potter, Roman Historian, 106–17. Note his comments on “the ubiquity of notes” in the compositional process (p. 113). See also George A. Kennedy, “Classical and Christian Source Criticism,” in The Relationships among the Gospels: An Interdis-ciplinary Dialogue, ed. William O. Walker Jr., Trinity University Monograph Series in Religion 5 (San Antonio, TX: Trinity University Press, 1978), 125–55, at 139–41; and van den Hoek, “Techniques of Quotation,” 225–27.

59. Kennedy, “Classical and Christian Source Criticism,” 131. For a convenient summary of the evidence for note-taking, see Christopher D. Stanley, Paul and the Language of Scripture: Citation Technique in the Pauline Epistles and Contemporary Literature (Cambridge: Cambridge University Press, 1992), 74–78.

60. Stanley, Language of Scripture, 67–74, 77–79, citing pp. 78, 74.


would be rather cumbersome and expensive for Christian missionaries to carry scrolls of their favorite scriptures on their journeys, but very easy for them to bring notebooks containing testimonia: scriptural passages which they interpreted as prophecies attesting to Christ. Stanton argues further that Christian letter-writers likely used notebooks to prepare drafts of their letters and to keep their own copies. The pseudonymous author of 2 Timothy imagined Paul keeping such notebooks and therefore made him instruct Timothy to retrieve for him his parchment notebooks (4.13). Notebooks are also the most obvious form in which Christians would have kept collections of Jesus’ sayings. Very early and widespread use of codex notebooks among Christians would go a long way toward explaining the extraordinary fact that most of the earliest extant Christian manuscripts of the gospels and other scriptures are in the form of the codex, which was the standard format for notes, rather than the scroll, which was the standard format for finished books.61 In light of the fact that the Letter to Theodore depicts Mark following the standard procedure used in his day for composing texts, the fact that this procedure accords with the conclu-sions about written sources that modern scholars deduced from the gos-pels themselves can hardly be deemed suspicious.

What, then, of Carlson’s observation that early traditions about the creation of the Gospel of Mark “are tantalizingly brief and usually stress the role of Mark’s memory”? That claim is simply untrue. Carlson offers two patristic traditions to substantiate this generalization, namely, Papias’s comments in his lost Interpretation of the Oracles of the Lord, recorded in Eusebius’s Church History 3.39.15 (LCL 153:296), and Clement’s com-ments in his lost Hypotyposeis, recorded in Eusebius’s Church History 6.14.6–7 (GCS 17:197; LCL 265:48). I am not sure how these two tradi-tions could be described either as brief or as typical in their references to Mark’s memory, but Carlson creates that impression by quoting only one sentence from each62 and neglecting a wealth of traditions about Mark

61. Graham N. Stanton, Jesus and Gospel (Cambridge: Cambridge University Press, 2004), 177–78, 182–89. On Paul’s likely use of a codex notebook to keep copies of his letters, see Richards, “Codex.”

62. Carlson’s reproductions of the Greek texts and English translations of Papias and Clement appear in Gospel Hoax, 58, 121–22 n. 39 and n. 40. With no indication of omission, Carlson quoted twenty of the seventy-six Greek words from Papias, and nineteen of the forty-nine Greek words from Clement. Carlson’s extract from Papias is limited to Papias’s statement of what “the Presbyter [John] used to say,” wherein the words ˜sa §mnhmÒneusen (“all that he remembered”) might actually refer to Peter’s memory. The subsequent explanation of what the elder meant, which is more


that do not refer to his memory, including Irenaeus’s Against Heresies 3.1.1, Clement of Alexandria’s two other extant traditions about Mark (discussed below), Origen’s first commentary on the Gospel of Matthew (cited in Eusebius, Church History 6.25.3–6 [5]), John Chrysostom’s Hom-ily 1.7 and Homily 4.1, Epiphanius’s Panarion 51.6.10–13, the so-called Anti-Marcionite Gospel Prologue, the Monarchian Gospel Prologue, and Jerome’s Famous Lives 8 and preface to Commentary on Matthew.63 Not only do these other traditions about the evangelist Mark say nothing about his memory, they also vary considerably in length, ranging from a few words to a paragraph or more. How much an author said about Mark had a great deal to do with his purposes in saying it. In accordance with the letter-prefaces discussed by Jay, the Letter to Theodore attempts to combat the circulation of inauthentic versions of a text by explaining how both the legitimate and the illegitimate editions were produced.64 That purpose dictates more than the usual amount of attention to Mark’s literary activities.

Carlson’s generalizations about authentic patristic traditions inexplica-bly omit the fact that Clement related two other traditions about Mark’s composition of his gospel, neither of which mentions Mark’s memory. One of these other traditions was recorded in the same book that Carlson cites,

plausibly attributed to Papias, is completely omitted from Carlson’s excerpt. The full tradition related by Papias is as follows:

And the Presbyter used to say this, “Mark became Peter’s interpreter and wrote accurately all that he remembered, but not, indeed, in order, of the things said or done by the Lord.[”] For he had not heard the Lord, nor had he followed him, but later on, as I said, followed Peter, who used to give teaching as necessity demanded but not making, as it were, an arrangement of the Lord’s oracles, so that Mark did nothing wrong in thus writing down some points as he remem-bered them. For to one thing he gave attention, to leave out nothing of what he had heard and to make no false statements in them.[”] (LCL 153:297; trans. Kirsopp Lake; the bracketed quotation marks indicate uncertainty over where the Presbyter’s comments end)

63. For the text of these traditions in English translation, see C. Clifton Black, Mark: Images of an Apostolic Interpreter (Columbia, SC: University of South Caro-lina Press, 1994), on which I am relying in this section. One might argue that Justin’s references to the gospels as “reminiscences” (épomnhmoneÊmata) implies an emphasis on memory. However, Justin offers no tradition about Mark the evangelist, and in the only place where Justin appears to be referring specifically to the Gospel of Mark, the antecedent of “his reminiscences” is Peter (Dial. 106.3). There is no mention of Mark in this context.

64. See Jay, “Epistolary Framework.”


Clement’s Hypotyposeis. Both traditions from this book were excerpted by Eusebius in his Church History. The third tradition was preserved in Latin by Cassiodorus. They read as follows:

When Peter had preached the word publicly at Rome, and by the Spirit had proclaimed the gospel, those present, who were many, implored Mark (as one who had followed him for a long time and remembered what had been spoken) to record what was said; and he did, and shared the gospel among his petitioners; and when the matter came to Peter’s knowledge, he neither actively prevented nor promoted it.65

So brilliant was the light of piety that shone on the minds of Peter’s hearers that they were not satisfied with only a single hearing or with the unwritten teaching of the divine proclamation; but with all sorts of appeals earnestly entreated Mark, whose gospel is extant, who was Peter’s follower, to leave them in writing a memorandum of the teaching relayed to them by word; nor did they let up until they had prevailed upon the man, and thus they became the occasion of the scripture called the gospel according to Mark. And they say that the apostle, who knew through spiritual revelation to him what had been done, was delighted with the men’s zeal, and authorized the scripture for concourse in the churches. Clement cites the story in the sixth book of the Hypotyposeis, and the bishop of Hierapolis, named Papias, corroborates his witness, which calls to mind Peter’s reference to Mark in the first Epistle.66 This also, they say, was composed in Rome itself, as signified by his own figurative reference to the city as Babylon in this way: “The co-elect in Babylon sends you greetings, as does my son Mark.”

This Mark, who was the first, they say, to be dispatched to Egypt to preach the gospel which he also composed, was in addition the first to establish churches in Alexandria itself.67

Mark, Peter’s follower, while Peter was preaching publicly the gospel at Rome in the presence of certain of Caesar’s equestrians and was putting forward many testimonies concerning Christ, being requested by them that they might be able to commit to memory the things that were being spoken, wrote from the things that were spoken by Peter the gospel that is called, “According to Mark.”68

65. Clement, Hypotyposeis, as cited in Eusebius, Church History 6.14.6–7 (GCS 17:197; Black, Mark, 138).

66. Some scholars view the subsequent information as dependent upon Papias and anonymous tradition; I consider it more likely that “they say” continues to refer explicitly to Papias and Clement. See Brown, Mark’s Other Gospel, 59–60.

67. Clement, Hypotyposeis, as cited in Eusebius, Church History 2.15.1–16.1 (LCL 153:142, 144; Black, Mark, 138, 157).

68. Clement (possibly also the Hypotyposeis), as preserved in Cassiodorus, Adum-brationes ad 1 Peter 5.13 (GCS 17:206; Black, Mark, 139).


Only the first of these three traditions mentions Mark’s memory—in pass-ing. The focus in all three traditions is on Mark’s writing activity, as it is in Eusebius’s lengthy quotation from Papias (see n. 62) and in the Letter to Theodore.

When all of Clement’s undisputed traditions about Mark are taken into account, the problem with objecting that the Letter to Theodore “dif-fers from Clement’s own explanation” becomes obvious, for the three accounts are not entirely harmonious. They explicitly disagree on Peter’s reaction to Mark’s writing and on whether the persons who petitioned him to leave a written record consisted of “those present, who were many” or only “Caesar’s equestrians.” Despite these discrepancies, few scholars deny that these traditions were related in (substantially) these forms by Clement, for we are dealing with received traditions about Mark’s writ-ing, not with Clement’s personal opinions. The fact that these traditions are in conflict suggests that Clement (and Eusebius) did not attempt to eliminate the differences.

C. Clifton Black has written the best study to date of the patristic tra-ditions about Mark. He notes “five features that recur in most of the tra-ditions surveyed”:

(1) Typically, Mark is portrayed as a “literary figure” or evangelical author, the conveyor of oral proclamation, possibly in the form of preliminary memoranda (see Papias, perhaps, and Clement), ultimately in the form of the Second Gospel of the Christian canon (e.g., Adamantius and Jerome). (2) Normally, Mark is associated with the apostle Peter, often as the latter’s adherent, interpreter, or transcriber. (3) Usually, the linkage of Mark with an apostle reflects widespread patristic concerns about the fidelity of the Markan tradition to the church’s proclamation of Jesus. (4) Commonly, Rome (in particular), Italy (more generally), is identified as a locale either for Mark’s literary activity or for his association with Peter. (5) Once introduced into its traditions (at least as early as the late second century), an association of Mark, though usually not Mark’s [canonical] Gospel, with Alexandria is typically maintained in Eastern Christianity.69

The Letter to Theodore displays all of these typical features, as Black’s discussion demonstrates. In his opinion, the letter “enriches . . . Clem-ent’s other, better-known statements.”70 Of particular interest are Black’s observations (derived from George Kennedy) about the reference to a

69. Black, Mark, 185–88. I removed the italics and combined the initial sentences of his discussions of these points into a list.

70. Black, Mark, 140.


memorandum (ÍpÒmnhma) in the Hypotyposeis and its possible connection to the notes (ÍpomnÆmata) mentioned in the Letter to Theodore. The for-mer tradition tells us that Mark’s hearers compelled him to “leave them in writing a ÍpÒmnhma of the teaching relayed to them by word” and that their efforts “became the occasion of the scripture called the Gospel according to Mark.” What intrigues Black is the fact that the ÍpÒmnhma is not clearly equated with the gospel. Indeed, the gospel itself is mentioned incidentally in this context in order to identify the Mark whose literary activity this tradition recounts, which is rather odd if this tradition is about his writ-ing of that gospel. Black thinks it likely that the term ÍpÒmnhma denotes something preliminary to the gospel, noting that the four-stage procedure of listening to eyewitnesses, gathering their recollections into notes, fash-ioning those notes into an ordered narrative, and finally polishing that narrative with beauty and eloquence was recommended by the second-century writer Lucian of Samosata.71 In this scheme, the word ÍpÒmnhma denotes the unordered “notes,” described as “a body [of material] still charmless and disjointed,” and the word tãjiw describes the preliminary draft. Black also calls attention (again, following Kennedy) to the possibil-ity that Papias’s tradition about Mark may not actually describe Mark’s creation of his gospel but of the preliminary notes. Like the tradition of Clement that refers to a memorandum, that of Papias is introduced not as a tradition about how Mark wrote his gospel but as a tradition “about Mark, the one who wrote the Gospel.”72

Certainly the elder’s statement that Mark wrote “accurately all that he remembered, but not in order, of the things said or done by the Lord” could describe an unordered notebook or memorandum. In fact, if we take the words oÈ tãjei seriously, the statement should denote something without a narrative order, such as a notebook, for as Alistair Stewart-Sykes noted, “Papias does not say that Mark’s taxis is poor [or wrong], he says that there is none; and taxis is absent because none is to be expected in a collection of chreiai.”73 One must ask, though, why Papias would

71. Lucian of Samosata, How to Write History 47–48 (LCL 430:60). Black, Mark, 141–42; 173 nn. 15, 18; 178 n. 53; Kennedy, “Classical and Christian Source Criti-cism,” 147–52.

72. Black, Mark, 90.73. Alistair Stewart-Sykes, “Tãjei in Papias: Again,” JECS 3 (1995): 487–92, at

490, 492. Contrary to my use of his arguments, Stewart-Sykes was not supposing that Papias was referring to a collection of notes, but that Papias thought of Mark itself as “a collection of chreiai” with “no arrangement, chronological or otherwise” (p. 491), a position that I find implausible given the numerous chronological markers in


be interested in the lack of order of Mark’s notes. It seems more likely that Papias would be concerned with defects in the sequence of materials in Mark’s gospel. Yet the two issues could be connected. Whereas Ken-nedy read the whole excerpt as referring to Mark’s notes, I suggest that where Papias concluded that “Mark did nothing wrong in thus writing down some things as he remembered [or recorded] them [¶nia grãcaw …w épemnhmÒneusen],”74 Papias was now talking about Mark’s fashioning of a connected narrative from his unordered and incomplete notes (or from the memories those notes sustained). The first verb refers to the prelimi-nary draft of the Gospel of Mark; the second, to the notes (or memories) that formed its basis. Papias’s reasoning would be as follows: since Peter used to recount traditions about Jesus anecdotally in his teaching without attempting to produce an ordered account (sÊntajin) of the Lord’s oracles (log¤vn), Mark did not learn the correct order (whether chronological or artistic) from Peter. Mark therefore cannot be faulted for reproducing some things in the fortuitous order in which they occurred in his notes, for it was very important to him to leave nothing out (and to state nothing false).75 As additional support for this interpretation, one may note that Papias’s comments about Matthew’s literary activity, which follow Eusebius’s excerpt on Mark, do not seem to describe the Greek Gospel of Matthew as we know it but could well describe a preliminary, albeit ordered, work that was used by others: “Matthew systematically arranged the oracles [tå lÒgia sunetãjato] in the Hebrew language [ÑEbra˝di dial°ktƒ], and each interpreted them as he was able.”76

If, as seems likely, the term “memorandum” in Clement’s Hypotyposeis does in fact denote an unordered set of notes, then the letter adds little new information to Clement’s conception of Mark’s compositional pro-cedure. It does add that Mark excerpted from his notes in order to create his two gospels, which is an ordinary procedure. And it tells us that he used Peter’s notes along with his own (1.19–20), which is a bit surprising.

Mark (especially in the “seams” between pericopae) and its preference for materials that are longer than what is ordinarily understood by the word chreia as a literary-critical term.

74. On the translation of ¶nia, see Terence Y. Mullins, “Papias on Mark’s Gospel,” VC 14 (1960): 216–24. I am not convinced by the argument he based on this translation.

75. If Papias is contrasting Mark’s order with Matthew’s, the materials that Papias supposes are out of order in Mark might be the ones that occur in a different place in Matthew.

76. Eusebius, Church History 3.39.16; Black, Mark, 83.


The uniqueness of this reference to Peter’s notes does not make it dubious, for all of Clement’s traditions about Mark contain some unique informa-tion. The notion that Mark used two sets of notes was not, moreover, congenial to Morton Smith’s hypothesis that the authors of the canoni-cal Gospel of Mark, the longer Gospel of Mark, and the Gospel of John shared a single structured proto-gospel, a hypothesis that explains the extensive sequential parallels Smith detected between Mark 6.32–15.47 and John 6.1–19.42. Only a sequentially organized source such as a gospel or proto-gospel could explain how the story of the raising of the dead man in Bethany appears in the same place within the chronological frameworks of both longer Mark and John. Hence, Smith did not appeal to the letter’s reference to sets of notes as support for that theory or any of his other historical reconstructions.77 Instead, in his commentary under the word ÍpomnÆmata he conjectured that the letter’s reference to Peter’s notes helped substantiate the Alexandrian church’s claim to possess secret traditions going back to the apostles.78

THE LETTER TO THEODORE CONTRADICTS CLEMENT’S USE OF SALT AS A METAPHOR

Argument: There is one last argument to address in Carlson’s case against Clement’s authorship of the letter. If I understand him correctly, Carlson contends that the letter and the Stromateis contain contradictory elabo-rations of the second half of Matthew’s saying, “You are the salt of the earth; but if salt has lost its taste, how shall its saltiness be restored?” (RSV). The Letter to Theodore 1.13–15 compares the mixture of true and false things in the Carpocratian longer gospel with salt that has lost its savor, thereby associating loss of salt’s savor with the addition of false teachings. Str. 1.8.41.3–4, on the other hand, compares Christians who have left the church with fish that grew up in salt water yet still “need salt in their preparation,” thereby associating loss of salt’s savor with separa-tion from a Christian environment. “The interpretation of Theodore is therefore un-Clementine.”79

Assessment: The premise that there is something unusual and therefore suspicious about a Christian author applying a biblical metaphor to more than one phenomenon is curiously naive. It is easy to find examples of this

77. See Smith, Clement, 158–63.78. Smith, Clement, 28–29. 79. Carlson, Gospel Hoax, 59–60.


sort of thing in Clement’s undisputed writings. Consider, for instance, the metaphor “the light of the world,” which is set in poetic parallelism with the metaphor “the salt of the earth” in Matt 5.13–14. Clement interpreted this light metaphor at least twice. In Str. 4.11.80.3 (GCS 52:283) he said that “the light of the world” denotes Christian faith as exhibited in mar-tyrdom, whereas in Quis dives salvetur 36 (GCS 17:183) he said that this metaphor denotes “the elect of the elect” (the most advanced Christians whom he called the true gnostics), for whose sake “both the visible and invisible things of the world have been created.” The applications are dif-ferent and inconsistent, but given the multivalent nature of metaphorical language, this is not surprising.

In addition to its false premise, the argument supporting the conclu-sion that the letter’s salt metaphor is “un-Clementine” is predicated on a misreading of the salt passage in the Stromateis. John Ferguson’s transla-tion reads as follows:

That is the nature of these contentious fellows, whether they follow a sect or whether they practice dialectical craft. These are people who, in the words of Scripture, stretch the warp without weaving anything. They set their hearts on useless labor, which the Apostle [Paul] called human cunning and craftiness “directed to leading people astray” [Eph 4.14]. “There are many people,” he says, “who refuse to take instructions, talk nonsense and spread false ideas” [Titus 1.10]. So everyone was not being addressed in the words “You are the salt of the earth” [Matt 5.13]. Some of the hearers of the Word are like fish of the sea. They have grown up in saltwater from birth and even so need salt in their preparation.80

Carlson paraphrases the last two sentences as follows: “Just as a fish taken out of salt water does not taste salty and still needs to be salted, so do people brought up in the teaching of the church lose their saltiness when they are removed from it.” Carlson’s inference that this passage describes a loss of salt contradicts the logic of what came before. Clement’s quota-tion of Titus 1.10 describes headstrong Christians who will not submit to Christian instruction. Such Christians, Clement claims, are not among the people Jesus had in mind when he said, “You are the salt of the earth,” because, unlike “the salt of the earth” (and “the light of the world”), they do not have a positive effect on others. As persons explicitly excluded from the category “the salt of the earth,” the stubborn Christians have not lost their saltiness, for they cannot lose a quality they never possessed. Thus

80. Str. 1.8.41.2–4 (GCS 52:27; Ferguson, Clement, 52–53).


the second half of this saying, “but if salt has lost its taste, how shall its saltiness be restored?” logically does not apply to them either. It can apply only to Christians who once possessed this quality, that is, to people who once were “the salt of the earth.” In Clement’s interpretation, “the salt of the earth” are the true gnostics whom he distinguished from “all the faithful” in Quis dives salvetur? 36 as being “more elect than the elect.” The general point of the salt metaphor is therefore clear: although these stubborn Christians have grown up in a “salty” environment (i.e., they have been educated by true Christian teachers), they do not possess this quality themselves and therefore need to be “salted.”81

Because the recalcitrant individuals whom Clement likened to fish in Str. 1.8 are not the salt of the earth, it follows that Clement was not inter-preting the same biblical metaphor “if salt has lost its taste . . .” that is alluded to in the Letter to Theodore 1.13–15. Nor is it likely that he was describing what happens to Christians when they are removed from a “salty” (Christian) environment. Like the author of Titus, Clement was talking about members of the church who resist authority and engage in sophistry and speculative thought, substituting human machinations for sound doctrine. The notion that these “hearers of the Word . . . need salt in their preparation” suggests that they remain within the true church (otherwise they would not be “hearers of the Word”) and within the orbit of God’s grace but require special attention (the imagery of salting and of preparation imply improvement rather than abandonment), either from supernatural forces or from Christian teachers.

It is the negative in the clause “So everyone was not being addressed in the words ‘You are the salt of the earth’” that demonstrates that Clem-ent was not comparing the recalcitrant Christians to salt that has lost its savor. Since Carlson reproduced Ferguson’s translation of Str. 1.8.41.3–4 before he interpreted it, how is it that this problem escaped the notice of his readers? The answer is quite simple. Carlson omitted the word not (oÈ) from this clause (oÎkoun oÈ pçsin e‡rhtai: “Íme›w §ste ofl ëlew t∞w g∞w”), thereby changing the logic of this passage. Due to this inconspicuous omis-sion, it now conveys the notion that because all Christians are the salt of the earth, those who, like fish, still need to be salted in their preparation have lost their salt. In other words, this mistake supplied the basis needed for comparing this passage and the words “even the salt loses its savor”

81. The meaning of “their preparation” remains unclear. Perhaps Clement had in mind Mark 9.49 (“For everyone will be salted with fire”) and was picturing some kind of eschatological purification through divine “fire” or testing.


in the Letter to Theodore. It is certainly remarkable that this error was not corrected, for it turns Clement’s deduction concerning “the salt of the earth” into a non sequitur (some Christians are recalcitrant; therefore all Christians are good) and spawns a contradiction with Clement’s opinion that “the salt of the earth” refers only to the true gnostics (q.d.s. 36), a passage which Carlson quotes a few pages after telling us that Clement would not contradict his use of salt imagery.82

CONCLUSIONS

As is true of Carlson’s evidence connecting the Letter to Theodore with Morton Smith, his reasons for denying Clement’s authorship of the letter do not bear scrutiny. In most cases, the very things that Carlson asserts are uncharacteristic of Clement or of an authentic ancient letter are precisely what one might expect of a real letter by Clement, including the opening comments identifying the Carpocratians as heretics predicted in scripture, the tactic of refuting heretical exegesis by quoting the disputed verses, dis-cussing their literary context, isolating interpolations, and providing the true explanation, the apparently greater accuracy in reproducing the dis-puted excerpts in comparison to Clement’s didactic gospel citations, and the detailed picture of Mark creating a text for theologically advanced students by carefully selecting appropriate materials from his “notes.” In the remaining cases, the significant omissions in Carlson’s presentation of the evidence create a false impression that the letter differs from Clem-ent. Thus, the claim that the letter’s salt metaphor is “un-Clementine” is founded on the omission of the word “not” in Str. 1.8.41.3, and the claim that the letter’s tradition is too long and lacks the expected reference to Mark’s memory is established by omitting all of Papias’s own comments about Mark, all but one sentence of Clement’s three undisputed traditions about Mark, and all other early traditions about Mark.

Scott G. Brown lives in Barrie, Ontario, where he is working as a freelance copy editor

82. Carlson, Gospel Hoax, 62–63.


APPENDIX: THE ACCURACY OF CLEMENT’S VERBATIM QUOTATIONS

This appendix describes in more detail how Clement’s katå l°jin quotations compare with the nine instances in which the same passage has been preserved independently by other authors. My method of quantifying agreements is heuristic rather than scientific. I counted the number of individual letters in Clement’s katå l°jin quotations that agree with the comparison text, divided that number by the total number of letters in Clement’s quotation (plus the letters of any missing words, unless Clement was intentionally abbreviating), and multiplied by 100 in order to produce a percentage. Where differences in word order occur, I did not treat the affected words as absent, since they are merely in a different place and such reversals rarely affect the meaning, although they sometimes affect the emphasis.83 I did count as differences variations in spelling. This method is certainly not perfect, but is adequate for our purposes.

The first instance of katå l°jin introduces Clement of Rome’s First Epistle to the Corinthians 48.5 (Str. 1.7.38.8; GCS 52:25; sixteen words/seventy-seven letters). Ignoring accents, this quotation agrees with the critical text produced by Annie Jaubert in every respect except the word ègnÒw, for which Clement has gorgÒw (SC 167:178, 180). Since, however, his quotation of the same verse in Str. 6.8.65.3 (GCS 52:464) also has gorgÒw in this position, that was probably the reading of his manuscript. Accordingly, the quotation qualifies as exact. The next instance of katå l°jin introduces the first half of Acts 2.41 (Str. 1.18.89.4; GCS 52:57; eight words/forty-five letters). Clement’s quotation agrees 100 per-cent with the UBS4. Clement’s quotation of the Epistle of Barnabas 16.7 (Str. 2.20.116.3–4; GCS 52:176; thirty-seven words/173 letters) differs from the critical text of Robert A. Kraft by six letters.84 Specifically, it is missing the word …w, the second iota in daimon¤vn, and the prefix to katoikhtÆrion (SC 172:192). The agreement is roughly ninety-seven percent.

The expression katå l°jin next introduces a string of four synoptic sayings in Str. 4.6.34.3–6 (GCS 52:263). Whether Clement intended this expression to characterize all four sayings is unclear. The first saying is Luke 12.15 (seventeen words/eighty-four letters). The agreement between Clement’s quotation and the USB4 is eighty-four percent. The differences consist of the addition of to¤nun, the absence of §k and aÈt“, two differences in case resulting from the absence of §k, and the inversion of phrases around an equative verb. The second saying is Matt 16.26 (twenty-two words/106 letters). Its divergence from USB4 is lim-ited to a different form of the verb »felhyÆsetai, resulting in an agreement of

83. On the pervasiveness and insignificance of differences in word order within ancient quotations, see Whittaker, “Art of Misquotation,” 72–76, who conjectures that authors usually made these changes in their quotations deliberately in order to make familiar passages sound fresh.

84. As a point of comparison, our earliest manuscript witness for the Epistle of Barnabas, Codex Sinaiticus, differs from the critical text by eight letters.


ninety-eight percent. The third saying, Luke 12.22–23 (twenty-seven words/113 letters), omits one word (Ím›n), adds another (Ím«n), has peribãlhte instead of §ndÊshsye (synonymous verbs), and contains single-letter differences in the spelling of ple›on (omega instead of omicron) and §st¤n (no nu). The concordance with UBS4 is eighty-three percent. The accuracy of the fourth saying, Matt 6.32b–33 (twenty-two words/117 letters), is difficult to determine. The words t«n oÈran«n following tØn basile¤an disagree with USB4 but are included by Augustine and Cyril and occur in Paedagogus 2.12.120.2 (GCS 12:229), so it is likely that they appeared in one of Clement’s manuscripts of Matthew. Clement’s omission of ı oÈrãniow likewise agrees with known manuscripts (a* 28) and with his citation of these words in Paed. 2.10.103.5 (GCS 12:219) and Eclogae propheticae 12.2 (GCS 17:140). The words aÈtoË ka¤ are missing after tØn dikaiosÊnhn. After this point, Clement’s quotation is interrupted by an exegetical expansion that led him to paraphrase part of the ending of this saying (“and all these things” becomes “the things that are small and pertain to this life”) and omit the word pãnta. The quotation resumes with “shall be added to you.” The full ending without the paraphrase occurs in Paed. 2.12.120.2 and Ecl. 12.2. One’s impres-sion of the accuracy of this saying will depend on whether one supposes, as I do, that Clement’s manuscript included t«n oÈran«n and omitted ı oÈrãniow. That is, the agreement could be described as being as high as eighty-nine percent or as low as seventy-six percent. The average accuracy over the total number of words in this string of quotations is eighty-nine percent. Cosaert’s analysis of these four quotations is correct: they are noticeably “more exact” than is usual for Clement’s gospel quotations but still “reveal a slight degree of modification” in comparison to Clement’s non-gospel NT quotations that have comparable introductory formulas.85

The next two instances of katå l°jin occur in quotations of Plato. Clement’s quotation of Theaetetus 173c–174a (Str. 5.14.98.5–8; GCS 52:391; 120 words) is very accurate. Of its 573 letters, twenty-six differ from the critical text produced by E. A. Duke et al., and ten letters from the critical text are missing.86 The agree-ment is ninety-four percent. The important difference is that Clement’s quotation is abbreviated. It omits ten phrases of two or more consecutive words. Similar observations apply to Clement’s quotation from Plato’s Epistle to Hermeias and Erastus and Coriscus (Str. 5.14.102.3–4; GCS 52:395; twenty-nine words/149 letters). Twenty-one letters differ from the critical text edited by John Burnet, and eight are missing. The agreement is eighty-two percent.87 More specifically, Clement’s quotation has two words not present in the standard text (a‡tion ka¤), two substitutions of synonyms (Ùry«w instead of ˆntvw; §ãn instead of ên), and two verbs that differ in person. It omits one seven-word phrase and three

85. Cosaert, Text of the Gospels, 28–29.86. E. A. Duke et al., eds., Platonis opera, Scriptorum classicorum bibliotheca oxo-

niensis (1900; reprint, New York: Oxford University Press, 1995), 1:325–26.87. John Burnet, ed., Platonis opera, Scriptorum classicorum bibliotheca oxoniensis

(Oxford: Clarendon Press, 1905), vol. 5.2 at Epistolae 323d.


isolated words (ka¤, t«n, te). So again Clement was reproducing the wording in a manuscript but in a condensed form. Clement’s objective in these places was to demonstrate parallels to Christian ideas in Plato’s writings, so it was necessary to quote these passages fairly accurately, but not in their entirety.88

It might have surprised Clement to learn that his “word for word” quotation from Xenophon’s Memorabilia 4.3.13–14 (Str. 5.14.108.5; GCS 52:399; forty words) is not from the original work. It is likely from an anthology in which the same passage was imitated.89 The same text, however, was incorporated into the Anthologus of Joannes Stobaeus (2.1.33), so we are able to compare Clement’s quotation with that of Stobaeus.90 Clement’s quotation contains one substitution (dÉ §st¤n instead of d¢ tÆn) and an extra moveable nu. It also omits the word …w and puts the word aÍtÒn (sic) in a different place. Fortunately for us, Clement used the same quotation in his Prot. 6.71.3 (GCS 12:54); by comparing the two, we can determine which of these differences derive from his exemplar. The location of aÍtÒn, the absence of …w, and the optional nu are the same in both of Clement’s quotations and may reasonably be ascribed to Clement’s exemplar. Where the Stromateis has dÄ §st‹n instead of d¢ tÆn, the Protrepticus has d° tiw, indicating that Clement’s manuscript was defective and possibly indecipherable at this point. This leaves a difference of three letters in the Protrepticus that are attributable to either Clement or a subsequent scribe (aÈtÒw instead of otow and the addition of tã before pãnta), and no differences in the Stromateis. The entire quotation consists of 192 letters in both the Stromateis and the Protrepticus. Thus the latter quotation is at least ninety-seven percent accurate, and the former is as accurate as it could be.

The last two “verbatim” excerpts that we can check are Clement’s brief quota-tion of Homer, Iliad 17.53 in Str. 6.2.26.1 (GCS 52:442; seven words/thirty-five letters), which agrees in every respect with the Greek text of the Loeb Classical Library (171:234), and his quotation of 2 Chron 6.18 (Fragment 36; GCS 17:218; ten words/forty-six letters), which agrees completely with the Septuagint except for the addition of êra. Its accuracy, therefore, is roughly ninety-four percent.

The accuracy over all nine quotations (355 words) is (at least) ninety-three percent. We can compare these results with Clement’s quotation of Mark 10.17–31 in Quis dives salvetur 4.4–10 (GCS 17:162–63; 265 words/1309 letters), which is the passage most often cited as evidence of Clement’s propensity to loose quotation. Its agreement with the UBS4 is seventy percent.

88. It is also possible that Clement was relying on anthologies in which these pas-sages from Plato were abbreviated or that he was quoting these passages from his notes, wherein he had already condensed these passages.

89. Alain Le Boulluec, ed. and trans., Clément D’Alexandrie, Les Stromates. Stro-mate V, t. II, commentaire, bibliographie et index, SC 279 (Paris: Éditions du Cerf, 1981), 330–31.

90. O. Hense and C. Wachsmuth, eds., Joannis Stobaei Anthologium (Berlin: Weidmann, 1884), 2:15.

The Letter to Theodore: Stephen Carlson’s Case against Clement’s Authorship

Documents

Transcript of The Letter to Theodore: Stephen Carlson’s Case against Clement’s Authorship