Where the world comes to study the Bible

7Q5: The Earliest NT Papyrus?

Related Media

Review of
Carsten Peter Thiede,
The Earliest Gospel Manuscript?
The Qumran Fragment 7Q5 and its Significance for New Testament Studies1
(London: Paternoster, 1992)
74 pp. + 6 pp. bibliography


In 1962 M. Baillet, J. T. Milik, and R. de Vaux published the text and plates of manuscripts from six Qumran caves (caves 2, 3, 5, 6, 7, 10).2 The seventh cave, in particular, had some interesting materials in that this was the only cave with exclusively Greek fragments. For most of these manuscripts, including 7Q5, the editors did not have a clue as to their textual identity. (7Q5 is a papyrus scrap with writing only on the recto side, having just five lines of text with parts of no more than twenty letters visible.3 The only complete word that can be detected is καιv—hardly a confidence-builder when it comes to a positive identification.)

Ten years later, in 1972, the Spanish papyrologist José O’Callaghan published a controversial article, “¿Papiros neotestamentarios en la cueva 7 de Qumrân?”4 in which he argued that the fifth manuscript from the seventh cave of Qumran was a fragment from the Gospel of Mark (6:52-53). This produced a spate of scholarly reviews5 and interactions—most of which rejected O’Callaghan’s identification. This rejection rested on three grounds: (1) principally, the papyrus itself was so fragmentary that any identification would be tenuous at best (not to mention the fact that there were several textually intrinsic problems with O’Callaghan’s proposal); (2) since the Qumran community almost certainly disbanded in 68 CE—and hence the MS must be dated before that time (in fact, most likely, no later than 50 CE)—the majority of NT scholars felt that even the original draft of Mark’s Gospel was not this early, obviously precluding the possibility that a copy of Mark could have existed before the fall of Jerusalem; and (3) the differences between the Qumran community (usually considered to be identical with the Essenes) and the nascent Christian community are so pronounced that contact between the two seemed improbable (and a literary contact, as O’Callaghan proposed, seemed to imply that not only was there communication between the two groups, but open and somewhat friendly communication).

O’Callaghan defended his views against virtually every assailant. But until 1982 he found few, if any, real followers. In that year Carsten Peter Thiede, a German scholar, began to publish in defense of the O’Callaghan hypothesis. In the last dozen years, in fact, he has surpassed his mentor in periodical proliferation. The book under review is, in many respects, the culmination of his efforts. The Earliest Gospel Manuscript?, Thiede’s first book in English on the subject, has been written to appeal to a wider audience (since his earlier writings have almost completely fallen on deaf German ears). There is today both interest in and sympathy toward the O’Callaghan hypothesis—especially now that it has a fresh advocate in Thiede.6 Indeed, at the ETS national meeting in November 1992, even Alan Johnson pleaded the case for Thiede’s volume.7

Why all the furor? What is at stake? A number of things: (1) If this identification is correct, it would be the earliest NT MS by some 50-100 years;8 (2) on paleographical grounds, since the upper limit of its date is 50 CE, this would put Mark in the 40’s at the latest; (3) one consequence of such an early date for Mark would be to virtually silence advocates of Matthean priority; and (4) finally, it would suggest, perhaps, that at least some of the New Testament documents were regarded highly enough to be copied soon after publication—a view which lends itself to an early recognition of the NT as canon.9


Body of Review

There are five chapters to this slender volume. The first, “Introduction,” is both a selective tracing of the history of the discussion and a rebuke of the scholarly community for not really listening to the arguments put forth by O’Callaghan. Chapter 2 (“Ì52—The Most Famous Papyrus”) is, in essence, an implicit yet not-so-subtle attempt to argue from similarities: since Ì52 is accepted by the entire community of NT scholars as a fragment of John’s Gospel from the first half of the second century10—even though it has itacisms and variants from the standard text—we should also accept 7Q5 as a fragment of Mark, and dated no later than 68 CE, since it has similar textual “glitches.” One telling argument that the two are not that similar is the fact that, as Thiede concedes, the identification and dating of Ì52 were “accepted without argument” (p. 12) by the scholarly community, while 7Q5’s identification has not been. Thiede spends an exorbitant amount of space demonstrating that 7Q5 should be dated no later than c. 50 CE. An interesting concession by the author, however, is the fact that C. H. Roberts, on whose expertise he relies, gives a variance of 100 years for the date of this MS: from 50 BCE to 50 CE. Obviously, the earlier the date, the less likely is the possibility that this fragment comes from the NT at all.11 Even the most conservative NT scholars do not date the Gospel of Mark as early as this upper limit set by Roberts.

Chapter 3 (“7Q5—The Earliest New Testament Fragment?”) is the most substantial of the booklet, covering nineteen pages (23-41). Thiede puts forth a meticulously argued and somewhat technical case for the identification of this fragment with Mark 6:52-53. He points out, among other things, that even though at most ten of the twenty letters can be positively identified, (1) the three-letter space before καιindicates the beginning of a new paragraph (a not uncommon feature in ancient MSS), corresponding to the content break at Mark 6:53, and (2) line 4 apparently has the unusual combination of letters, ννησ (although the first and last letters are quite difficult to make out), corresponding to γεννησαρετin Mark 6:53.12

Thiede also responds at length to the three most common (and most serious) objections to this identification: (1) 7Q5 has a tau where Mark 6:53 has a delta (τι[απεράσαντες] vs. διαπεράσαντες); (2) in order to make the lines be of somewhat equal length and correspond to Mark’s text, the ἐπὶ τὴν γῆν of v. 53 must be omitted—even though no extant MSS omit this expression; and (3) a number of O’Callaghan’s identifications of the partially readable letters are quite improbable. To those involved in the debate over 7Q5’s identification, Thiede’s argument is more summary than new insight. In essence, he argues that (1) there are frequent interchanges between tau and delta in koine Greek,13 rendering such a possibility here hardly surprising; (2) other early papyri (e.g., Ì52, Ì45) omit material at times, even though such an omission is a singular reading; and (3) if O’Callaghan’s critics had taken the time to look at the fragment instead of a photograph, their objections about his letter reconstructions would have vanished.

These counter-charges by Thiede are not as substantial as he supposes. We shall approach them chiastically. First, both the original editors of this fragment and most who have followed disagree with several of O’Callaghan’s letter reconstructions. At every point in which the enlarged photograph of the fragment at the end of Thiede’s booklet (p. 68) seems to disprove O’Callaghan’s reconstructions, Thiede discounts the empirical evidence which he himself provides and renders his own judgments untouchable by any who have access only to a photograph. In other words, he is saying, “You don’t have a right to criticize O’Callaghan’s reconstruction because you haven’t seen the fragment.” Such a stance is elitist at best; at worst, it moves the entire discussion from a scholarly dialogue to a fideistic statement: Thiede basically says “Trust me.” A constant refrain is that O’Callaghan’s reconstructions are possible. Perhaps this is so, but such are also highly unlikely. In particular, an unbiased reader looking at the photograph will almost certainly disagree with O’Callaghan’s reconstructed nu in line 214 and agree with the original editors’ judgment about epsilon, sigma in line 5 (against O’Callaghan’s sigma, alpha). Thiede is quite right that examination of a document firsthand is to be preferred to examination of a photograph.15 And this is precisely where his and O’Callaghan’s approach falters: others have looked at the MS firsthand and have disagreed with O’Callaghan.

Second, although it is certainly possible that ἐπὶ τὴν γῆν is legitimately omitted in O’Callaghan’s stichometric reconstruction,16 it strikes me as too convenient for the hypothesis: in order to make this papyrus fragment fit the text of Mark, the non-recoverable portion of the text needs to be altered. This again makes the proposal non-falsifiable. Further—and this still looms as an important consideration—such an omission is unattested in any other MS for this verse.

Third, most damaging for O’Callaghan’s identification is the tau in the place of a delta. Although, admirably, both O’Callaghan and Thiede provide examples of such interchange in koine Greek due to the similar sound of the two letters (e.g., τεfor δεv), none of the examples produced involve the preposition διαv, whether standing alone or in compound. Illustrations such as the interchange of τεfor δέ do not help the case, because both were real words with some semantic overlap. And Thiede’s example of the interchange between δρύφακτονand τρύφακτον (pp. 28-29) is not very convincing, because such a rare word would be expected to have variant spellings. The preposition διά, however, has no semantic overlap with τια (there is, in fact, no such word) and is so common that a schoolboy would have learned its correct spelling. Such a misspelling as O’Callaghan and Thiede envision this scribe as producing would be analogous to a modern author writing “tiameter” for “diameter.” In light of this, surely it is an overstatement for Thiede to assert that “one might go so far as to say that the peculiarities themselves support this view [that 7Q5 = Mark 6:52-53]” (p. 31).

One final point about chapter 3 can be mentioned. In his final footnote of the chapter (n. 31, pp. 40-41), Thiede states that “a more recent computer check [than K. Aland’s], using the most elaborate Greek texts (Ibykus [sic]) has failed to yield any text other than Mark 6:52-53 for the combination of letters identified by O’Callaghan et al. in 7Q5.” In other words, using a very powerful software search engine17 which is able to scan over 64 million words in hundreds of ancient Greek texts in a matter of minutes, Thiede could not find any text, besides Mark 6, that fit this Cinderella’s shoe.

At first glance, this sounds very impressive. But Thiede overlooked two things. First, the restriction of “letters identified by O’Callaghan” assumes O’Callaghan’s problematic letter reconstructions to be correct. But this manifold assumption is exceedingly gratuitous. It is like observing a sheet of paper that has been left out in the rain. Only a handful of letters can be made out clearly; all else is up for grabs. Now suppose I come along and say that one or two of the clear letters need to be changed. And of the unclear letters, I propose three or four nearly impossible suggestions. I do this because I have a certain text in mind that I want this sheet to be a copy of. Would it be so surprising when my Macintosh spits out that very text—after I have programmed it do so? In doing this kind of thing, Thiede has fallen prey to the very argument he just leveled against Kurt Aland in the same footnote!18

Second, when one allows for different possibilities than just O’Callaghan’s for the partially legible letters, the Ibycus program19 does, indeed, seem to permit other texts to be identified with 7Q5. In my own cursory examination of the TLG via Ibycus, I found sixteen texts which could possibly fit (though only if one stretched both his or her imagination and the textual evidence).20

Third, even if none of these is as impressive as is Mark 6:52-53 (a point I would readily concede), there is no necessity in identifying 7Q5 with any known text.21 As possible as the O’Callaghan/Thiede proposal is, it remains far more plausible to see 7Q5 as a copy of some unknown text—just like other papyri in cave 7.

Chapter 4 (three pages in length) is an attempt to show, by analogy with two other fragments, that positive identification of 7Q5 can be made in spite of the paucity of letters.

The fifth chapter (“The Seventh Cave at Qumran—Its Text and Their Users”) (pp. 45-63) answers the historical question: Why would Christian documents be concealed in a Qumran cave? Thiede summarizes O’Callaghan’s case that some of the other fragments in this cave are portions from the NT (e.g., 7Q6 = Mark 4:28; 7Q15 = Mark 6:48; 7Q8 = Jas 1:23-24; 7Q9 = Rom 5:11-12; 7Q10 = 2 Pet 1:15; 7Q4 = 1 Tim 3:16-4:3).22 Such equations were pursued by O’Callaghan because he had already felt that his identification of 7Q5 was certain. As would be expected, he has received quite a bit of criticism for these speculations. Some of the arguments against his proposals are that (1) the fragments involved have as few as three or four clearly identified letters; (2) one of the documents, 7Q6, has two fragments, yet O’Callaghan assigned the first to Mark 4, the second to Acts 27; (3) on higher critical grounds, that 2 Peter and 1 Timothy especially could have had copies by 68 CE seemed impossible;23 (4) four fragments identified as copies of Mark by four different scribes seemed to go beyond even the realm of “Phantasie”;24 (5) textual emendations and/or less than probable reconstructions of letters were forced on the fragments to make them fit the theory; and (6) 7Q4 (= 1 Tim 3:16-4:3) is, paleographically, so much like 7Q5, that it should likewise be dated no later than 50 CE—and this is an impossible date for any pastoral epistle. In my judgment, Thiede does not adequately address these concerns (many of which are completely ignored).

Regarding the historical situation, Thiede devotes ten pages (54-63) to his defense of a Christian cave among the Qumran caves. He builds an ingenious case for geographical contact between Christians and the Essenes in Jerusalem, with many of his points containing an element of truth. From this he extrapolates that when the Christians left Jerusalem for Pella (c. 66 CE), they would have “entrusted them [their sacred documents], or some of them, to their Essene neighbours for safekeeping, and they, in turn, [would have] hid them in a separate cave at Qumran” (p. 58). Although this reconstruction is in the realm of possibility, it is barely so.

Even if we were to grant geographical contact between Christians and Essenes in Jerusalem, it is too much to assume that there was a friendly familiarity between the two communities. Two considerations seem to argue against this. First, the Essenes were the most extreme separatists of any Jewish sect in the first century—so much so that they established a celibate community away from Jerusalem. If they hardly communicated with other Jews, how much less would they do so with Christians? Second, the Essenes were extreme legalists.25 The Christians were at the other end of the spectrum. And it is significant that five of the fragments found in cave 7 are allegedly from Mark and Romans—two books which are about as anti-legalistic as can be found in the NT canon. In light of these two considerations, is it really plausible that the early Christians “entrusted [these documents] to their Essene neighbours for safekeeping”?

The book concludes with several illustrations (including 7Q5, Ì52, et al), inviting the reader to see exactly what it is that the experts have been debating.



To sum up: Not only are O’Callaghan and Thiede arguing that 7Q5 is a fragment from Mark’s Gospel, but they are also appealing to Kurt Aland to list this document officially as a NT papyrus: “Future editions of the Greek New Testament will have to include 7Q5. It should, at long last, receive a ‘p’ number, it must be recognized in the apparatus, with its variants” (p. 41). Here is no detached plea; rather, it is an indictment. And this not-so-subtle indictment takes on parabolic overtones in the concluding statement of the book, where Thiede comments about the alleged early Christians who orchestrated the burying of these documents in Qumran’s Cave 7 (p. 63):

Using papyrus instead of the more expensive parchment, these first Christians were eager to share the first fruits of their own literary harvest with those who were hungry for the good news. When it was a question of promoting the gospel about Jesus they showed a spirit which was at the same time innovative and open-minded. Of them, it could not be said what Mark writes, preserved in 7Q5, about the first disciples after the feeding of the five thousand: ‘Their minds were closed.’

Putting all this in perspective, we conclude this review by addressing two concerns: evidence and attitudes. First, what is the hard evidence on which O’Callaghan’s identification is based? A scrap of papyrus smaller than a man’s thumb with only one unambiguous word—και. Only six other letters are undisputed: τω (line 2), τ (line 3, immediately after the και), νη (line 4), η (line 5). To build a case on such slender evidence would seem almost impossible even if all other conditions were favorable to it. But to identify this as Mark 6:52-53 requires (1) two significant textual emendations (tau for delta in a manner which is unparalleled; and the dropping of ἐπὶ τὴν γῆν even though no other MSS omit this phrase); and (2) unlikely reconstructions of several other letters. Add to this that the MS is from a Qumran cave and that it is to be dated no later than 50 CE and the case against the Marcan proposal seems overwhelming. If it were not for the fact that José O’Callaghan is a reputable papyrologist and that C. P. Thiede is a German scholar, one has to wonder whether this hypothesis would ever have gotten more than an amused glance from the scholarly community.

Second, regarding attitude, I find it disturbing that many conservatives have been so uncritically eager to accept the O’Callaghan hypothesis. 7Q5 does not, as one conservative put it, mean “that seven tons of German scholarship may now be consigned to the flames.”26 On the other hand, I find it equally disturbing that many liberal scholars have uncritically rejected O’Callaghan’s proposal without even examining the evidence. Higher criticism must of course have a say in this discussion; but it must not preclude discussion. Both attitudes, in their most extreme forms, betray an arrogance, an unwillingness to learn, a fear of truth while clinging to tradition, a fortress mentality—none of which is in the spirit of genuine biblical scholarship. When the next sensational archaeological find is made, should not conservatives and liberals alike ask the question: Will we fairly examine the evidence, or will we hold the party line at all costs? 27



1There is some confusion over the title. The title listed above is what appears on the book’s cover. However, on the title page “Papyrus” has replaced “Fragment.”

2Les ‘Petites Grottes’ de Qumrân, DJD III.

3Its dimensions are, in Thiede’s words, “at the most 3.9 cm high and 2.7 cm wide. At most, visible text covers an area measuring 3.3 cm high and 2.3 cm wide” (p. 25). In other words, 7Q5 is smaller than two standard U.S. postage stamps.

4Bib 53 (1972)91-100. Translated into English in the JBL 91 (1972) supplement no. 2.

5See Thiede’s bibliography for a listing of the reviews, which are in any case too numerous to mention in a footnote. Among the specific reviews in scholarly journals alone (i.e., neither books nor essays where 7Q5 is only a part of the discussion), New Testament Abstracts lists more than thirty—not to mention one dozen responses by O’Callaghan!

6Fifteen years ago, David Estrada and William White, Jr., argued his case in The First New Testament. In 1980, Wilbur Pickering added his support in his The Identity of the New Testament Text , 2nd ed. (Nashville: Nelson), 155-158. This supports his majority text theory of textual criticism in the following way: “That someone should have such a collection of New Testament writings at such an early date may suggest their early recognition as Scripture and even imply an early notion of a New Testament canon” (158).

7One should note at the outset that this work is marred by scores of not insignificant typographical errors, including grammar and spelling mistakes, several misquoted statements, and worst of all, a discrepancy in the very title of the book. Such a casual approach to the form of presentation can give the reader a natural temptation to see an equally imprecise handling of the data on Thiede’s part. A second, corrected edition ought to be published as soon as possible, if for no other reason than to remove an unnecessary stumbling block for the viewpoint espoused.

8Ì52 is to be dated c. 100-150 CE, while 7Q5 is dated c. 50 BCE-50 CE.

9Another possible implication would have to do with the ending of Mark. Since the fragment 7Q5 was written only on one side, it was doubtless a scroll rather than a codex. If so, then the original of Mark would most likely have been a scroll. And if this is the case, it is extremely unlikely that the ending of Mark would have somehow become lost—since the ending of a scroll would, under normal circumstances, be the most protected part of the document. In this case the most plausible scenario for the ending of Mark is that the author chose to end his gospel at 16:8. This argument can certainly be sustained without 7Q5, though it would not hurt to have this MS lend its voice.

10Thiede makes the repeated assertions that this papyrus should be dated “to the first quarter of the second century (at the very latest)” (p. 2; cf. also p. 21), in spite of the fact that most textual critics today would be more comfortable dating it more generally, c. 100-150 CE (cf. B. M. Metzger, Manuscripts of the Greek Bible: An Introduction to Greek Palaeography [Oxford: Clarendon, 1981] 62). One of the reasons for this is that a scribe’s handwriting is not going to change very much over the duration of his career. Thus, on palaeographical grounds, it is difficult to pinpoint the date of a MS within a period smaller than 50 years (ibid., p. 50).

11Thiede makes the remarkable statement that “leaving theological arguments aside, the earliest possible date for this gospel, historically speaking, is AD 30, the year of the last event recorded in it, the resurrection of Jesus” (p. 25). Thiede’s assessment that higher critical reconstructions—especially as regards the synoptic problem—are merely “theological arguments” strikes me as a bit naïve and ought to signal the reader to Thiede’s antecedent eagerness to accept O’Callaghan’s identification of 7Q5. No reputable NT scholar—regardless of his theological underpinnings or views of gospel priorities—dates Mark this early.

12On the basis primarily of these two points Thiede asserts: “Even without considering other aspects of the fragment in detail, it should be clear to any unbiased observer that on the basis of these findings, the indentification [sic] of the fragment as Mk 6:52-53 is more than merely probably [sic]” (p. 27).

13See especially O’Callaghan, “El cambio δ>τ en los papiros biblicos,” Bib 54 (1973) 415-16, as a demonstration of this point. O’Callaghan finds twenty places in biblical papyri (18 for LXX, two for NT) where this interchange takes place.

14See especially G. D. Fee, “Some Dissenting Notes on 7Q5 = Mark 6:52-53,” JBL 92 (1973) 109-12.

15Actually, the ideal is to examine both the original document and a photograph side-by-side. The advantages of a photograph involve enlargement and contrast especially. I recently discovered this in a fresh examination of Ì26: the photograph revealed at least eight more letters than could be detected by looking at the papyrus alone.

16With the omission of the expression, the letters per line are as follows: 20/23/20/21/21. If the phrase is left intact, the lines are 20/23/29/21/21.

17The search engine was in fact Ibycus, but the database being searched was the Thesaurus Linguae Graecae, developed by the Packard Humanities Institute. There are now several search engines available to search the TLG, both for Mac and Windows platforms, but there is only one TLG.

18Thiede points out that “Aland used the computer at his institute in Münster in order to analyse two different combinations of letters which he thought were possible “minimal” readings of the fragment 7Q5 . . . But . . . Aland’s efforts had to fail for a methodological reason . . . : no existing edition of the Greek text of Mark has the variant tau for delta in the ‘diaperasantes’. Thus, Aland’s computer programme of the Greek New Testament, based here on the delta, had to miss Mark 6:52-53 as a possible passage, and it promptly did.”

19A lexical search engine canvassing over sixty million words in Greek literature (based on the Thesaurus Linguae Graecae), from Homer to 1453 CE.

20The search involved the following pattern: των, καιτ, ννη, corresponding to lines two, three, and four of 7Q5 (and even allowing O’Callaghan his nu in line 2). The passages found include Ezek 23:36; Josephus, Vita 42-3; Vita 236; Bellum 5.528; 7.380-1; Philo Cher. 44; 119; Plant. 135; Plant. 136; Mut. 173; Thucydides, Hist. 1.10.2; 1.60.1; 3.109.2; 4.67.4; 5.82.5; 8.55.1. I would not be so rash as to suggest that 7Q5 is a copy of any of these passages, but just that the identification with Mark 6 is not unparalleled. Almost all of these passages—like Mark 6—involve what I consider to be insuperable problems: date (in the case of the Josephus texts), length of line, and manipulation of partially legible letters. With a little imagination, however, I was able to emend several of the texts (even finding plausible homoioteleuta, metatheses, etc.) and make the data fit. In fact, in one text this was not even necessary. In Philo, Plant. 135 the three lines of text can be reconstructed, without any textual emendation, in a 16/14/16 stichometry:

θωματων απαντων αρ (16)

ιστον  και τελειο  (14)

τατον γεννημα ο εισ (16)

τον πατερα . . .

There is a certain advantage of this text over Mark 6: whereas O’Callaghan’s reconstruction involves twenty or twenty-one letters per line as the norm—including line 3 which has a three-letter gap and ought therefore to have fewer letters, the Philonic text has two letters fewer in line 3, taking into account the gap in 7Q5 at this point.

Of course, there is still the problem of forcing the partially legible letters into the theory—but this suffers no disadvantage over against the Marcan proposal.

21Other potential identifications have been suggested on occasion. Cf., e.g., Gordon D. Fee, “Some Dissenting Notes on 7Q5 = Mark 6:52-53,” JBL 92 (1973) 109-112; Conan DiPonio Parson, 7Q5: An Ancient “Honey Do” List? (Snowflake, Saskatchewan: Technasma Press, 1975); Kurt Aland, “Über die Möglichkeit der Identifikation kleiner Fragmente neutestamentlicher Handschriften mit Hilfe des Computers,” in Studies in New Testament Language and Text, ed. J. K. Elliott (Leiden: E. J. Brill, 1976) 14-38; V. Spottorno, “Una nueva posible identificación de 7Q5,” Sefarad 52 (1992) 541-43.

22O’Callaghan’s most certain (in his mind, that is) identification was that 7Q4 = 1 Tim 3:16-4:3.

23This is true even if one holds to apostolic authorship. Some date Paul’s death at 67 CE, and Peter’s at 68.

24So K. Aland, “Neue neutestamentliche Papyri III,” NTS 20 (1974) 363.

25So much so that they even refused to urinate on the Sabbath, regarding even that as “work”! Ironically, in Thiede’s own reconstruction the Essenes’ latrine wall was in close proximity to where the Christians met for prayer. One can only wonder if friendly associations should truly be implied from such evidence.

26Anonymous, “Eyewitness Mark?Time, 1 May 1972, 54.

27 This essay was originally delivered at the Evangelical Theological Society Southwestern Regional Meeting, held at John Brown University in March, 1994.

Related Topics: Canon, Textual Criticism