A Case against the Longer Ending of Mark

This is the second article on Mark’s ending. It responds to an argument for the authenticity of Mark 16:9–20.

I am grateful for the opportunity to respond to James Snapp’s article. He argued that we should regard Mark 16:9–20 as the ending of Mark’s Gospel. I will be arguing instead that we should regard Mark 16:8 as the ending of Mark’s Gospel.

For clarity I simply note that I will use the term Longer Ending for Mark 16:9–20, and the term Shorter Ending for the sentences which were never given verse numbers, but which follow on from “for they were afraid” (v. 8) in some manuscripts with the words: “But they reported briefly to Peter and those with him all that they had been told. And after this Jesus himself sent out by means of them, from east to west, the sacred and imperishable proclamation of eternal salvation.”¹ I will argue that neither the Shorter Ending nor the Longer Ending were the initial ending of Mark’s Gospel, but that we should regard Mark 16:8 as the original ending of Mark’s Gospel.

What Counts as Evidence

Before we get into the evidence, we should note that one of the issues in this debate concerns what actually counts as evidence, especially what counts as evidence against the Longer Ending. It is reasonably clear what counts as evidence for the Longer Ending, or at least evidence that the Longer Ending was known and used as the ending of Mark’s Gospel—manuscripts (in Greek and other languages) in which Mark’s Gospel ends with 16:20 are the clearest evidence, as well as Church Fathers who clearly quote from portions of the Longer Ending.

But it is not as clear what will count as evidence against the Longer Ending. Church Fathers who don’t quote from any passage in the Longer Ending (like Origen or Clement of Alexandria, just for the sake of argument), are not providing us with evidence that their copy (or copies) of Mark ended at 16:8. It is not even an argument from silence, it is an argument from absence. But this doesn’t really seem like solid evidence at all. Of course, it would be relevant if a Church Father preached through Mark and his sermons finished with 16:8; but of course, sermon series and commentaries on Mark are pretty much completely absent from the patristic period.

So, what counts as evidence against the Longer Ending then? Only manuscripts (in Greek or other languages) which end Mark’s Gospel at Mark 16:8, manuscripts which preserve evidence for earlier manuscripts which lacked the Longer Ending, and Church Fathers explicitly discussing the ending of Mark. At least that narrows our discussion somewhat.

Points of Agreement

James Snapp’s argument for the Longer Ending reflects his long-standing commitment to the investigation of this problem, as well as his learned impatience with inaccurate generalizations about the status of the manuscript and patristic evidence in some commentaries and English translations.

But it also reflects a particular approach to which bits of evidence are actually decisive in this discussion. At various points I will obviously disagree with his judgment on that point—otherwise the conveners would not have asked me to write on this topic! But before I disagree, I want to begin with a whole series of his points with which I am in substantial agreement.

Irenaeus provides indisputable evidence for the Longer Ending in the 2nd century. Source

The vast majority of extant Greek manuscripts of Mark (copied between the fifth and the fifteenth centuries) contain the Longer Ending of Mark. I agree that over 99 percent of all manuscripts of Mark in this period contain the Longer Ending though I also think that Snapp doesn’t always acknowledge the slim strand of evidence showing that this ending was also disputed.

So, on these six points I am in substantial agreement.

Of course, in part, this agreement highlights the problem of the Longer Ending: it was absent from the earliest manuscripts, it doesn’t fit when it does appear in the later manuscripts, and it has unusual stylistic features; but it was widely adopted as the ending of Mark, is quoted regularly in ecclesiastical writers, and became the almost universal ending of Mark in later manuscripts. This is the textual problem—a combination of external and internal evidence—that scholars are trying to investigate, explore, and ultimately explain and resolve.

Differences of Judgment

Within a discussion taking in many different manuscripts, versions, church fathers, and such there is ample room for differences of judgment. We may distinguish between disagreements which are fundamental and those which are more marginal.

For example, I am not as confident about finding evidence for the Longer Ending of Mark, as part of Mark, within the second-century witnesses such as Justin, Tatian’s Diatessaron, and the Epistula Apostolorum, and perhaps a little more confident that the Gospel of Peter should be considered as a witness to the ending of Mark at 16:8.² These are not, however, crucial differences, just areas where scholars have often come to different judgments on questions where the evidence is inconclusive. Take, as one example of these, the question as to whether Justin Martyr, an important mid-second-century Christian leader in Rome, knows the Longer Ending of Mark.

It is difficult to agree on a method for determining when and where second-century Christian writers show knowledge of particular New Testament texts. We can perhaps agree on a statement of the facts of the matter: Justin Martyr, on one occasion in his First Apology 45 (as cited by Snapp) used three words—relatively quite common terminology, about the apostolic mission—that are also found in Mark 16:20, not in the order they are found in Mark, in a vaguely similar context, but without other strong connections with the context in Mark.

So I disagree with Snapp’s confidence in seeing this as proof that Justin knew the Longer Ending of Mark. At best I would rate this as “possible.” But it is also possible that the three-word quasi-agreement is coincidental (I suppose a common source could not be excluded). Thus, I remain more at home with Westcott and Hort on this: “the evidence is slight.”

It would be similarly possible to come to more cautious conclusions about Tatian’s Diatessaron—where the task of reconstructing Tatian’s work is obviously complex and the problem can be posed simply by noting that Snapp’s evidence for this second-century harmony actually comes from a sixth-century Latin manuscript and a fourth-century Syriac commentary.³ But differences in these matters (or in relation to the Epistula Apostolorum) are not particularly decisive; they are disagreements in judgment.

More Substantial Disagreements

There are four areas, however, requiring more significant discussion.

1. Internal Evidence

In relation to the style and content of the Longer Ending, I agree with Snapp that the following features of the Longer Ending provide “compelling evidence” that this is not Mark’s originally intended ending. I shall note Snapp’s evidence and then explain:

The reintroduction of Mary Magdalene. Mary Magdalene has already appeared three times in the latter sections of Mark: at the cross (15:40), at the tomb seeing where Jesus’ body was placed (15:47), and coming to the now empty tomb on Sunday morning (16:1). Because of this, it is incongruous to introduce her in 16:9 as “the one from whom seven demons had been expelled” (a phrase that comes from Luke’s introduction of her in Luke 8:2)
The restating of the day and time. Mark 16:1 states in emphatic manner that the women came to the tomb “very early on the first day of the week, when the sun had risen,” while 16:9 states the time of Jesus’ resurrection as “early on the first day of the week” in a way that is both unnecessarily repetitive and also in different wording to what was used in 16:1.
The lack of any meeting in Galilee predicted in Mark 14:28 and 16:7. This is a major problem with the Longer Ending of Mark—it doesn’t deliver what both Jesus and the angel promised would take place!

But I think the internal issues go further than this in the case of both style and content. In relation to style, given that twelve verses are a small sample, two features suggest a different author: the frequent use of the pronoun “that” or “those” (ἐκεῖνος) referring to people (v. 10: ἐκείνη; v. 11: κἀκεῖνοι; v. 13: κἀκεῖνοι, ἐκείνοις; v. 20: ἐκεῖνοι), and the general shift in connectives away from a simple “and” kai (καί) to the post-positive “but” de (δέ; vv. 9, 12, 14, 17, 20 [in a μέν … δέ construction])—Mark generally uses de to signal a change of subject, but in 16:9–20 it becomes the default connective.

In relation to content there is a significant issue that the Longer Ending draws upon parallel material in the other Gospels.

In relation to content there is a significant issue that the Longer Ending draws upon parallel material in the other Gospels.⁴ The individual appearance to Mary Magdalene (Mark 16:9–11) parallels John 20:14–18; the appearance to two people walking in the country (Mark 16:12–13) parallels the two disciples on the road to Emmaus in Luke 24:13–35; the appearance to the eleven while reclining (Mark 16:14) parallels Luke 24:36–43; the commissioning (Mark 16:15) parallels Matthew 28:19–20; and the mention of the ascension (Mark 16:19) parallels Luke 24:50–51. This synthesizing feature of the content of the Longer Ending has long been recognized as reflecting a different relationship to the other Gospels than is reflected within Mark’s Gospel.⁵

2. Church Fathers

I’m not persuaded that Snapp does justice to all the patristic evidence in favor of ending the Gospel at Mark 16:8, including the important evidence of Eusebius (and Jerome). It seems clear to me that in his letter to Marinus, in discussing the problem of harmonizing Matthew 28:1 and Mark 16:9, Eusebius not only quotes from the perspective of the person who would regard the Longer Ending as spurious—where he reports that “the accurate copies” end at 16:8—but that he himself also affirms that this was the ending of Mark “in nearly all the copies” (to Marinus 1).⁶ Eusebius does also attempt a harmonization, for the sake of argument, assuming that the Longer Ending was part of the text of Mark, but Eusebius is offering a complex double sort of approach here.⁷ What is also clear is that, in the construction of his influential Canon Tables, Eusebius did not include the Longer Ending of Mark.⁸

A page of Eusebius’s canon tables in Walters Manuscript W.538, fol. 8v (12th c.). Public domain

To this we could add the testimonies of Hesychius of Jerusalem (5th c.), who notes that Mark’s Gospel ends after the appearance of the angel to the women, and Severus of Antioch (465–538), who echoes Eusebius’s comment that “in the more accurate copies the Gospel according to Mark ends at the [passage] ‘for they were afraid.’”⁹ The comments of Eusebius (who travelled widely but was especially associated with Caesarea) and Jerome (mostly associated with Rome), show that awareness of the manuscript support for ending the Gospel at Mark 16:8 was present in intellectual and major ecclesiastical centers of the fourth and fifth centuries.

3. External Evidence

The straightforward evidence of the two great fourth-century codices, Sinaiticus and Vaticanus, is unduly minimized by Snapp. These are our earliest complete Greek manuscripts of Mark, and they both end the Gospel quite clearly at Mark 16:8. In Codex Sinaiticus the text ends at Mark 16:8 with a decoration and an end-title (“Gospel according to Mark”). In Codex Vaticanus the text ends at 16:8 with a decoration and an end-title (“according to Mark”), followed by an empty column.

This essentially quite straightforward evidence is subjected to some obfuscation by Snapp on the basis of his interpretation of the decorations of Codex Sinaiticus and the unusual empty column in Vaticanus. It is well known that the end of Mark in Sinaiticus (from Mark 14:54) and the beginning of Luke (up to Luke 1:56) are written on a replacement sheet by scribe D. (This is one of three such replacement sheets within the New Testament.) Investigations of the possible problem corrected by Scribe D (who appears to be the senior scribe, regularly correcting the work of Scribe A, and copying very carefully) have pointed to the likelihood of problems in the text of Luke. As Dirk Jongkind states, supporting prior scholarship: “the so-called longer ending of Mark could never have fit on this sheet.”¹⁰

In addition, it is not only the earliest Greek manuscripts which end at Mark 16:8. There is good evidence that the earliest form of the Gospel of Mark, as translated into Latin, Syriac, Sahidic Coptic, Christian Palestinian Aramaic, Armenian, and Georgian, all consisted of the text of Mark ending at 16:8. This is further very significant confirmation of the testimony of Eusebius as to the state of the text of Mark in the manuscripts of the fourth century. Here is a brief listing of the relevant evidence.¹¹

The ending of Mark in Codex Bobiensis, the earliest Latin manuscript of Mark. Source

The oldest Old Latin manuscript (VL 1 = Codex Bobiensis, from the fourth or fifth century) concludes at Mark 16:8 with a version of the Shorter Ending and lacks 16:9–20
The oldest Syriac manuscript (the Sinaitic Syriac, from the fourth century) ends at 16:8
The oldest Sahidic manuscript (sa 1 = P. Palau-Ribes Inv. Nr. 182, from the fifth century) ends at 16:8
The earliest evidence we have for the Christian Palestinian Aramaic version of Mark (Codex Sinaiticus Rescriptus in St Petersburg, Syr. No. 16) ends at 16:8
The oldest Armenian manuscripts (going back to the ninth century) end at 16:8
The oldest Georgian manuscripts (translated from the Armenian) end at 16:8

In each of these language groups, later witnesses include the Longer Ending, but that does not detract from the force of this observation. The general direction of travel in the manuscript evidence as we have it for Greek, Sahidic, Latin, Syriac, Christian Palestinian Aramaic, Armenian, and Georgian, moves from an original, shorter Mark towards incorporating a version of Mark with the Longer Ending.

In other words, the Greek manuscript evidence, and the evidence within these six linguistic areas, works in the opposite direction to that proposed by James Snapp (and others). In short, it is not evidence that an original long form of Mark was subsequently edited down, but is in fact evidence for the opposite: the earliest form of Mark known in these areas ended at Mark 16:8, and this was subsequently supplemented with one or more of the available additional endings.

4. Method

Finally, in terms of method, it is a general principle within New Testament textual criticism to work on the principle that the reading which explains the other readings is to be preferred. Snapp attempts to explain the ending at Mark 16:8 as an editorial emendation by “overly meticulous scribes,” that is, as a deletion of material within their exemplars.

But evidence for this sort of speculative conjecture is lacking. A stronger argument is that an ending at Mark 16:8 explains the origin of the other readings. It is an unusual and abrupt ending, which gave rise to a natural desire for a clearer ending, and this is evident in both the Shorter and the Longer endings to Mark. This is the tendency of the textual tradition as already noted.

The Significance of the Longer Ending

This argument is obviously my own, but broadly speaking this is the sort of argument which stands behind the views of many textual critics over the past two centuries (e.g., Lachmann, Tischendorf, Tregelles, Westcott & Hort, Nestle, Aland, Metzger, Parker), it explains the decision to print a primary text ending at Mark 16:8 in the contemporary critical editions of the Greek New Testament (N1–NA28, UBS1–5, THGNT, SBL, ECM), and it is the viewpoint reflected in practically all the recent scholarly commentaries on Mark (e.g., Yarbro Collins, Hooker, France, Gundry, Pesch, Gnilka). If this judgment is correct, one final but important question remains: what should we do with the Longer Ending?

One final but important question remains: what should we do with the Longer Ending?

To be honest, I hesitate at precisely this point (and I appreciate that I am channeling Eusebius here). Some days I want to argue, with the full courage of my textual convictions, for simply printing the text of Mark up to Mark 16:8 and then closing (with Sinaiticus and Vaticanus). Positively, this would be a clear expression of the conclusions drawn from internal and external evidence. It would expose readers to the surprising nature of the original Markan ending. And the Longer Ending could take its rightful place, not among the words of the inspired authors of Scripture, but among the writings of the Apostolic Fathers, Christian writings from the generations after the apostolic period.

But on other days, I think of the 99 percent of Greek manuscripts which contain the Longer Ending, I consider the breadth and depth of Christian reception of the Longer Ending in the Church Fathers from the time of Irenaeus, and its ubiquitous presence in primary historical vernacular translations of the New Testament. On these days I think of it as inhabiting a quasi-canonical space, somewhat similar to the place of the Apocrypha in the Anglican tradition: not inspired Scripture, not for deciding on doctrine, but interesting, useful, and even edifying to read and ponder.

Notes

1
The Editio Critica Maior (ECM) identifies the Shorter Ending as Mark 16:8/38–104. In this system of reference every word in the text is assigned an even number, and each space between words is assigned an odd number. It is a little cumbersome, but it is made for precision! The ECM for Mark has recently been published and I shall refer to it in some of the following notes. Novum Testamentum Graecum Editio Critica Maior. I.2 Das Markusevangelium (Stuttgart: Deutsche Bibelgesellschaft, 2021; in three parts).
2
On these, see now N. Kiel, “Die frühen Kirchenväter als Zeugen des kurzen and langen Markusschlusses” in Editio Critica Maior. Das Markusevangelium. Teil 3. Studien, 105–132. Kelhoffer is more positive about both Justin and Tatian in Miracle and Mission, 170–175.
3
Within this Syriac commentary, the only evidence for the Longer Ending of Mark comes in the form of Jesus’ commission: “Go forth into the whole world, and baptize in the name of the Father, and of the Son, and of the Spirit.” This admittedly, does seem like a conflation of Mark 16:15 and Matt 28:19. But that is the only direct evidence. Quoted from C. McCarthy, Saint Ephrem’s Commentary on Tatian’s Diatessaron: An English Translation of Chester Beatty Syriac MS 709 with Introduction and Notes (JSSS 2; Oxford: OUP, 1993), 289.
4
See Kelhoffer, Miracle and Mission, 123–150, for an overview; see the charts on pp. 138–139. He concludes: “The numerous allusions to Matthew, Luke and John—especially to the ends of these writings—demonstrate that the author of the LE [Longer Ending] wrote with knowledge of copies of these writings” (p. 150). Cf. also C.B. Amphoux, “La finale longue de Marc: un epilogue des quatre évangiles” in C. Focant (ed.), The Synoptic Gospels, Source Criticism and the New Literary Criticism (BETL 110; Leuven: Peeters & LUP, 1993), 548–55.
5
For a survey of the history of scholarship see S. L. Cox, A History and Critique of Scholarship Concerning the Markan Endings (Lewiston: Mellen, 1993); Kelhoffer, Miracle and Mission, 5–46.
6
See R. Pearse (ed), Eusebius of Caesarea, Gospel Problems and Solutions Quastiones ad Stephanum et Marinum (CPG 3470) (Ipswich: Chieftain, 2010), 96, 97. See also J.A. Kelhoffer, “The Witness of Eusebius’ ad Marinum and Other Christian Writings to Text-Critical Debates concerning the Original Conclusion to Mark’s Gospel,” ZNW 92 (2001), 78–112. Jerome’s later comment to the effect that the Long Ending “appears scarcely in [copies of] the Gospel, while almost all books in Greek do not have this pericope at the end, especially since it seems to narrate things different and contrary to certain evangelists” (Ep. CXX.3, ad Hedybiam; AD 406–407) does bear some relationship with Eusebius’ discussion (translation from Kelhoffer, “Witness,” 99).
7
C. J. J. Coombs, A Dual Reception: Eusebius and the Gospel of Mark (Minneapolis: Fortress, 2016).
8
See M. R. Crawford, The Eusebian Canon Tables: Ordering Textual Knowledge in Late Antiquity (OECS; Oxford: OUP, 2019), 182.
9
Both are cited in the ECM apparatus. Texts and discussion in Kelhoffer, “Witness,” 101–104 (citation from p. 103).
10
D. Jongkind, Scribal Habits in Codex Sinaiticus (T&S III.5; Piscataway: Gorgias Press, 2007), 39–57 for general discussion, citation from p. 45; cf. H. J. M. Milne and T. C. Skeat, Scribes and Correctors of the Codex Sinaiticus (London: British Museum, 1938), 9–11. For more on Scribe D, see P. M. Head, “Some Observations on Various Features of Scribe D in the New Testament of Codex Sinaiticus” in Codex Sinaiticus: New Perspectives on the Ancient Biblical Manuscript (eds. S. McKendrick, D. C. Parker, A. Myshrall, and C. O’Hogan; London: British Library, 2015), 127–137.
11
Latin: Turin, Bibl. Nat. Univ. 1163. C. Cipolla (ed), Il codice evangelico k della Biblioteca Universitaria Nazionale di Torino (Turin: Molfese, 1913). See especially C. Clivaz, “Mk 16 im Codex Bobbiensis. Neue Materialien zur conclusion brevior des Markusevangeliusms,” ZNT 47 (2021), 59–85 (with reference to earlier scholarship).
Syriac: George A. Kiraz (ed), Comparative Edition of the Syriac Gospels Aligning the Sinaiticus, Curetonianus, Peshitta and Harklean Version (NTTS 41.1–4; Brill: Leiden, 1996), vol. 2, 251–52. Taylor notes, “The general consensus is that this manuscript represents a generally earlier form of the Old Syriac Gospel text than the Curetonian manuscript,” D. G. K. Taylor, “New Developments in the Textual Study of the Old Syriac Gospels” in At One Remove: The Text of the New Testament in Early Translations and Quotations (eds H. A. G. Houghton and P. Montoro; Piscataway: Gorgias, 2020), 1–42, at pp. 11–12.
Sahidic: Edition: H. Quecke, Das Markusevangelium saïdisch: Text der Handschrift PPalau Rib. Inv.-Nr. 182 mit den Varianten der Handschrift M 569 (PCS&T 4; Barcelona: Papyrologica Castroctaviana, 1972). This is taken to be the primary witness to the earliest form of the Sahidic translation of Mark by A. Boud’Hors and S. Torallas Tovar, “Towards a Textual History of the Gospel of Mark in Sahidic Coptic. Prolegomena to a New Critical Edition” in Editio Critica Maior, Das Markusevangelium. Teil 3. Studien, 203–220. ECM notes sa 393var—with text from Mark 1:1–2 and 16:8—as another witness to the Sahidic version ending at Mark 16:8, see S. G. Richter and K. D. Schröder, “Zur koptischen Markus-Überlieferung” in Editio Critica Maior, Das Markusevangelium. Teil 3. Studien, 185–202 (at p. 200). For this amulet, containing readings from Matt 1:1 and 28:20; Mark 1:1–2 and 16.8; Luke 1:1 and 24:53; John 1:1 and 21:25 (Freiburg/Schweiz Bible and Orient Museum ÄT 2006.8), see G. Emmenegger, “Ein koptisches Amulett als Beleg für den kurzen Markusschluss,” ZNW 103 (2012), 142–145. For an earlier stage of scholarship see P. E. Kahle, “The End of Mark’s Gospel: The Witness of the Coptic Versions,” JTS 2 (1951), 49–57.
Christian Palestinian Aramaic: C. Müller Kessler & M. Sokoloff, The Christian Palestinian Aramaic New Testament Version from the Early Period: Gospels (CCPA IIA; Groningen: Styx, 1998), 97 (fol. 103/40v); this is designated as CSRP^cin the ECM following C. Müller-Kessler, ‘Codex Sinaiticus Rescriptus (CSRG/O/P/S). A Collection of Christian Palestinian Aramaic Manuscripts’ Le Muséon 127 (2014), 263–309.
Armenian: Out of 220 manuscripts studied by Colwell, 99 manuscripts (which are generally earlier manuscripts) end Mark at 16:8, 33 manuscripts add the Long Ending after an end-title for the Gospel of Mark (or other closing signal), while 88 manuscripts (generally later ones) include 16:9–20. He concludes that 16:9–20 were not present in the original Armenian translation. E. C. Colwell, “Mark 16:9–20 in the Armenian Version,” JBL 56 (1937), 369–386. One Armenian manuscript from AD 989 (Echmiadzin Patr. Libr. 229) attributes 16:9–20 to Ariston the elder: Colwell, “Mark 16 9–20 in the Armenian Version,” 373, 383. Cf. F. C. Conybeare, “Aristion, the Author of the Last Twelve Verses of Mark,” The Expositor Fourth Series VIII (1893), 241–254.
Georgian: Blake argued that the Adysh Gospels (dated to AD 897 and representing a translation from perhaps the fifth century) represented the earliest translation into Georgian, with later versions being adapted to Greek texts: R. P. Blake, The Old Georgian Version of the Gospel of Markfrom the Adysh Gospels with the Variants of the Opiza and Tbet’ Gospels. Edited with a Latin Translation (Patrologia Orientalis XX.3; Paris: Firmin-Didot, 1928); cf. also D. M. Lang, “Recent Work on the Georgian New Testament,” BSOAS 19 (1957), 82–93; for general orientation: J. W. Childers, “The Georgian Version of the New Testament” in The Text of the New Testament in Contemporary Research: Essays on the Status Quaestionis, 2nd ed. (eds. M. W. Holmes and B. D. Ehrman; NTTSD 42; Leiden: Brill, 2012), 293–327.

Peter M. Head

Peter Head (PhD, Cambridge University) is New Testament Tutor at Wycliffe Hall and Member of Faculty of Theology and Religion at the University of Oxford. He previously served as the Sir Kirby Laing Senior Lecturer in New Testament, a joint appointment between Tyndale House and the Faculty of Divinity in the University of Cambridge. He is the author of Christology and the Synoptic Problem and How the New Testament Came Together and numerous articles.

Topics

Levels

Most Read

A Case against the Longer Ending of Mark

An argument that Mark 16:9–20 is not original and so not inspired Scripture

What Counts as Evidence

Points of Agreement

Get new articles and updates in your inbox.

Differences of Judgment