Chapter 1. Mark 16:8 as the Conclusion to the Second Gospel

Chapter 1

Mark 16:8 as the Conclusion to the Second Gospel

Daniel B. Wallace

Comedian George Burns once pontificated on the key to homiletical success: "The secret of a good sermon is to have a good beginning and a good ending, then having the two as close together as possible." This essay will abysmally fail at least that third criterion, but I suspect that Mark's Gospel might have succeeded on all three fronts.

The question we are entertaining at this symposium is this: When does Mark's Gospel end? For those who have studied the issue at all, there is an ancillary, though equally relevant question: When does the debate over the ending of Mark's Gospel end? This particular debate will be over tomorrow, but that won't be the end of the story. This conference, in fact, is intended to stimulate your thinking about the issues, getting you to wrestle with the questions far beyond tomorrow. Perhaps that is what Mark intended for his Gospel, too.

Introduction: Presuppositions

Before I get into the details of this notorious textual problem, I need to address the issue of presuppositions for a few minutes. Two of mine are that, of the four Gospels, Mark wrote first and John wrote last. As well, I hold to the Doddian school that John was not at all dependent on the Synoptic Gospels; in fact, he was most likely unaware of their specific contents, possibly even of their existence. This means that both Mark and John were writing, in a sense, a new literary genre that would later be called Gospel. Yet, they are radically different.

There are several ironies here: If Mark wrote first, he created a genre that would be followed by Matthew and Luke, thus giving some vindication to what he had done—since mimicry is a high form of flattery. Yet, Mark's Gospel was the least copied of the four, perhaps because it was almost entirely swallowed up by Matthew, with only the less appetizing parts left out. John, on the other hand, had no real literary followers, yet his is by far the most copied of the Gospels in early Christianity. The irony doesn't stop there. If Mark intended to end his Gospel at 16:8, as I will argue, it's a Gospel that leaves the reader hanging, wanting more. John, on the other hand, not only finishes his Gospel, he keeps on writing after he finished it. He finishes his Gospel twice! After his conclusion in chapter 20, he goes on with an appendix. It's almost as if John is a Baptist preacher who wants to add one more chorus of "Just As I Am" before he can conclude the service. Mark is rather more like a Methodist preacher who cuts his sermon short—giving a sermonette that contains deep thoughts, to be sure, but leaves his congregation with as many questions as answers.

Now, back to presuppositions. All of us have come with a set of them. Some have such strong presuppositions in one area that, regardless of the evidence in another, they are unwilling, or unable, to allow the second realm to cast doubt on the first. I want to bring to the conscious level some of the presuppositions that may be driving your view of this lengthy and difficult textual problem. I hope that you will think through these related disciplines as you wrestle with the question about Mark's ending. There are at least three important presuppositions to address.

First, one's view on source criticism plays a large role in deciding this particular issue. Let me put this plainly: If you hold to the Griesbach Hypothesis, or Matthean priority, you may have some trouble believing that Mark's Gospel ended at 16:8. The reason is not that the textual evidence is compelling for the Long Ending (LE), but a prior commitment about the synoptic problem is. If Matthew wrote first and Mark wrote last, would Mark really reject the rich material in Matthew 28 or Luke 24, and write a truncated Gospel that has no resurrection appearances by Jesus? It's much easier to believe that if Mark is last, he combined snippets from the other Gospels and wrote 16:9-20 than that he decided to excise the post-Resurrection narratives that were in Matthew and Luke.

Perhaps the most scholarly defense of the LE of Mark was written by William Farmer, a man who was already committed to Matthean priority when he wrote The Last Twelve Verses of Mark in 1974. In that monograph, he mentions with approbation the textcritical work of Harry Sturz in a lengthy footnote. Sturz considered the major texttypes to have all originated in the second century, and all independently of one another. His view is known as "the independent texttypes theory." His method was to determine the archetypal reading of each texttype by examining the best witnesses and determine the wording of the original based on a majority of texttypes—or, more exactly, a majority of the reconstructed archetypes of those texttypes. So, for Sturz, it wasn't a majority of manuscripts that had priority, but a majority of archetypes. Practically, though not theoretically, his views looked very much like the majority text theory. That is why he was one of the editors of the first published Majority Text.

Farmer does not say that he is following Sturz's method in his book, but one gets the distinct impression that he was heavily influenced by it. And in the draft for the second edition of The Last Twelve Verses of Mark, in his conclusion of the external evidence Farmer explicitly speaks against the "text critical tradition of... Westcott and Hort." In short, Farmer started with the Griesbach Hypothesis—in fact, he was the world's leading Matthean prioritist of the twentieth century—but he recognized that the Short Ending of Mark was a roadblock. So he sought out a text-critical method that would allow him to continue to maintain Matthean priority. At first, this method was most likely not clearly formulated in his mind, but twenty-six years later, he had more consciously and explicitly rejected reasoned eclecticism. Does this mean he had fully embraced Sturz's views? Ironically, the answer is no, because if he had, it would weaken his argument for the LE of Mark. In short, Farmer's source-critical views seemed to drive his text-critical decisions.

A second presupposition that may influence your view of Mark's ending is the whole field of textual criticism. Some of you have already become convinced of a particular text-critical theory, even though you may be uncertain about source criticism. If your views of the text are settled, you may be here simply to get more arguments for your conclusions about this textual problem. The proposed solutions to Mark's ending will probably not alter your theory, but your theory may dictate—or at least heavily influence—your solution to this textual conundrum.

Third, one's presuppositions about bibliology could have a large impact on how he or she views this particular problem. For example, if you believe in the doctrine of preservation—that God has preserved the Scriptures so that there must always be manuscript testimony to the original text—then you will not be open to the view that the original ending of Mark was lost. Now, it is possible that you would be open to the Gospel intentionally ending at v. 8, but if that interpretation is not rock solid, you just might opt for the LE because it puts you on safer bibliological ground.

Note, for example, what Dr. Wilbur Pickering said, when he was the president of the Majority Text Society, concerning the possibility that the ending of Mark was lost:

Are we to say that God was unable to protect the text of Mark or that He just couldn't be bothered? I see no other alternative—either He didn't care or He was helpless. And either option is fatal to the claim that Mark's Gospel is "God-breathed."... If God was powerless to protect His Word then He wouldn't really be God and it wouldn't make all that much difference what He said.... If God permitted the original ending of Mark to be lost then in fact we do not have an inspired text.

Nearly a century before Pickering made his statement, John Burgon, the famous antagonist to Westcott and Hort, wrote: "I am utterly unable to believe... that God's promise has so entirely failed...." He articulated two presuppositions in this one sentence that are relevant for the ending of Mark. First, his belief in divine preservation—and of a peculiar kind—prevented him from even entertaining the possibility that the LE was not authentic. Second, he assumed that there actually are explicit promises in the Bible about its preservation—in spite of the likelihood that none of the texts that are so used are speaking about the preservation of the written word.

If, however, the doctrine of preservation is not part of your credo, you would be more open to all the textual options. I, for one, do not think that the real ending to Mark was lost, but I have no theological agenda in this matter because I don't hold to the doctrine of preservation. That doctrine, first formulated in the Westminster Confession (1646), has a poor biblical base. I do not think that the doctrine is defensible—either exegetically or empirically. As Bruce Metzger was fond of saying, it's neither wise nor safe to hold to doctrines that are not taught in Scripture. I may be wrong in my view of preservation, but this presupposition at least keeps an open door for me for all the options in Mark 16.

My point in this preliminary treatment is to underscore the fact that we all bring a lot of presuppositions to the table that influence how we hear the evidence being presented; indeed, such presuppositions may even keep us from hearing the evidence.

Let me illustrate this point with a personal anecdote. I've mentioned Harry Sturz already. He was my first Greek professor at Biola University. At the time, he was teaching the only year-long college-level course on NT textual criticism offered in the country. As a young impressionable student, I came to embrace his views heartily. And I took several courses from him. When I came to Dallas Seminary, I studied under Harold Hoehner. He taught the Griesbach Hypothesis—that Matthew had written first. That view fit into my preunderstanding of textual criticism well. I also studied with Zane Hodges and learned the majority text position from him. In practice, I was a majority text man, though in theory I held to the independent texttypes view.

I joined the Dallas Seminary faculty and taught both Matthean priority and the independent texttypes theory. I went on to teach at another seminary, and I continued to teach both views. I studied the literature, and saw it only through a particular lens. I read some essays that were devastating to my theories, but I was able to brush them aside. And I thought that my arguments against them were decent.

A few years later, I returned to Dallas for my doctorate. While in the program, I wanted to take the course on textual criticism. Unfortunately, there was no one on faculty who could teach it since Zane Hodges had retired. So, I was asked to lead a doctoral seminar in textual criticism. Thus, when I asked if a class was going to be offered, I was told that it would be as long as I offered it! In retrospect, it's a good thing I didn't ask about a course in Egyptian hieroglyphics!

I had my work cut out for me, because I knew I could no longer give glib answers. I started by reading over 10,000 pages of text-critical material—to quickly get up to speed. I had to think through the entire paradigm. The more I studied, the more I came to the conclusion that Sturz's views were faulty—that the Byzantine text was not equal to the Alexandrian and Western. In short, I became a reasoned eclectic. It was an enormously painful methodological shift for me, since I had held to Sturz's view for seventeen years. Nevertheless, I did not completely reject the Byzantine text, but felt that it still had a place at the table.

Two months after my text-critical foundation had crumbled, I abandoned Matthean priority. In my mind, the two were not connected so much by their interdisciplinary nature as by their similar method of investigation. Both my text-critical and source-critical views had put a premium on the external data—and frankly, on a very mechanical approach to them. But, as Günther Zuntz noted, "At every stage the critic has to use his brains. Were it different, we could put the critical slide-rule into the hands of any fool and leave it to him to settle the problems of the New Testament text."

I began to read the same journal articles, monographs, and Festschriften in a different light. What had appeared to me to be dangerous and subjective viewpoints because they depended so much on the scholar's ability to get into the head of the scribe or evangelist, now looked like compelling arguments. What changed were not the arguments, but my presuppositions. I came to the deep conviction that evangelical scholars must be in the business of pursuing truth, regardless of where it takes us, rather than protecting our presuppositions. That has been the most liberating conclusion I've drawn in my academic career.

We each have various convictions about interlocking disciplines that affect our take on the last twelve verses of Mark. This means that this symposium will not really settle the issue for most of you. But it also means that it's good to take a step back and reflect on source criticism, your overall text-critical views, and how your bibliology impacts your view of the text. And I want to challenge you to wrestle with these interlocking presuppositions, and to be open to approaches that may be outside your comfort zone.

I don't doubt the integrity or scholarship of any of my colleagues, and I hope they don't doubt mine. But even though we are all looking at the same evidence, we are not all coming to the same conclusion. In part, we certainly do read the evidence differently. In part, we also each bring certain presuppositions to the discussion that impact our view of this textual problem.

With that introduction, I now turn to an investigation of the ending of Mark's Gospel. The overall objective will be to determine the reading that best explains the rise of the others. I will limit my examination to two aspects to the problem. First, I want to consider the external evidence—Greek manuscripts, ancient translations, and patristic writings. I will not only look at the hard data, but try to construct some reasons for why it looks the way it does. There are curiosities concerning Mark's ending that simply cannot be ignored. Second, I will look at internal evidence—though much more briefly. In particular, the focus will be on what the author was likely to have written. Internal evidence is not nearly as subjective as it may at first appear; likewise, external evidence is not nearly as objective as some might think.

The End of Mark's Gospel—External Evidence

As for the external evidence, the raw data can be quite deceiving: at least 95 percent of all Greek MSS and ancient versions have the LE. In fact, that number may be too low. I used to be impressed by the sheer volume of MSS on one side of a textual problem, but our investigation must take us deeper than that. In particular, the major question we need to answer is this: Which is more likely—that scribes would intentionally omit vv. 9-20 or that they would add these verses?

The Long Ending

The LE of Mark is not found in the oldest MSS, but it is found in the majority of MSS. And it is found in all the major texttypes—Western, Caesarean, Byzantine, and even the secondary Alexandrian. Thus, there is a broad geographical spread for these verses.

As well, it is witnessed to by several church fathers, most likely beginning in the late second century with Irenaeus. But there is a curious feature in this patristic evidence. Farmer makes a case that the LE was omitted by Alexandrian scribes because of two reasons. First, there seems to be a discrepancy between v. 9 and the other Gospels' accounts regarding the time of Jesus' resurrection. Second, vv. 17-18, with their promise of drinking poison and handling snakes without harm, caused embarrassment to early Christians. These two texts would have been sufficient reason for certain scribes to omit the entirety of vv. 9-20, according to Farmer.

There are severe problems with both of these arguments, however. Regarding the first issue, if any Gospel was the "odd man out" it would be Matthew, not Mark: Mark 16:9 is actually in agreement with Luke 24:1 and John 20:1 on the time of the Resurrection, while Matthew's wording seems to be at some variance. Mark, Luke, and John all say that Jesus was raised from the dead "early on the first day of the week," while Matthew says it happened "late on the Sabbath" or "after the Sabbath." In other words, Mark 16:9 is no more problematic than the Resurrection narratives in the other Gospels on this score. If scribes were prone to omit a verse because of this difficulty, it would have been Matt 28:1, not Mark 16:9.

So, v. 9 is not really the problem that Farmer thinks it is. This leaves vv. 17-18 as the sole reason for excision of all twelve verses by some scribes.

Second, then, regarding the drinking of poison and the handling of snakes, Farmer thinks that these verses would have been so problematic for many Christians that some even took the radical course of omitting all twelve verses from the Gospel. As the dust jacket of The Last Twelve Verses affirms, "Professor Farmer traces the history of the text tradition for omission back to Egypt, and argues that one important factor contributing to their omission was the dangerous teaching they seemed to contain: they appear to encourage Christians to handle deadly snakes and drink poisons to prove their faith...."

But if the locus of embarrassment was the second half of vv. 9-20, we might expect to see scribes deleting just these verses and retaining vv. 9-14. There is, however, no evidence at all that this happened. In fact, one could argue that just the opposite was the case: the early patristic writers allude to the second half of this pericope (vv. 15-20) far more often than they do the first half (vv. 9-14). Farmer gives the evidence himself: at least ten fathers quote from or allude to vv. 15-20 in a steady stream from the second to the fifth century, while no fathers mention the first half of the pericope until the fourth century! The trend, in fact, is so one-sided that Eta Linnemann saw this as evidence that Mark 16:15-20 was part of the original text of Mark, but vv. 9-14 were not, an argument similar to what Walter Schmithals would later make.

But the point of all this is simply to note that, contrary to Farmer's supposition that the early scribes excised the latter half of these twelve verses because of offensive language, the evidence shows that precisely the opposite happened: vv. 15-20 are quoted much earlier and more frequently than vv. 9-14. Yet, if as we have seen, v. 9 would cause no unusual turmoil, what shall we make of Farmer's argument? It is evident that he has not really discovered a reason why scribes would omit these twelve verses.

And this contention brings us to the major question that we are wrestling with: Why would the scribes do what they did to this text? The motive for excising the passage is too negligible to explain the short ending. And it almost certainly cannot explain why important, early, and diverse witnesses lack these verses.

If there is no adequate explanation for why some scribes would omit the LE, then the only alternative is that other scribes added it. Can a reason be found for this? Would scribes really be motivated to add these twelve verses? The reason seems so palpably obvious that it is almost needless to mention: If Mark's Gospel ends at 16:8, there are no Resurrection appearances by Jesus to his disciples. The Gospel marches on toward the Resurrection, beginning with Jesus' response to Peter's confession that after his crucifixion he would be raised from the dead (8:31). After his transfiguration, he again tells his disciples that he would die and rise from the dead (9:31). But again, his disciples "didn't understand and were afraid to ask what he meant." Finally, in 10:34, he again tells them that after his crucifixion he would be resurrected. In each of these scenes, the disciples are slow to get the message. We will come back to this issue later, but for now I only wish to point out that the Resurrection of Christ is prophesied by the Lord to his disciples three times in Mark. Yet, Mark's Gospel remarkably ends without any post-Resurrection appearances that fulfill the prophecy. Is this not reason enough for some early scribes to want to add something to this Gospel—anything!—that would have him appear before the disciples? Indeed, it is not just something but several "somethings" that were added after v. 8.

The Short Ending

Now, let's consider the evidence for the "short" ending—that is, the MSS that conclude the Gospel at 16:8. This discussion won't take much space because there aren't very many of these documents. But MSS should be weighed rather than counted, so we will want to look at their pedigree as well.

On the side of the short ending we have the following:

These two codices are the oldest Greek MSS for Mark 16. They also are the only "primary" Alexandrian witnesses to Mark 16 in Greek. This is a relatively pure form of the text that must be given full consideration when making any textual decisions. Metzger-Ehrman note that "textual witnesses connected to Alexandria attest a high quality of textual transmission from the earliest times. It was there that a very ancient line of text was copied and preserved.... the Christian scholars of Alexandria worked assiduously to preserve an accurate form of text."

Yet Codex Sinaiticus and Codex Vaticanus contain several thousand differences in the Gospels alone, suggesting that their common ancestor must be several generations back. In addition, we know that the Alexandrian text-form existed in the second century since nearly a dozen Alexandrian papyri are from that period, and because early versions and patristic writers utilized a Greek Vorlage that must have its roots deep in the second century. This MS, versional, and patristic evidence all points to the early second century as the beginning of the Alexandrian text, and it's a threefold cord that is not easily broken.

Further, although no papyri witness to Mark 16, one might cautiously enlist the support of 75 here. The text of B is closer to that of 75 than it is to any other MS. And 75 is a MS that antedates Vaticanus by at least a century. However, this papyrus only contains portions of Luke and John. Nevertheless, David Parker argued that since "we have no reason to consider the text of B in Matthew and Mark to be inferior to that in Luke and John,..." Vaticanus surely represents a text that already existed in the late second century. It can be added that the text of B, when compared to the text of 75, actually gives evidence of having more primitive readings than 75 does. Therefore, B is not a descendant of 75 but reaches back to their common ancestor. Thus, it is probable that B, even by itself, represents a text that existed early in the second century. In combination with , this likelihood is made stronger. In any event, even Sturz would argue that the reading of and B is a second-century reading.

One other comment regarding Vaticanus is in order: There is a large gap at the end of Mark in this MS. Vaticanus has three columns per page; Mark's Gospel ends at the bottom of the second column. The third column is left blank and Luke starts on the next page. The gap is clearly too small to allow for the LE, though Farmer tries to argue from this lack of evidence that the scribe knew about the LE "but disapproved" of it. A more sober assessment is provided by William Lane. He makes the case that the blank column after the conclusion of Mark is "a wholly singular phenomenon, for in Codex B a new book follows in the next column as soon as possible." He argues from this that the scribe of B knew of what is sometimes called the "Shorter Ending" or the "Intermediate Ending" (which we will discuss below), and made room for it.

However, even Lane's understanding of the gap in B is in doubt. First, this is not a "wholly singular phenomenon." Although it is certainly Vaticanus' normal custom to begin a new book at the top of the column following the conclusion of the previous book, this MS breaks that rule on four occasions. Tobit ends with one and a half columns to spare; 2 Esdras (Nehemiah) has only two lines in the first column, followed by the entire rest of the page blank; and Daniel concludes half way down the first column, with the rest of the page blank. All three of these books leave larger blanks than Mark does.

Second, it was recently discovered that the scribes of Vaticanus have indicated knowledge of textual variants by using two horizontal dots in the margin next to a line of text where a variant occurs. There are more than 700 such "umlauts" in the NT of Vaticanus, forty-three of which are in Mark alone. Thus, Codex B marks out half as many variants as the UBS text does! It's almost as if this is an ancient UBS Greek text. This is a remarkable discovery whose implications have yet to be fully explored. But, significantly, there is no umlaut at 16:8. Thus, the non-unique gap at the end of Mark and the lack of an umlaut here both seem to indicate that the scribe knew only that Mark's Gospel ended at 16:8. To put this another way: of the three other gaps in Vaticanus, not one is used to indicate knowledge of textual variation. So, to argue that this must be the case for the gap at the end of Mark is hardly compelling. And the fact that Vaticanus uses umlauts to indicate textual variants and that no umlaut is present at Mark 16:8 adds significant weight to the likelihood that the gap at the end of Mark does not imply knowledge of any other ending.

Codex 304, an otherwise unremarkable twelfth-century Byzantine MS, also ends with ἐφοβοῦντο γάρ in v. 8.

Besides the Greek MSS, there are a few ancient versions that lack vv. 9-20.

The Sinaitic Syriac MS (Syriacs) displays the oldest form of the Gospels in the Syriac language. This MS dates from the fourth century, but it represents a text from the late second or early third century, for the most part following the Western texttype. The Sinaitic Syriac MS concludes the Gospel at v. 8.

Approximately 100 of the Armenian MSS, including almost all of the earliest ones, lack the LE. It has been conclusively demonstrated that the original Armenian text did not have these verses. Even William Farmer admitted as much. The Armenian has roots in the early fifth century. Joseph Alexanian, one of the world's leading Armenian scholars, notes that all extant Armenian NT MSS belong to the second revision, a revision that was based on Greek MSS brought from Constantinople shortly after the Council of Ephesus in 431. As well, he notes that the text-form of the earliest of these MSS is either Caesarean or proto-Byzantine (aka "Early Koine"). Almost all Byzantine MSS extant today have the LE, but the Armenian version demonstrates (i.e., if it is truly Byzantine instead of Caesarean) that this was not always the case.

The two oldest Georgian MSS end at v. 8. This version finds its roots in the fifth century.

One Sahidic MS lacks any material past v. 8. The Sahidic is the oldest version of the Coptic NT, originally produced in the early third century, though, on the surface, whether this lone MS represents that text is more difficult to assess. However, the other Sahidic MSS, as we will soon see, confirm that the Sahidic original did not go beyond v. 8.

In addition to the Greek MSS and versions, there are several early patristic writers who may be witnesses to the Gospel ending at ἐφοβοῦντο γάρ. Origen is silent about the LE, even though he had opportunity to mention these verses. Before him, Clement of Alexandria is also silent about the LE, though he also never quotes from Matthew 28. Regarding the argument from the silence of both Clement and Origen, Parker notes:

Those who wish to argue for the originality of the Long Ending point to the weakness of this argument, and not unreasonably. It cannot be argued that the Long Ending was unknown to or rejected by Clement and Origen. But there is a more important point: neither can silence indicate that the Short Ending was unknown. The presence of evidence for the Long Ending is demonstrable. That there cannot be similar evidence for an absence of text requires us to accept that there is no evidence against the existence of the Short Ending in the second century.

In other words, since this particular textual problem involves the addition of a dozen verses, rather than the alteration of the same, there is simply no way to verify whether Clement and Origen knew of the LE. All we can say is that there is no evidence that they did.

Once we get to the fourth century, however, the situation looks decidedly different. Eusebius mentions MSS that end at v. 8 and those that end at v. 20. In discussing the differences between Matthew 28:1 and Mark 16:9, he says:

This can be solved in two ways. The person not wishing to accept [these verses] will say that it is not contained in all copies of the Gospel according to Mark. Indeed the accurate copies conclude the story according to Mark in the words... they were afraid. For the end is here in nearly all the copies of Mark.

Eusebius thus indicates that most of the MSS in his day—the early fourth century—ended at v. 8; indeed, "the accurate copies conclude" here. That this is his opinion seems to be evident from the fact that the Eusebian Canons made no provision for Mark 16:9-20.

As might be expected, Farmer vigorously debates whether the statement about the accurate copies of the Gospel ending at v. 8 really represented Eusebius's view of things, or whether Eusebius was simply quoting from an earlier source. On the one hand, Farmer is not giving Eusebius enough credit for owning what is clearly stated as his opinion; on the other hand, if Eusebius is borrowing from an earlier source, then this shows that MSS with the short ending were in the majority even before the early fourth century. Farmer, in fact, suggests that Eusebius's statement about the more accurate and more plentiful MSS went back to Origen, a concession that implies that the MSS ending at v. 8 were in a majority at least a century earlier than the time of Eusebius.

At the beginning of the fifth century, Jerome also notes that the LE is found in "scarcely any copies of the Gospel—almost all the Greek codices being without this passage...." His statement has been discounted since it is evidently a paraphrase of Eusebius's statement and since he did include the LE in the Vulgate. There are two reasons, however, why we should accept Jerome's statement as his own opinion. First, he adds information not found in Eusebius—viz., that almost all of the Greek MSS that he was acquainted with lacked the LE. This is an important point because Jerome's major work was the Latin Vulgate. He was very familiar with both Greek and Latin MSS. Yet he qualifies Eusebius's statement to refer only to the Greek MSS. The implication may be that the Latin MSS he knew often—or at least more frequently than the Greek—had the last twelve verses.

Second, Jerome was well acquainted with several MSS of Mark's Gospel. For example, he quotes from some verses that were found between vv. 14 and 15 of the LE, noting that he has found this material "in some exemplars and especially in Greek manuscripts of Mark in the end of his Gospel...." Until the twentieth century, the only evidence we had for such verses was Jerome's quotation of them. But when Codex W was discovered, we finally had concrete proof of the verses that Jerome quoted. (These verses are known as the Freer Logion because they are found in Codex W, discovered by Charles Freer.) Jerome had seen this passage "especially in Greek manuscripts" but also in others. He thus makes a quantitative statement about three endings to Mark's Gospel: most of the Greek MSS ended at v. 8; some of the Greek MSS as well as a few others added material between vv. 14 and 15; and a few MSS, almost none of them Greek, included vv. 9-20. All this shows that Jerome was well aware of the variations at the end of Mark's Gospel precisely because he had access to numerous MSS. Thus, just because he was paraphrasing a statement from Eusebius is not sufficient reason to think that the evidence he was describing did not apply in his day.

Jerome's statement has also been discounted because he included the LE in the Vulgate. Why would he do that? Perhaps for the same reasons that it is included in Bibles today—call it antiquity, tradition of timidity, or not wanting to rock the boat too much. In ad 400, a riot broke out in Tripoli when Jerome's translation of Jonah 4:6 was read publicly. He used the word "ivy" (hederem) instead of the traditional "gourd" (cucurbita) to describe the plant that gave Jonah shelter. Augustine wrote to Jerome about the situation, pleading with him to temper how much he tampered with the traditional text. Even though Jerome wrote a defiant letter back, it is likely that there were limits to his alterations. If a riot had broken out over the description of a plant, how much more chaos could result if Jerome had omitted Jesus' appearance to his disciples in Mark 16?

Finally, Victor of Antioch, in the fifth or sixth century, notes that "very many copies" of the Gospel ended at v. 8, and "very many copies" ended at v. 20. He weighs in at this point and says that the more accurate MSS included vv. 9-20. Victor is important because his commentary was extremely popular, becoming the "established commentary on Mark for the later church."

The patristic testimony thus reveals a very interesting trend: from the earliest discussion on the authenticity of this passage, the fathers indicate that most of the copies of Mark ended at 16:8. Yet, in later centuries, the short ending was increasingly looked on unfavorably, and in the standard commentary on Mark of the Middle Ages the short ending was rejected. Putting this on a trajectory, it takes little imagination to realize that what became the majority reading in the Middle Ages started out as a minority reading.

If this were all there were to the external evidence, it should be enough to convince us that Mark's Gospel did not originally contain the LE. However, there is much more evidence that this is the case. First, there are alternative endings to the Short and Long Endings that made their way into the MSS. Codex Bobbiensis (itk), an Old Latin MS written in c. ad 400, shows evidence of having been copied from a second-century papyrus. This is the most important Latin MS of the Western texttype for the Gospels. This MS does not have the LE, but instead has what is known as the Short Ending (SE) or the Intermediate Ending (IE) after v. 8, as follows: "But they reported briefly to Peter and those with him all that they had been told. And after these things Jesus himself sent out through them, from east to west, the sacred and imperishable proclamation of eternal salvation."

This ending is patently secondary. It is simply added to the abrupt ending of v. 8 as a way to conclude the Gospel. But it is a colorless conclusion. Why would it have been added to the Gospel? The obvious answer is that the Gospel MS that the scribe had in front of him ended at v. 8. Thus, Bobbiensis should be added to the witnesses for the short ending. In the least, it does not support the LE.

In several other MSS this Intermediate Ending is added before the LE. The list includes "four uncial Greek manuscripts of the seventh, eighth, and ninth centuries (L Ψ 099 0112 al),... the margin of the Harclean Syriac, several Sahidic and Bohairic manuscripts, and not a few Ethiopic manuscripts...." In none of these witnesses is the Intermediate Ending added after the LE. Yet the Intermediate Ending is both a conclusion to the Gospel and is implicitly contradictory to the LE, because the Intermediate Ending speaks of Mary telling the disciples that Jesus had risen, resulting in his commissioning of them, while the LE has Mary also telling the disciples of his Resurrection, resulting in their continued disbelief! If the Intermediate Ending were added after the LE, this tension would at least be more relaxed. Yet no MS does this.

The MS situation, therefore, points to three important implications. First, the MSS that added both the Intermediate and Long Endings imply that their ancestors only had the Intermediate Ending. As Metzger observed, "No one who had available as the conclusion of the Second Gospel the twelve verses 9-20, so rich in interesting material, would have deliberately replaced them with a few lines of a colorless and generalized summary." He rightly concludes that these witnesses should thus be added to those that end in v. 8.

Second, once a reading made its way into the text, it was very hard to dislodge it, even if that reading ended up contradicting what was added later. The fact that the Intermediate Ending is found in eight Greek MSS, all the Sahidic MSS (except for one that ends at v. 8), and several other versional witnesses, is concrete testimony to this principle.

And third, the presence of the Intermediate Ending demonstrates that scribes were not satisfied with the Gospel ending with "for they were afraid."

Again, we ask the question of motive: Which is more likely—that scribes would intentionally omit vv. 9-20 or that they would add some ending because Mark 16:8 ends without Jesus appearing to his disciples? The Intermediate Ending speaks eloquently to both of these alternatives: scribes were prone to add material, even if it created contradictions, to a Gospel that ended with the fearful women.

In addition to these witnesses, there are several MSS that indicate some doubt about the authenticity of the LE. They do this in one of two ways: First, the scribe may add a note after v. 8, such as that found in Codex 22: "The end. In some copies the evangelist ended here, but in many this also"—referring to vv. 9-20 which follow. Similar notes are found in key members of Family 1 (1 209 1582), as well as in several other codices. Altogether, a dozen MSS have such a note. Second, the scribe might simply place an asterisk or obelisk in the margin, indicating doubt about these verses. Such a symbol is found in at least five MSS. That these same scribes nevertheless included vv. 9-20 reveals a basic principle that ancient copyists followed: if in doubt, don't throw it out!

As many textual critics have observed, one of the key evidences that a text is unstable is a high degree of variation and displacement. If so, the ending of Mark's Gospel is a poster child for this principle. David Parker points out six different endings for Mark, along with multiple variations within them. I have only scratched the surface of the textual fluctuation in my presentation of the external data. But if we pause for a moment and take it in, the question that we raised earlier surfaces again: Why are there so many differences in the MSS here? The issue is not simply the LE vs. the SE, as Farmer seems to imply. If that were the case, then we might well expect all the MSS to line up in either camp. Instead, some MSS end at 16:8, others add an Intermediate Ending, while others indicate doubts about the authenticity of vv. 9-20. To be sure, the vast majority of MSS include vv. 9-20, but the diversity and age of the witnesses against the LE have to be accounted for.

Why is it that this Gospel, and only this Gospel, has major textual upheaval at the end? The reason can't be due to discrepancies about the Resurrection accounts because whether Mark is in the mix or not, there are still tensions in these accounts. Mark 16:9-20 is hardly the most difficult post-Resurrection narrative to harmonize with the other Gospels. And it can't be because of the promise about handling snakes and drinking poison because this portion of the LE is the most secure, patristically speaking, of the entire pericope. As we have seen, the only alternative is that scribes were uncomfortable with a Gospel ending without any Resurrection appearances.

Which is easier to believe—that scribes cut out rich theological material because they didn't like the abrupt change between vv. 8 and 9, even though that would obliterate any Resurrection appearances by Jesus to the disciples in this Gospel? Or that they felt that the Gospel came to a close too abruptly and needed to have a proper ending in which the Resurrection appearances of Jesus were included? The answer should be obvious: When scribes find small segments of text that are problematic, they do some plastic surgery. But they don't amputate the leg because of athlete's foot!

One final comment is needed on the external evidence. The MSS, versions, and patristic evidence on behalf of the short ending are early and widespread. They represent the Alexandrian (especially the primary Alexandrian), Western, and Caesarean texttypes, possibly even the proto-Byzantine. The short reading is found in the best witnesses of the three most important early versions as well—the Latin, Coptic, and Syriac. There is no obvious connection between all these diverse witnesses, showing that this reading must go back to a much earlier stage in the transmission of the text. Further, as the centuries rolled on, the LE was at first considered inauthentic, then timidly accepted, then finally considered to be part of the original text. Yet even then, many scribes registered their doubts about this ending. If the internal evidence on behalf of the LE looks at all suspect, we should consider the matter closed: Mark's Gospel did not originally have vv. 9-20.