Mark's Ending: The Long and Short of It
I have long been interested in the debate over the authenticity of Mark 16:9-20, known as the long ending of Mark (LEM). A good overview of this debate is contained in Perspectives on the Ending of Mark: 4 Views by Daniel Wallace, David Alan Black, Keith Elliott, Maurice Robinson, and Darrell Bock. Each author takes a different perspective on the ending of Mark:Wallace = LEM is not original. Mark ended his gospel at 16:8. (In Bock’s closing summary of the book, he noted his agreement with this position.)
Elliott = LEM is not original. Original ending has been lost.
Robinson = LEM is original.
Black = LEM is original, but was added by Mark as part of a “second edition” to round our Peter’s lectures.
Of the four, I think Wallace presented the most convincing case, and Black the least convincing. I will summarize the evidence/arguments for and against the LEM in hopes that this will help you sort through this issue as much as it helped me.
External evidence for the ending of Mark
Greek manuscript evidence
- Manuscripts (MSS) containing the LEM
- The LEM can be found in at least 95% of all Greek and versional (translations) MSS.
- The LEM is represented in all major Greek text-types (including secondary Alexandrian MSS). Such broad geographical distribution/representation argues in favor of its authenticity.
- Alexandrian: C L Psi 083 099 33 157 579 892 1241 2427 L-1602 bo
- Western: D aur c dsupp ff2 l n o q vg syc
- Caesarean: W Delta Theta f1 f13 28 565 700 armmss
- Byzantine: A Byz Lect syp syh sypal
- MSS lacking the LEM:
- Vaticanus (B)
- Sinaiaticus and B = 2 oldest Greek copies of Mark’s Gospel. They are dated early- to-mid-4th century.
- Scribe D of Sinaiticus is likely one of the scribes who copied B, which would demonstrate that these two manuscripts came out of the same scriptorium.
- Vaticanus was likely copied from a 2nd century text, since its readings are more primitive than p75 (early 3rd century MS).
- 304 (12th century commentary on Mark)
- Two other manuscripts end at 16:8, but only because the codices are missing pages, so they do not count.
- MSS containing an intermediate ending (IE) after 16:8, followed by the LEM:
- L, Psi, 083, 099, 274mg, 579, l, 1602
- None is older than 7th century
- MSS containing the LEM with an expansion (“Freer Logion”) following 16:14:
- MSS indicating doubt about the authenticity of the LEM:
- 12 MSS have a marginal note indicating the scribe’s doubt about authenticity, such as Codex 22: “The end. In some copies the evangelist ended here, but in many this also” (he went on to copy vs 9-20).
- At least five MSS have an asterisk or obelisk in the margin to indicate doubt (138, 264, 1221, 2346, 2812), and yet they included the verses, demonstrating the basic scribal principle of ‘when in doubt, don’t throw it out.’
- The Syniatic Syriac manuscript (sys) is the oldest Syriac MS. It’s a 4th century MS, but copied from a 2nd or early 3rd century MS. Lacks the LEM.
- ~100/220 Armenian MSS (early 5th century) lack the LEM.1
- 4 MSS place the LEM elsewhere, including at the end of Luke, the end of John, and even after “according to Mark” at the end of the Gospel (ancient MSS included the title of the book at the end, rather than the beginning).
- The Armenian MSS reflect a Byzantine text, which would show the Byzantine did not always contain the LEM.
- One Armenian MS—add. 21932—lacks the LEM, but writes 16:7-8 in very large letters as if to use up space that had originally been calculated so as to possibly include the LEM.
- Two oldest Georgian MSS (5th century roots, translation of Armenian) lack the LEM
- 1 Sahidic MS (oldest Coptic, early 3rd century) lacks the LEM
- Codex Bobbiensis (itk) is a 4th century Old Latin MSS that lacks the LEM, but adds the IE after v.8.
- All Sahidic MSS (except one) and several other versional witnesses contain the IE. None is older than the 7th century.
- Justin Martyr may have alluded to Mk 16:20 in Apology 1.45. He used a phrase consisting of three, rather stylized words that appear in Mark’s text (“having gone forth, preached everywhere” = exelthontes pantachou ekeruxan), albeit in a different order than Mark’s Gospel (he transposed the last two words).
- In the way of rebuttal, it’s doubtful that Justin Martyr was alluding to Mk 16:20. He was summarizing vast amounts of material. Pantachou is more common in Luke (appears only 1x in Mk), and given that Acts shows the apostles going everywhere, it’s not hard to see how Justin could have come up with the idea/phrase without any knowledge of Mk 16:20.
- The first clear reference to a verse from the LEM is from Irenaeus in ~A.D. 180. He quoted Mk 16:20 to show how the gospel ends. We have evidence, then, that the LEM existed at least as far back as the mid 2nd century.
- Eusebius is the first church father (4th century) to specifically mention the fact that some manuscripts contain the LEM while others do not. He noted that the LEM is lacking in “nearly all copies of Mark.” He considered the MSS with the short ending to be the “accurate copies of the Gospel according to Mark.”
- Jerome (5th century) notes the same as Eusebius, though he seems to have recognized that the Latin MSS usually contained the LEM (and hence he put it in his Latin Vulgate, which became the only translation used by the Catholic Church for ~1500 years). He also noted how some Greek MSS and exemplars added material between vs. 14-15 (the “Freer Logion” as evidenced in MS W).
- Victor of Antioch in the 5th/6th century noted that “very many copies” of Mark ended at 16:8, and “very many” at 16:20. He thought the LEM was more accurate.
- 10 Church Fathers from the 2nd to 5th centuries cited portions of the LEM.
Internal evidence for the ending of Mark
Internal evidence for the authenticity of the LEM
- Total words in the LEM = 166
- LEM words occurring elsewhere in Mk:
--in identical form = 106 (63.8%)
--in parallel parsing or declensional forms = 39 (23.5%)
--in parallel compounded or non-compounded forms = 9 (5.4%)
- Words in LEM with some related parallel elsewhere in Mk = 154 (92.7%)
- Some words that appear in 16:9-20 appear in compounded or non-compounded forms elsewhere in the gospel.
- In the 12 verses preceding 16:9-20 (15:44—16:8) there are 17 words, phrases, or stylistic usages not used elsewhere by Mk. Many other segments of Mark can be looked at, and they turn up a similar number of “non-Markan” words.
- Black argues that the internal evidence only shows that the LEM is different than the rest of Mark. While interesting, it does not prove whether Mark wrote it. Internal evidence is too subjective, and can only serve as corroborative evidence.
- Most of the non-Markan words are rare even in the other gospels.
Internal evidence for the inauthenticity of the LEM
- “There is not a single passage in Mark 1:1—16:8 comparable to the stylistic, grammatical, and lexical anomalies in 16:9-20.”—Wallace
- There is a strange jump between verse 8 and 9. The opening words of the LEM in verse 9 suggest that Jesus is the subject in the preceding context rather than Mary, and Mary is introduced in verse 9 as if for the first time. Furthermore, verses 19-20 do not continue with, or explain what appears in 16:1-8. It looks as if the two are not connected either literarily or grammatically, which implies that 16:9-20 is a later addition.
- Matthew and Luke both used Mark as a template, so they both have similar content. But when it comes to their resurrection accounts, they look very different from one another. The best explanation for this divergence is that they lost their shared template in Mark at that precise point; i.e. the copy of Mark they were working from lacked the LEM (which contains resurrection appearances).
- Verse 8 ends with the particle gar (“for”). Some argue that Mark could not have ended with gar, thus the original ending must be lost.
Assessing the Evidence
When it comes to the manuscript evidence, how should we best judge? If we are just counting manuscripts, the LEM wins the authenticity contest hands-down. But that is from our own temporal perspective. While the extant MSS containing the LEM far outnumber the MSS lacking the LEM in our day, the opposite was true in the 4th and 5th centuries according to Eusebius and Jerome. If authenticity is determined solely by counting MSS, we end up with an ironic situation in which someone living in the 4th century would have to conclude that the LEM is an interpolation, while we in the 21st century have to conclude that the LEM is original. Who’s to say which century should be considered the standard?
In addition to this practical difficulty, there is a more fundamental problem with this approach: in principle, counting MSS cannot tell us which reading is original because it is always possible that an errant MS was copied more than the original. In the same way that repeating a lie 100 times does not make it true, a particular reading is not shown to be original simply because it is copied hundreds or thousands of times.
It’s clear that both the LEM and the short ending of Mark (SEM) have early attestation. Our earliest Greek MSS and versions lack the LEM. And yet, we know the LEM appeared in at least some manuscripts as early as the mid-2nd century since Irenaeus was aware of it. But given the testimony of Eusebius and Jerome, the SEM appears to have been the most common reading in the early church (it’s important to note that Eusebius and Jerome were outside Egypt, proving that the SEM was not some aberration unique to the Alexandrian tradition). By the end of the 5th and beginning of the 6th centuries, however, Victor of Antioch indicates that the MSS evidence for both readings was about equal. And we know from the extant MSS that from that time on, the LEM eventually became the majority reading. It would appear, then, from both the MSS, versional, and patristic evidence, that the SEM was the reading of the majority of MSS in the early church, and only came to be a minority reading by the 7th century. So it took ~500 years for the SEM to be superseded by the LEM. If the SEM was the majority reading early on, it is more likely to be the original reading.
Vaticanus’ blank space
Normally, if a book ended mid-column or mid-page, scribes would continue copying the next book immediately after. No space was spared. Vaticanus, however, has a blank space between the end of Mark and the beginning of Luke. Some, such as Robinson, argue that this space was left blank because the scribe was aware of the LEM, and left space for its later inclusion if necessary.
Most of Vaticanus was transcribed using three columns per page. Mark’s Gospel ends in the middle of the 2nd column. Rather than beginning Luke’s Gospel immediately after, however, the scribe began copying Luke on the next page. Could this space have been intended for the LEM? No. There is not enough room to fit the 163 word contained in the LEM, and too much room for the IE. If the scribe was truly leaving room for the LEM, he should have begun his copy of Luke somewhere toward the end of the first column on the next page. The blank space following Mark’s Gospel, then, does not discount Vaticanus as a witness against the LEM.
It should be noted that this is not the only section in Vaticanus in which a blank space is found following the conclusion of a book. This phenomenon occurs three other times as well:
- Tobit has 1 ½ blank columns
- The end of Nehemiah only takes up 2 lines of the 1st column. The rest of the page is blank.
- The end of Daniel only takes up ½ of the 1st column. The other 2 ½ columns are blank.
Why would the scribes of Vaticanus leave spaces between these books, but not others? Perhaps it is due to shifts in genre (history to poetry, etc.).
The gap after Nehemiah is explicable: Psalms follows Nehemiah, but Psalms was transcribed into two columns as opposed to three, so it was best to begin Psalms on a new page. Daniel’s 2 ½ column gap is explained by the fact that Daniel ends the OT. It is not clear why the scribe left a gap between Mark and Luke, however, it is unlikely that he did so because he was leaving space for a particular ending, whether it be the LEM, the IE, or some alternative ending. In fact, it would appear that the scribe was not aware of any other ending for Mark. Throughout Vaticanus one will find umlauts. Scholars have determined that these marks were meant to indicate the scribe’s knowledge of textual variants. Interestingly, no umlauts appear after 16:8, and thus there is no reason to believe the scribe left the space blank to acknowledge textual variants in the ending of Mark.
Origin of differences
There are six different endings to Mark in our extant MSS. Everyone, regardless of what they think about the original ending of Mark, must account for the multiplicity and diversity of the endings. One must explain how, if Q was the original ending, it gave rise to R, S, T, U, and V. Likewise, if one believes T was the original ending, they must explain how T could have given rise to readings Q, R, S, U, and V. The reading that can best explain the origin of the alternative endings is most likely to be original. I am persuaded that the SEM, rather than the LEM, best explains the origin of the alternative readings.
What is more likely – that scribes would intentionally add verses not original to the manuscript they were copying, or omit 12 verses that appeared in their exemplar (163 words)? An examination of the MSS tradition shows that scribes generally followed the practice of ‘when in doubt, don’t leave it out.’ That’s why some scribes included both the IE and the LEM in their MSS (even though the IE and LEM contain contradictory details). They reasoned that something that does not belong in the text could always be struck out later, but if one eliminates some questionable text that does truly belong there will be no room to add it in later, possibly causing a whole lineage of errant manuscripts to make their way into the MS tradition. This shows us just how difficult it was to remove a reading once it had been introduced into the MSS tradition. Given our knowledge of scribal practices, it stands to reason that they were unlikely to omit the LEM, and thus it is unlikely that the LEM is original.
Despite the evidence of scribal practices, those who believe the LEM is original maintain that the SEM arose because some scribe omitted the LEM from his MS. But how could this happen? It couldn’t be due to an oversight. It’s hard to miss 163 words! Some scholars, such as Robinson, suggest that the scribe(s) intentionally omitted the LEM because it contains questionable and embarrassing theological claims (particularly the sign gifts). This hypothesis is unlikely, however, because the theologically questionable claims are limited to the last few verses of the LEM, not scattered throughout the periscope (and ironically, the church fathers who quoted from the LEM always quoted from the second half, not the first half, so they clearly didn’t find the passage theologically troublesome). So why wouldn’t the scribe(s) just eliminate a verse or two that he found questionable and embarrassing, rather than eliminate the entire pericope? When “activist” scribes found something they considered to be problematic in the text, they performed plastic surgery, not amputation.
Some have suggested that a scribe might have omitted the LEM from his MS because it contains details that appear to contradict the accounts of the other Evangelists. This is unlikely for two reasons. First, the same sort of tension exists between the other Evangelists’ accounts. There are tensions between John and Luke, between Luke and Matthew, and between John and Matthew, and yet we don’t find scribes eliminating pericopes from Matthew, Luke, and John. Why would they only do so in Mark?
Secondly, why wouldn’t the scribe just amend/remove the portions of text containing the discrepancy? Why delete the entire periscope when the vast majority of it was not problematic? We have plenty of examples in which scribes would amend the wording of a text to harmonize it with other gospels, but no examples of them eliminating large pericopes. Indeed, there was no need for them to do so.
While there is no good reason to omit the LEM if it was original to Mark’s Gospel, there are good reasons to add the LEM if Mark’s Gospel ended at verse eight. First, Jesus predicted His resurrection three times in Mark’s Gospel. While 16:6 depicts the angels announcing Jesus’ resurrection from the dead, one would expect for the book to end with someone actually seeing Jesus alive from the dead. Strangely, no one does. We would naturally expect an ending in which the triumphant Jesus proves to His disciples that His prophetic announcements have been fulfilled. The story seems incomplete. It ends abruptly, like a cliff-hanger used in a TV show that announces at the end, “to be continued.”
Secondly, all of the other canonical gospels contain pericopes about Jesus’ resurrection appearances to His followers. Mark’s account, in contrast, does not. It is the odd-man-out, motivating someone to “fix” Mark by adding a periscope containing Jesus’ resurrection appearances so that it conforms to its canonical brothers.
Thirdly, the story ends on a rather sour note. The last picture we are provided is of the women followers trembling in fear and silence. This doesn’t seem like a fitting ending to a story about the Son of God’s triumph over death.
Finally, the angel told the women to instruct the disciples to journey to Galilee to see Jesus alive from the dead. One would naturally expect for Mark to chronicle that meeting, and yet if Mark ends at 16:8, no such report is provided. Again, the account seems incomplete.
Surely these four reasons can explain why some scribes would be motivated to construct an ending for Mark’s gospel. And because they were adding text rather than deleting text, it also explains why different endings appear in the MS tradition: different scribes in different locales – each being dissatisfied with Mark’s ending – constructed different endings to “complete” Mark’s Gospel. Several of these alternative endings made their way into the MS tradition. Only if the LEM is not original can we explain the rise of multiple endings.
The seam between verses 8 and 9
In Mark 16:1-8 we read about Mary Magdalene, Mary the mother of James, and Salome journeying to Jesus’ tomb early Sunday morning to anoint Jesus’ body. When they arrive, the tomb stone was rolled away, and they are greeted by a young man who informs them that Jesus is no longer in the tomb because He has been raised from the dead, and that Jesus would appear to His disciples in Galilee.
One would expect for this periscope to be immediately followed by a report of the women running off to tell Jesus’ disciples what they had seen and heard. Instead, we read: “Early on the first day of the week, after he arose, he appeared first to Mary Magdalene, from whom he had driven out seven demons” (16:9). It’s not until 16:10 that the story picks up where 16:8 left off.
If Mark was the author of 16:9-20, why would he mention that it was “early on the first day of the week” since the timing of the periscope had already been stated in 16:2? And why would Mark speak of Mary Magdalene as if she was being introduced for the first time in the story (“Mary Magdalene, out of whom he had cast seven devils”), when she was introduced in 15:40, and spoken of again in 15:47 and 16:1? The detail that Jesus had cast seven devils out of Mary would be more fitting for the first mention of Mary, not the last. It’s as if 16:9-20 was written independently of Mark’s Gospel, and simply appended to the end by a scribe who may have thought it to be a better ending than Mark’s original ending.
Why do Matthew and Luke Differ So Much?
Most scholars have concluded that Mark was the first Gospel to be written, and that Matthew and Luke based their accounts on Mark’s. Throughout the gospels of Matthew and Luke, one can see striking similarities in both content and wording. Only 132 of Mark’s 11,025 words find no parallel in either Matthew or Luke. A whopping 97% of Mark’s gospel is contained in Mt (about 612 of 662 verses), and 88% in Luke. But when it comes to Matthew and Luke’s resurrection accounts, they differ from each other as much as they differ from Mark. Why did they diverge so much from Mark and each other precisely at this point? The best explanation for this divergence is that they lost their shared template in Mark at that precise point; i.e. the copy of Mark they were working from lacked the LEM (which contains resurrection appearances).
Could it end at 16:8?
Some scholars, such as Elliot, do not accept the LEM as original, but neither do they believe that Mark intended his gospel to end at 16:8. They argue that Mark wrote an ending that has since become lost. There are literary and grammatical arguments for this position. We just rehearsed many of the literary arguments above (abrupt ending, sour note, loose ends not wrapped up, etc.). As for grammatical arguments, it is noted that in the Greek text, 16:8 ends with the particle gar (for), and argued that no piece of Greek literature would end with a particle. But an ancient Greek work ending in gar was discovered in 1972, proving that such a grammatical phenomenon, while uncommon, was not impossible. Even if this work had not been discovered, this grammatical argument is not very persuasive since we have a lot of evidence for Greek sentences ending in gar. If a sentence could end with gar, who’s to say a book cannot? After all, books end in with a sentence!
Why would it end at 16:8?
Why would Mark end his gospel at 16:8? Of all the Evangelists, Mark tends to leave it up to the reader to form his own opinion about Jesus. Mark takes them right up to the angelic announcement of Jesus’ resurrection, but leaves it up to the reader to decide for himself whether or not Jesus had risen as He said He would.
Aren’t such “cliff-hanging” literary techniques a modern phenomenon? While such endings are rare in the ancient world, they did exist in Graeco-Roman literature, and even the NT. For example, Acts leaves off without telling us the result of Paul’s preaching in Rome and the result of his appeal. In Jesus’ parable of the prodigal son, we are left wondering what happened to the brother. Even Mark leaves us hanging in 9:32 and 16:8. Interestingly, both pericopes use the same imperfect verb, ephobounto (afraid) we find in 16:8. According to Bock, “fear” is a theme in Mark. While some people’s fear leads to faith, other people’s fear leads to rejection of Jesus. In Bock’s words, “In Mark fear can paralyze or lead to faith. The choice is with the one who fears.” And thus, Mark ended his book with disciples in fear. Will their fear lead to faith, or unbelief? The women represent the position of the reader. They too must decide whether to put their faith in Jesus, or reject Him in fear.
How likely is it that Mark penned an extended ending to His Gospel that has since been lost? Not very. While the most vulnerable sections of a scroll were the beginning and the end, usually the beginning suffered because it was exposed. The end of the document was rolled up and protected from exposure. The same was not necessarily true of codices, however. In codices, both the beginning and ending were vulnerable. But the church did not start using codices until the second century. It is unlikely that if Mark’s Gospel did not end at 16:8, that the original ending could have been stamped out of the tradition so quickly after having been copied for 50+ years with the original ending. Not even a single church father was aware of it. Again, that seems unlikely. It is more reasonable to believe Mark ended his gospel at 16:8.
While the SEM and LEM are both old readings, the manuscript, patristic, versional, literary, and grammatical arguments favor the SEM as the original reading. While the LEM ultimately became the majority reading, it started out as the minority reading in the early- to mid-2nd century. While many of the details found in the LEM are consistent with the pericopes found in the other Evangelists, the LEM itself does not appear to be original to Mark’s gospel, and thus should not be viewed as inspired Scripture by Christians, or used for doctrinal purposes.
1. All the versions have not been examined yet, but among the 200 Armenian MSS that Colwell examined, 99 lacked the LEM, 33 indicated that vs 9-20 were doubtful, and four placed the LEM elsewhere.
IBS | Statement of Faith | Home
| Browse by Author | Q
Links | Virtual Classroom | Copyright | Submitting Articles | Search