Why can't the gospel of Mark be as late as 110 CE?

I thank LangLangC for his answer, which helpfully points out that at least some modern scholars, motivated by the hypothesis of Marcionite priority, are willing to consider much later dates for Mark. Although the answer doesn't point to any actual dating arguments or evidence, and Marcionite priority seems to be a bit of a fringe theory, the answer does suggest where to look for such arguments and evidence, which would be in recent scholarly debates about this hypothesis. Using this lead, I did manage to turn up the following.

One of the two most visible proponents of the hypothesis, Markus Vinzent, has a blog post in which he quotes a lengthy footnote by Udo Schnelle, in James W. Thompson, 'Die ersten 100 Jahre des Christentums', an English-langauge 3rd ed., p. 466-467, note 41. (Vinzent shows some irritation at being dismissed in a footnote.) If Schnelle is an expert on these topics, and is trying his best to shoot holes in Vinzent's theory by using dates, then presumably this would be a good summary of current arguments that would limit how late Mark can be.

Schnelle's note contains two different types of arguments. First, he tries to show that other writers refer to the canonical gospels earlier than the lifetime of Marcion:

The common claim that before Marcion (in Rome ca. 140; in his home in Sinope ca. 120 CE) no evidence exists for a Gospel (e.g. Vinzent, Die Auferstehung Christi, 119-20) is not convincing. The Didache (ca. 120) presupposes the presence of the Gospel of Matthew and (indirectly) the Gospel of Mark (cf. Did. 15.3//Matt. 18:15; Did. 8.2//Matt.6:7-13; Did. 9.5//Matt. 7:6a; Did. 7.1//Matt. 28:19; Did. 8.1-2//Matt. 6:2, 5; Did. 11.7//Matt. 12:31-32; Did. 16.1//Matt. 24:42; 25:1-13'. For the detailed argument, cf. Wengst, Didache, 24-32. The Papias fragment (ca. 130) in Eusebius (Hist. eccl. 3.34.15-16) attests Mark and Matthew as writers of Gospels. Justin, the contemporary of Marcion, cites from the written text of Matthew (Matt. 11;17) and presupposes readings from the Gospels each Sunday (1 Apol. 67.3). Moreover, the Lord's Supper tradition in 1 Apol. 66.3 refers explicitly to the tradition of the Gospels (cf. Luke 22:19). The Gospel of John is attested in P52 (middle of the second half of the second century CE) in Egypt and must have been written a considerable time before that. For the Gospel of Luke, Marcion is the first (indirect) witness. However this fact certainly does not preclude its being written considerably earlier.

I have to say that I don't find any of this particularly convincing.

Re Papias, modern scholars seem to reject his attributions as attempts to lend authority to the canonical gospels, which would make him not a very reliable source -- and his motivation for these attributions would also be a motivation for him to represent the canonical gospels as being as early as possible.

The argument based on the Didache, if valid, would only limit Mark to 120 CE. And in any case Vinzent attacks this dating of the Didache as unreliable. (I'm not sufficiently knowledgeable to judge whether Vinzent is right here.)

The final couple of sentences provide a whole different line of argument.

The classic dating of the Gospels (and Acts) between 70 and 100 CE is evident in the situation presupposed in them and the history of early Christianity. Without the Jesus traditions of the Gospels, the expansion of Christianity in this period is not conceivable.

This seems extremely unpersuasive to me. There is no reason that Christianity can't have spread through some combination of other oral and written means, for many decades before the final forms were set for any of the texts we now have access to or can reconstruct.

So in answer to my own original question, my impression (based on the limited information I've been able to turn up) is that there is no firm reason at all why Mark can't be as late as 110 CE, or maybe even later. There are just indirect and circumstantial arguments, such as Σίμωνα τὸν Καναναῖον and speculations about the dynamics of the spread of early Christianity. I'm sure that people like Brandon and Schnelle have a good feel for the period, and are satisfied that their picture of the chronology is pretty close to the truth because, for them, it makes everything hang together. However, there doesn't seem to be much evidence of any more definite character.

After making a timeline of Jesus and early Christianity, it seems to me very difficult to date Mark gospel later than during the 80' decade. I'm not expert in the topic, but Mark's gospel fits well with early dating just after the Roman Jewish War, although as said, the window is open up to the early II century. I must say making such a timeline was difficut, many datings are not that clear and I'm sure this timeline contains errors.

The question needs a couple of frame challenges. Dating solutions for the gospel of Mark do range from ~40s–140s.

It is correct to say that the most common/popular/conventional dating would put Mark around 70, because a vaticinium ex eventu argument that says 'predictions are hard if they are about the future'. The thinking is: if the to-be-destroyed temple is mentioned, and was destroyed in reality, then this prediction is more likely not a prediction, but hindsight addition, and therefore puts that gospel as after the temple destruction.

This sounds convincing. The problem is of course that we do not have a complete Markan gospel from much later (even the oldest tiny fragment {\mathfrak {P}}137 was once announced as a 'First Century Mark' but is now dated as late as "later 2nd or earlier 3rd") and then dating the entire gospel on this one event/line of text conveniently overlooks that we are sure that the text was written and rewritten several times. The ending is even available in more than one version with 'the longer one' only evidenced by the second half of the second century (post 155). Therefore, also this one passage about a destroyed temple could have been inserted after the destruction, while most if not all of the rest might be in fact much older. One extreme position still held but not widely supported is 'very early dating', just after the crucifixion/resurrection (Easter), and the temple episode even being a real prediction. But this fixed point in time alone is just no sound argument for dating the entire text conclusively as one unaltered unit to such a seemingly precise date.

There is no clean solution to the synoptic problem, since all theories proposed have distinct advantages and disadvantages, none of those solving all problems and most leading to further questions.

We therefore still have to contend that either parts of the text, or 'source material', might have been written relatively early, that even a longer proto-Mark might have existed quite early on, but that at least the 'final' canonised version is perhaps of a quite late date.

This 'late dating' faction does propose a date for Mark that either approaches 110 or even later. This is a minority position even in European theology, with apparently most American scholars rejecting it wholesale, but growing from its initial presentation by Semler and then Harnack and now being developed into a full reconstruction of 'the oldest gospel' by Klinghardt since the early 2000s:

The core of that hypothesis is that the synoptic problem is supposedly 'solved best' not with Markan priority like in the two source theory, but with Markionite priority!

That would mean Marcion was only born in 85, either found, collected or wrote his proto-gospel, and presented this together with his collection of 10 letters of Paul as 'the canon' to the church in Rome, which then responded, or rather had to respond in writing more gospels.

The more traditional interpretation is that 'Marcion rewrote Luke' to make the text fit his 'arch-heretic views', but that presents more problems than it solves. The proposal that Luke would be a rewrite of Marcion on the other hand avoids many of those problems while having relatively few weaknesses on its own. That is: If it would not run counter to many beliefs in scholarship and force a very late dating of all canonised gospels!

First, the general picture confirms the critical arguments brought forward from both sides against their respective counterparts. On the one hand, "Q" is, indeed, "dispensable." The inclusion of Men avoids the methodological weakness of the 2DH with regard to the minor agreements and the hypothetical character of "Q": Compared to "Q", Men is clearly less "hypothetical", even though its text must be critically reconstructed from the sources and even though its place within the maze of the synoptic problem requires careful assessment. On the other hand, the basic observations that led to the hypothesis of "Q" in the first place, i.e. the bidirectional influence within the double tradition, are equally confirmed. The postulate of a single dependence of Luke on Matthew (or of Matthew on Luke) oversimplifies the complexities of the inter-synoptic relations. But it is neither possible nor necessary to establish such a single dependence. Instead, the inclusion of the "proto-Lukan" gospel which was used by Marcion easily explains the ambiguity of the material. Particularly with respect to the 2DH the burden of proof has shifted to those who suggest the existence of "Q" in order to explain the synoptic relations.

Second: What seems to make this picture complicated at first glance, indicates a major shift in methodology when compared to 19th century source-criticism. Although the 2DH tried to overcome the blockades of single dependencies, it is still basically oriented towards the simple usage of sources: it only augmented the number of relevant sources. Although the inclusion of Men is a similar augmentation of "sources", the evolving picture is different: whereas the 2DH tried to explain the complexity of the data by the addition of two basic sources (Mark + Q), the inclusion of Men demonstrates that both Matthew and Luke received their triple tradition material via two different routes: Matthew read Mark directly and in its revised edition in Men, and Luke used Men both directly and in Matthew's revised and enlarged edition. Since Luke, as it was demonstrated, also did know and use Mark, Mark was present in all stages of the synoptic tradition. The editorial procedure of both, Matthew and Luke, was not a mere addition of "sources" but a comparison of texts and concepts. This is fully consonant with the insight of the redaction history that the evangelists were ambitious and competent authors rather than mere editors. The mutual inter-dependencies create the complex maze of the synoptic tradition which, as a result, must be regarded as a much denser process than the 2DH suggested.

Finally, it is clear that this paper only intends to open the window for further discussion: I am fully aware that I am far from seeing all the implications and consequences of this suggestion, neither within the realm of the traditional issues of the synoptic problem nor the historical consequences that lie beyond it. But since this model provides a solution of the contentious issues of the present debate, it may help to break the deadlock in which the discussion of the synoptic problem seems to be caught for too long now.

— Matthias Klinghardt: "The Marcionite Gospel and the Synoptic Problem: A New Suggestion", Novum Testamentum , 2008, Vol. 50, Fasc. 1 (2008), pp. 1–27. jstor

This is then fully developed in the reasoned theory and even reconstruction of the lost Markionite gospel text:

Also for the dating of the preceding steps of tradition, i.e. the origin of the pre-canonical gospels (*Mk; *Mt; *Joh), there are hardly any clues. The oldest gospel looks back with certainty to the destruction of Jerusalem in 70 AD. Even if the lk editorship has brought this event even more into focus and interpreted it theologically, statements such as *21,5f.20 can probably only be understood in such a way that the destruction of Jerusalem already dates back. Although the destruction of Jerusalem is the most prominent historical event in the history of Judea in the second half of the 1st century and has therefore been used again and again for the dating of the Gospels, it does not provide the only clue. For the story of the possessed Gerasenes in Mk 5 I have shown elsewhere that the combination of the Decapolis city Gerasa with the many-headed demon called "Legion" as well as the drowning of the swine in the Sea of Galilee presupposes conditions that are not even conceivable before the end of the 80s of the 1st century: [...]

For Gerasa a Roman presence on a larger scale is epigraphically secured only in the context of the city expansion from the end of the 80s. Since *8,26-37 with the idiosyncratic localization of the event in Gerasa and the presence of "legion-demons" contains the same elements as Mk 5, the same conclusions are to be drawn for the dating of Mcn: Before the end of the 80s the writing of Mcn is more than unlikely. […]

However, these dating proposals represent only a majority opinion: The actual range of the seriously discussed datings is much broader and covers about 100 years from the 40s of the 1st century to the mid-140s.48 The widespread majority opinion with the datings between 70 and 90 AD is not only not very sustainable (which most authors do acknowledge), but also not very consistent, as can be seen especially in the determination of the terminus ante quem: Here, as a rule, it is not really evident how a dating even before the turn of the 2nd century can be justified. For the internal criteria that are usually given for this do not really allow these restrictions. Given these uncertainties, it is not advisable to add more to these rough estimates: There is simply not enough information for that.

However, two methodological notes are indicated. First, it has been shown that the canonical Lk only came into being in the context of the Canonical Edition. If this edition is to be seen in the context of Marcion's separation from the Roman community and at the same time Justin can be regarded as the earliest witness of this edition, one arrives at the decade between 144 and 155 A.D. for the origin of Lk.50 The discrepancy between this and the traditional dating should warn against the attempt to date the pre-canonical Gospels too precisely solely on the basis of internal criteria.

A second point of view arises from the great uniformity of the Gospel tradition: the close literary recourse of the individual stages of transmission to all the respective pre-texts indicates that the process of the formation of the pre-canonical tradition between Mcn and *Joh need by no means have extended over the entire period between 90 and about 144 CE: Months or a few years rather than decades may lie between the various pre-canonical transmission stages.

If one takes both considerations together, then a period suggests itself for the writing of the pre-canonical gospels, which must have lasted not longer than few years. Various scenarios are conceivable here. A continuous process is possible, which - in analogy to the widespread assumptions for the origin of the canonical gospels - calculates with longer intervals between the individual stages of transmission and fills the entire possible period. With an even distribution one would come for Mcn to the time from 90, for *Mk to the first decade of the 2nd century, for *Mt to the mid-120s and and for *Joh to the time shortly before 144. It would also be conceivable that the entire written gospel tradition would have originated in only a few years at the very beginning of the period (between 90 and 100) or even at its very end (130 to 144). And finally, it is also possible that Mcn existed as the only Gospel for a longer period of time from about the 90s, but that the pre-canonical updating from *Mk to *Joh took only a few years; such an updating could be imagined up to the turn of the century, but it could just as well have taken place only in the 130s. These considerations show on the one hand the great danger of arbitrary determinations, but on the other hand they make clear the scope for further research.

— Matthias Klinghardt: "Das älteste Evangelium und die Entstehung der kanonischen Evangelien", Texte und Arbeiten zum neutestamentlichen Zeitalter 60, Franke: Tübingen, 2015. gBooks, above: my translation

In that model, the relationships between the different gospels and their traditions look like these two schematics:

(first graphic from Klinghardt's German Wikipedia page, second from the English WP: Priority of the Gospel of Marcion)

This idea was of course quite heavily criticised (examples on the respective Wikipedia pages or especially here), not only for its consequences, but in some detail, while nevertheless prising some of its aspects. One critic who seems to have converted to this idea would be Vinzent:

Markus Vinzent is the head of the research project at the Max Weber Centre for Advanced Cultural and Social Studies on the place of Marcion of Sinope in the development of the gospel genre. The project, entitled ‘The Gospel of Marcion: The Beginnings of Christianity’, proceeds from the premise that the four canonical gospels of the New Testament were composed in the first-century and were canonised following the end of the second-century AD.

The project brings into question the scholarly consensus surrounding the dating and redaction history of the canonical gospels, to show that the teacher and naval merchant, Marcion, was the first to compile a work of literature recognisable today as belonging to the gospel tradition. The predominant language of Marcion’s home region, Sinope, a Bithynian-Pontian Roman province, is widely accepted to be Pontiac Greek, a dialect of Ancient Greek, although it remains yet unknown, whether or not Marcion was fluent in the language.

Vinzent (2014) argues, that the core writings of the New Testament were likely heavily redacted, if not outright created, in the second-century AD, namely the canonical gospels (as well as the non-canonical gospels), which stem from the middle of the second-century AD, the Pauline Epistles (going back to Paul, but severely redacted when further ‘Pauline’ letters and Acts were added).

Even further redactions followed in the second-century, when the texts were gathered to form the collection of the New Testament. The Roman teacher and merchant, Marcion of Sinope, is regarded by Christianity as one of its key thinkers and founders, and is associated with the first compilation of a gospel and 10 Pauline letters. Vinzent’s views stand apart from the conventional understanding of scholarship. The early Church first developed itself out of a ‘Jewish sect’, but it failed to shed its Jewish trappings until Marcion reinterpreted it in the years following 140 AD. For Vinzent, it was Marcion, who through the corpus of the ten Pauline Epistles and his one Gospel, helped to spread the belief in the Resurrection of Christ throughout Christendom.

Thus Marcion’s Gospel is taken to be the historical font for each of the four canonical gospels, because they consult Marcion as their source. In his own words, Marcion created the literary genre of the gospel tradition and gave his work its name, all whilst without historical precedent in the attachment of the name of this genre to the story of Jesus.

Thus, one can date Mark "as late as 110", and even later, but it's simply not the most popular position.