The place and significance of Marcion's Gospel in the formation of New Testament literature has been obscured by the persistence of what I call the Patristic Hypothesis about its origin, namely, the idea first put forward by Irenaeus and Tertullian that Marcion created it by means of ideologically motivated editorial subtractions from Luke. Their assertion is merely a hypothesis, since they were in no position to know what form Luke took prior to the time of Marcion, or in which direction the redactional relationship went.Footnote ¹⁹ This hypothesis, of course, has its roots in polemical assumptions, and in fact fits very poorly the actual data of Marcion's Gospel in comparison to Luke. Once we break with such assumptions and objectively examine the texts of the two gospels, it becomes immediately clear that Marcion's Gospel cannot be an ideologically motivated redaction of Luke, for the simple reason that the two gospels are practically identical in ideology. For every position of Marcion cited to explain an omission in the text of his gospel, the latter contains any number of passages sharing the idea of the omitted passage, and in tension with Marcion's own position. Close analysis of the content of Marcion's Gospel, therefore, makes it evident that this gospel, in the words of Judith Lieu, ‘is in many ways neutral: It can only have served to inspire and support [Marcion's] system to the extent that he interpreted it’.Footnote ²⁰ Clearing away the obscuring influence of the Patristic Hypothesis, and treating the text of Marcion's Gospel seriously as a precious early witness to the process of gospel formation, leads to a number of implications both for gospel history and for New Testament history more broadly.

We have all been guided by Occam's Razor and the desire to find a clean, neat, simple, uni-directional model of gospel relationships that explains all the evidence. We need to accept that such an ideal is unattainable. There are two principal reasons for this. First, we have no autographs of these texts, so we are always dealing with manuscripts that reflect various degrees of modification and exposure to other gospel texts. Our difficulty in identifying which elements belong to which layer of composition and later development is a major obstacle to establishing the original textual dependencies of gospels as originally composed by their original authors. Second, these texts underwent an ongoing fluidity of text that defies familiar understanding of what constitutes authorship and composition, on the one hand, and what constitutes emendation and corruption, on the other, due to the sub-literary character of gospels as cultic texts. They have been mishandled in scholarship when read as works of high literature, comparable to the treatises of Cicero or poems of Virgil, the product of an authorial act with an original text that can be clearly distinguished from later textual ‘corruption’ or distinct redactions at the hands of specific editor-authors. By contrast, the intertextual exposure and modification we see in gospel texts flows seamlessly from the kind of textual dependencies involved in their original formation as cultic instruments, for which authorship or a singular event of composition is largely irrelevant.Footnote ²¹

A concrete example of how Marcion's Gospel illuminates our understanding of these processes of gospel formation can be found in the so-called ‘Minor Agreements’ between Matthew and Luke against Mark in the clean, neat, simple, uni-directional Two-Source Hypothesis of Synoptic relationships. This phenomenon has caused a great deal of hand-wringing, and has led a significant number of scholars, including Matthias Klinghardt, to conclude that the Two-Source Hypothesis is wrong.Footnote ²² The seriousness of the problem depends upon whether the ‘Minor Agreements’ are an element of composition that existed in the original autograph of Luke (for some reason, it is always Luke, not Matthew), or were introduced subsequently as a textual corruption. The evidence of Marcion's Gospel aligns with the latter idea, that they were introduced in the process of transmission of the gospel text, since Marcion's Gospel contains between a half and two thirds fewer ‘Minor Agreements’ with Matthew than the current critical text of Luke does (a critical text that, due to axioms of text criticism, gives an absolute minimum of ‘Minor Agreements’ in Luke). In other words, the phenomenon of ‘Minor Agreements’ is reduced in Marcion's Gospel to such a small factor that one must doubt that it was a feature of the original text at all, and conclude that Luke has more of them due to the greater exposure to the text of Matthew in the process of its transmission – either from a longer period of exposure or from transmission in closer association with Matthew, or both. In the case of Marcion's Gospel, of course, exposure to the text of Matthew must have occurred before the gospel text reached Marcion and was sequestered within the Marcionite community, at which time exposure to Matthew in its transmission would have ceased. Nonetheless, two centuries of critical scholarship had to contend with the ‘Minor Agreements’ as if they were compositional elements that needed to be solved in the construction of our models of gospel interrelationships. Only now with the evidence of Marcion's Gospel can this whole problem be set aside.

It is the unique conditions of control afforded by the three Synoptics and by the fortuitous partial survival of a fourth Synoptic, Marcion's Gospel, that allow us to distinguish the stages at which certain developments of gospel texts occurred; but in countless other details of the gospel texts, where we do not have such controls, it is impossible for us to make similar distinctions. It is for this reason that we cannot insist on perfectly clean, neat, simple uni-directional models of gospel relationships with all elements accounted for and no flies in the ointment. We cannot insist on this because our manuscripts come too late in the transmission process to escape intertextual exposures and other changes that have altered the texts from their originally composed form.

Despite these challenging conditions of the materials we have to work with, neater, simpler, less-multi-directional models of gospel relationships are still to be preferred, as requiring less special pleading in their defence. Marcion's Gospel, as the Fourth Synoptic, adds a control that allows us to assess such models of gospel relationship. Matthias Klinghardt argues that canonical Luke derives from Marcion's Gospel by a process of additions to the text.Footnote ²³ His arguments are, on the whole, cogent and persuasive. But that does not necessarily mean that Luke is a post-Marcion, anti-Marcionite redaction. If Marcion's Gospel predates Marcion, so too might the redactional relationship between it and Luke. The signs of an anti-Marcionite purpose that Klinghardt and others point to are far too subtle. There is a fundamental continuity in ideology and ethos between Marcion's Gospel and Luke.Footnote ²⁴ If we were to think in terms of authorship and distinct redactions, it could even be suggested that Luke is a second edition of Marcion's Gospel by the same author. Be that as it may, there are few grounds for proposing ideologically distinct communities as the venue of use for these two gospels. Since there is no clear ideological tendency that distinguishes one from the other, I would suggest a pragmatic or cultural purpose behind the differences between the two texts, that is, culturally rather than ideologically distinct communities. Not every variation in early Christian life and literature was ideology-driven. Marcion's Gospel, which is relatively less engaged with the Jewish tradition, was suitable for use in Gentile-dominated communities, while Luke, relatively more engaged with it, could have been intended for use in communities with a stronger Jewish background.

The agreement between Klinghardt and myself that Marcion's Gospel is the earlier version, pre-Marcion in its composition, and not a tendentious derivative of Luke, leads to the implication that it is a closer witness to the textual dependencies of the Synoptic Gospels, and that is what I mean in calling it the Fourth Synoptic. As such, it should be included, and even given priority over Luke, in explorations of the Synoptic relationship. When we do that, my initial assessment differs from the conclusions of Klinghardt, who finds reason in the comparison of Marcion's Gospel with Luke to reject the Two Source Hypothesis and the role of the hypothetical text ‘Q’. In my judgement, however, the evidence of Marcion's Gospel strengthens the case for the general accuracy of Two Source Hypothesis of the Synoptic relationships, once we allow for the greater fluidity of text I described previously. I have already mentioned the disposal of the problem of the ‘Minor Agreements’, removing a major stumbling block to the hypothesis. A second problem with the hypothesis has been the reconstruction of ‘Q’. This strange hypothetical text, as currently reconstructed, starts out as a narrative, with Jesus baptised by John and enduring the Temptation, but then turns into a sayings source resembling Thomas. But if Marcion's Gospel is substituted for Luke in the reconstruction of ‘Q’, the problem disappears. No baptism, no Temptation. ‘Q’ emerges as a pure sayings source. With the evidence of Marcion's Gospel dispelling two of the major arguments against the Two Source Hypothesis, the latter is affirmed as fundamentally sound. The problems for the hypothesis created by the text of Luke on which it has been based stem from the fact that Luke is a relatively late redaction of the gospel that has been deeply impacted by intertextual exposures to the other gospels – John as well as the other Synoptics. Marcion's Gospel, therefore, solves problems in the Synoptic relationships that have been insoluble on the evidence of the canonical gospels alone.

Marcion evidently viewed the gospel (which he knew simply as ‘the Gospel’) as authoritative, and incorporated it along with ten letters of Paul into a corpus that for the first time elevated Christian writings to the level of scripture on a par with those used in Jewish synagogues. In this way he anticipated the later formation of the New Testament canon, and it is natural to raise the question of the degree to which his move was a catalyst for the latter. Irenaeus two generations later and Tertullian three generations later attest the progressing development of ideas of canon among the non-Marcionite Christians. Even for these authors, there is nothing like a full New Testament, with a fixed order of texts; yet Marcion had already achieved this level of ‘canonisation’.Footnote ²⁵ Scholars no doubt will perennially debate how inevitable the development of a New Testament was within Christianity, and therefore how necessary Marcion was to its impetus. Nevertheless, it cannot be doubted that Marcion anticipates not just the idea of a New Testament, but even the peculiar selection of texts involved – namely, gospels and epistles, and in particular Pauline epistles. This combination of genres poses real problems for those who want to see an organic continuity between the Old Testament and the New Testament as an evolving scriptural corpus that did not require Marcion as a catalyst. Nothing in the Old Testament resembles a gospel or an epistle. Early Christians did compose texts that look more like Old Testament texts, and in fact a large quantity of Jewish texts underwent Christian redaction and circulated widely. But the kind of texts Marcion chose for his Christian scriptures were unrelated to Old Testament literary forms. So it is remarkable that non-Marcionite Christians embraced these choices in gradually demarcating ‘scriptural’ from ‘non-scriptural’ texts. Even more remarkable is the dominance of Paul, which is totally unexpected against the background of the relative neglect of him in the proto-orthodox writers of the second century.

What then explains this peculiar selection that first Marcion and then his opponents considered Christian scripture? I have proposed that gospels and epistles closely correspond to the sort of texts one would find in the capsa, the chest of charter documents kept by Greco-Roman cultic associations.Footnote ²⁶ Epistolary correspondence played a major role in the chartering and operation of such associations, along with either a mythic or historical foundation-narrative for the cult. Both kinds of documents were read out in association meetings. What I am proposing, therefore, is that the Hellenised if not Hellene Marcion, rather than thinking in foreign and anachronistic terms of ‘scripture’, was working within this understanding of charter documents when he undertook a standardisation, or ‘canonisation’, of texts for his network of Christian associations. We need to cut through anachronism to recover this environment of sub-literary charter documents to get the nuance of terms quoted (albeit in Latin translation) from Marcion, such as instrumentum for his collection of gospel and apostle, which is precisely the Latin term for such charter documents; and concorporatio, which perhaps goes back to Greek sussōmatizō (cf. sōmatizo, sōmatismos), to combine or merge into a (single) legal instrument.

It is only when we reset gospels and epistles in this category of charter documents that we can understand the very different expectations about texts that shaped the composition and modification of this material, and can properly situate Marcion's textual and organisational actions. He is quoted by his opponents as reacting to the introduction of additional texts that he considered incompatible with the original charter documents.Footnote ²⁷ This innovation took the form of ‘reading from scripture’, which appears to have been spreading in Marcion's own time, and for which there is little evidence before that time.Footnote ²⁸ This explains his move to ‘canon’, in order to set limits to the Christian instrumentum and to undo the concorporatio of an expanded one that included Jewish scripture, by delimiting and closing the set of accepted charter documents. His opponents, meanwhile, allowed the concept of ‘scripture’ carried by the Jewish Tanakh to gradually but substantially alter their relationship to text, and their view of gospels and epistles, to the one that is assumed today both in the practice of Christianity and in the modern academic study of the Bible.