Christians since antiquity have grounded their faith on its authentic attestation in the gospel of Jesus Christ received from his first apostles. This grounding is already explicit in the Bible itself and has remained an uncontroversial aspect of historic Christian praxis and worship since antiquity.
Throughout their history, churches of virtually every stripe have—for all their tacit or fiercely contested differences—shared a core conviction about Jesus of Nazareth as in some sense both a human being in history and yet also “God with us.” Jesus has always been encountered and experienced in a variety of ways. Most prominent since antiquity have been practices of prayer and common worship that include a liturgical meal celebrating both his memory and his presence, accompanied by the public reading of the four gospels—authoritative writings about his teachings and ministry received in the names of his earliest disciples.
But the early Christian use of gospels also has a fascinating dynamic of its own, operating in theologically powerful and yet surprisingly polyvalent ways in diverse periods and communities.
The term “gospel” surfaces in the earliest tradition as characterizing Jesus’ message. Matthew and Mark both present “the gospel” (to euangelion) as the radical message and praxis of Jesus about the imminent coming of God’s kingdom (see esp. Mark 1:14–15; 8:35; 10:29; 13:10; Matt. 4:23; 9:35; 24:14). Luke, who is more aware of the public, imperial context of his writing, does not seem to like this noun, for reasons that will become apparent in a moment. He never uses it in his gospel, and in Acts it appears only once each on the lips of Peter and of Paul (Acts 15:7; 20:24). The verb “to announce good news” (euangelizomai), on the other hand, occurs frequently in both Luke and Acts.
Even Matthew and Mark, however, already show a transition in meaning that evidently occurred at a very early stage in the tradition—it is in fact already complete in the Letters of Paul, which predate all four New Testament gospels. Whereas “the gospel” in Matthew and Mark almost invariably reports what Jesus himself preaches and enacts, even here there are signs that by the time of these evangelists “the gospel” has become the content of the message he entrusts to his disciples, and indeed the message about him. So Matthew’s Jesus himself can promise that “this gospel of the kingdom” will be proclaimed throughout the world after his death (Matt. 24:14; 26:13). And Mark 1:1 opens with the words, “The beginning of the gospel of Jesus Christ”—a famously ambiguous phrase that leaves unresolved whether the gospel here in view is Jesus’ message (as in 1:14), the message about Jesus (e.g., 13:10; 14:9), or perhaps even—by a kind of metonymy—Mark’s own book that sets forth this message. But it clearly involves the person of Jesus, including his message and ministry as well as his death.
Additionally, and well before Mark writes his account, it is already clear that when in the early 50s Paul preached to the Corinthians the gospel by which they are saved, this entailed at a minimum a narrative passion and resurrection sequence involving “Christ died for our sins, … he was buried, … he was raised on the third day, … he appeared to Cephas, then to the twelve, then” to many others in succession (1 Cor. 15:1–6; cf. 2 Tim. 1:10; 2:8). There seems moreover to be continuity here with the similarly sequential narrative, quoted a few chapters earlier, of words and actions of Jesus “on the night when he was betrayed” (1 Cor. 11:23–25).
A few decades later, in a more retrospective account of Peter’s first preaching to the Gentiles during the mid-30s, the narrative of Acts has Peter assuring his audience at the house of Cornelius about “the word” God sent to the children of Israel, “proclaiming the good news [euangelizomenos]” of peace through Jesus Christ (Acts 10:36, my translation). That “word” (logos), he goes on to say, came to expression through the “message” (rhēma) associated with certain particular events that recently transpired in Jewish Palestine,
beginning in Galilee after the baptism that John announced: how God anointed Jesus of Nazareth with the Holy Spirit and with power; how he went about doing good and healing all who were oppressed by the devil, for God was with him. We are witnesses to all that he did both in Judea and in Jerusalem. They put him to death by hanging him on a tree; but God raised him on the third day and allowed him to appear, not to all the people but to us who were chosen by God as witnesses, and who ate and drank with him after he rose from the dead. (Acts 10:37–41, NRSV)
In other words, even the earliest stages of the tradition, both as attested in Paul and as attributed to the remembered Peter in Acts, envisaged the gospel to include a narrative about Jesus’ public ministry and message, culminating in his death and resurrection. (Significantly, Luke places a Mark-like apostolic gospel outline on Peter’s lips. This is despite its obvious divergences from the structure of Luke’s own gospel account with its addition of birth, infancy, and ascension stories.)
Readers familiar with the gospels and with cognate English words like “evangelical” are sometimes surprised to discover the extent of scholarly debate and controversy about the origin and precise meaning of the early Christian use of the term euangelion. One school of thought has long stressed the conviction that the term must be understood as originating in connection with the Hellenistic use of euangelia (Greek plural) to denote “happy news” or “good news”—as used in the eastern empire most publicly in relation to official Roman imperial announcements about good news like the accession, birthday, or victory in battle of the emperor as “Savior” (sōtēr, a word the New Testament uses much more sparingly than later Christian tradition). The most famous pre-Christian example is an inscription in praise of the birthday of Caesar Augustus that was erected at Priene and other cities in Asia Minor in 9 BCE. He is celebrated as “our God” whose birth “signified the beginning of happy news [euangelia] for the entire world.” Even without using the word “gospel,” the Roman poet Virgil’s famous Fourth Eclogue, composed around 42 BCE, deploys Isaiah-like imagery in anticipation of an age of eschatological peace and salvation associated with the birth of an unnamed child (though not perhaps identifiable as the hoped-for son of Mark Antony and his wife Octavia, as scholars used to think).
The notion of public good news had been common currency for many centuries, being attested ever since Homer (Odyssey 14.152, 166: euangelion, singular). Indeed the commonplace inflation of such terminology could even become the butt of jokes: the Athenian comic playwright Aristophanes (ca. 446–386 BCE) already had a sausage seller poking fun at bawdy market hyperbole by intoning, “Hey, Senators, I’m the first with tremendous news [euangelisasthai]: never since the war began have sardines been so cheap” (Knights 642–45; trans. Roche 2005). The familiarity of such terminology can be gauged too by its adoption as a Latin loan word: the Roman writer Cicero repeatedly and somewhat informally does this, as when writing to his friend Atticus in 60 BCE, “First, I have what I think is good news [euangelia] …” (Letters to Atticus 2.3.1).
One might think, therefore, that Christian talk of to euangelion, the good news, basically just recycled for Jesus a well-known cliché that could evoke little more than a yawning response. That would hardly convey the sort of grandly anti-imperial ambition which the claim of a Christian euangelion is sometimes said to advance. To be sure, resistance to the force of empire soon became at least a sporadic occurrence—and sometimes part of the very essence of what it meant to be a Christian, as stories about the trials of martyrs repeatedly affirm. But despite sometimes heated scholarly debate, it remains difficult to document in the New Testament any sense that the use of the term “gospel” serves a clear anti-imperial function.
A related line of argument has sometimes taken such early Christian terminology to imply the church’s origin not as a Palestinian Jewish messianic movement but as a Hellenistic divinized hero cult, drawing on culturally commonplace idioms and assumptions about heroes or rulers.
But to acknowledge the existence of such potential Hellenistic resonance is not yet to understand what a (or the) gospel conveys in the early Christian texts. Even for Greek-speaking Jews and Christians, gospel language must have carried a kind of dual significance. On one hand, there will have been at least an awareness of the secular use of “good news,” sometimes exploited in the service of ideological ends and propaganda. Jewish writers in Greek like Philo and Josephus repeatedly illustrate the currency of such a meaning of “good news.” Secular as well as religious overtones were indeed in the air, even for Jews.
On the other hand, however, we must recognize that the Greek terminology was also already part of a richly textured discourse of prophetic and divine communication in older, pre-Christian Jewish Greek Scriptures. In that respect the Greek words conveyed a Jewish, Old Testament meaning—often associated with the second part of the book of Isaiah, which announces the Servant of the Lord’s return to redeem Jerusalem (52:7) and speaks of “good news” to the afflicted and imprisoned (61:1, both times using the verb euangelisasthai). While the Greek Old Testament does not deploy the noun “gospel” in this fashion in either the singular or the plural, the formative role of widely influential texts like these in the early Christian understanding of the gospel of Jesus is clear. Other Jewish texts in Greek like Psalms of Solomon 11:1 clearly highlight such usage, and Paul quite confidently appropriates Isaiah 52:7 in speaking of the activity of the apostles as proclaimers of a message that is “the gospel” (see Rom. 10:15–16; cf. 1 Cor. 9:14; also Stanton 2013, 281–92 and passim and Horbury 2005, 2006).
Unlike the Greco-Roman use almost exclusively of the plural euangelia, the early Christian writers deploy the singular “gospel” (euangelion) consistently and uniquely in relation to the message of or about Jesus. That said, even here there is some evidence of semantic ambiguity from the start. As we saw earlier, Jesus’ message soon became the message about him (Mark 1:1; 14:9; and 16:15; note esp. Matt. 26:13; 24:14, “this gospel,” i.e., not only Jesus’ words and actions but evidently an account of that message and ministry—such as Matthew himself provides; cf. Stanton 2013, 95–98). Already in the corpus of Pauline Letters the term came to be used interchangeably for either the message or its content: the apostle speaks of both “the gospel” and “my gospel” (cf. Phlm. 13 with 2 Tim. 2:8 and Rom. 16:25).
As already noted, Luke never uses the noun “gospel” in his narrative of Jesus (but see Acts 15:7; 20:24), although he does deploy the cognate verb twenty-five times in Luke and Acts. In Acts 13:32–33 he places on Paul’s lips a definition of what it means to preach the gospel: “we bring you the good news that what God promised to our ancestors he has fulfilled for us, their children, by raising Jesus.” The New Testament’s Johannine writings avoid the Greek euangel- word group altogether except at Revelation 10:7 as well as at 14:6, where it denotes a message of judgment.
Matthew’s usage in particular evidently had a powerful influence on subsequent understanding of what the gospel might be. Very rapidly, its range of meaning expanded from Jesus’ kingdom message or (as in Paul, e.g., 1 Cor. 15) the message about Jesus’ death and resurrection to include accounts of his life, preaching, innocent death, and resurrection “for us.” As we saw earlier, an early narrative form of this is implied in Peter’s account in Acts 10:34–42, and in the writings of Ignatius (d. ca. 107) it is already evident that “the gospel” designates for him the crucifixion-resurrection message of Jesus (Smyrnaeans 7.2), quite possibly in its Matthean form (cf. Smyrnaeans 1.1 with Matt. 3:15; similarly cf. Didache 8.2 with Matt. 6:9–13; also 2 Clement 8.5, more loosely, with Luke 16:10–11; see further Hill 2006; Foster 2005).
Significantly, not later than the middle of the second century the notion of this gospel story “according to” one apostolic figure or another had become attached to gospel books— for example, in Justin, First Apology 66.3 (see Stanton 2013, 92–97). A little before this, Marcion had already identified his edition of Luke as “the gospel.” Similar examples can be found in other early documents: the form of the Didache’s reference to its source suggests that “the gospel” was already used to designate “a gospel writing, almost certainly Matthew, some decades before Marcion” (thus Stanton 2013, 77; cf. Kelhoffer 2014, 72).
If this is correct it follows, importantly, that known portions of one or more of the subsequently canonical gospels were known and cited as “the gospel” before any of the extant noncanonical gospels were composed. To some extent this is inevitably a judgment about a serendipitous state of affairs at this present time, which the discovery of new sources or compelling reassessments of existing ones might require us to revise. And absence of evidence is not evidence of absence. But in the meantime it matters for our assessment of recent and current claims that while specific literary identifications are sometimes difficult or textually ambiguous (e.g., Luke 16:10–11 in 2 Clement 8.5, cited above), no ancient author refers to any identifiable version of a noncanonical text like Thomas or Q as “the gospel.”
Further on this note, it has been repeatedly shown (e.g., Hengel 1984; Gathercole 2013) that while the titles of the existing New Testament gospels are clearly not from the pen of the original authors, they and the associated authorial attributions are nevertheless both stable and remarkably early, probably from the first half of the second century. Although in theory compatible with simplistic explanations in terms of wholesale deliberate “forgery,” as Ehrman (2013) prefers, such a date makes it difficult to rule out the possibility that these apostolic attributions are instead based in some fashion, whether correctly or in error, on an existing chain of collective or individual living memory.
Manuscript evidence suggests that the short forms “according to Matthew” or “according to Luke” are secondary abbreviations from an original longer form, “the gospel according to” Matthew or Luke; see, for example, Gathercole 2013. (Gathercole 2012b illustrates this same usage in the flyleaf of Matthew included with manuscript 4, dating from ca. 200. Thus Bovon’s assertion that “we have no codices [with inscriptio and subscriptio] of these gospels predating their canonization” [1988, 20–23] turns out to be an argument from increasingly partial silence, which will require fuller facts and rather more nuance. It is hardly the comprehensive refutation of the “extravagant claims of Martin Hengel” that Ehrman [2013, 53 and 53n55] imagines).
Hengel additionally observes that while the title “gospel” is routinely introduced in reference to other gospels like Thomas or Nag Hammadi’s Gospel of the Egyptians, it is never lost from a text that has once been so designated—even though gospel status itself seems to fade from interest for later compositions at Nag Hammadi, where “dialogues” and “revelations” predominate over narratives of the earthly Jesus. Among other things, this suggests that the relatively rapid successive publication of the Synoptic Gospels between the 60s and the 90s, designated within a few decades as “the gospel according to X,” may have established a compelling precedent for the choice of titles in later accounts of the teachings of Jesus. This precedent entailed both the term “gospel” and the name of an apostolic guarantor, as evidenced not only in John but also in several noncanonical gospels. (See Hengel 2008b, 110–11, 182–83.)
A related point concerns certain material aspects of conservation and innovation in gospel writing. As we will see, there appears from the start to be greater textual stability in the extant manuscripts of subsequently canonical gospels than in those of Thomas, Peter, and other apocryphal gospels. In relation to this it has been plausibly suggested that the more widespread copying, liturgical reading, and memorization would have had a stabilizing effect on the textual tradition, certainly allowing for the composition of new gospels (like Matthew or Luke) but largely eliminating the scope for successive textual recensions of the same text (Evans 2015, 36–37). While Evans’s related inferences about the longevity of New Testament autographs look a little problematic in their specificity, a manuscript lifespan of a century and a half was indeed a reasonable expectation (see, e.g., Houston 2014, 175, on Oxyrhynchus)—and might reinforce this stability for texts that circulated widely.
Except for scribal identifications in titles or colophons (i.e., concluding scribal comments), the term “gospel” itself is remarkably rare in the body of ancient gospel-like texts at Nag Hammadi or elsewhere. Leaving aside late works like the Gospel of Nicodemus (B 14.1) or the History of Joseph the Carpenter (1.2; 30.3), the small handful of examples from antiquity includes the Gospel of Mary (9; 18) and the Gospel of Truth (17.1–4; 18.11; 34.34–35) for the saving message about Jesus. Nag Hammadi’s Sophia of Jesus Christ, in a question about why “in the gospel” (evidently a text!) Sophia’s Son is called “human” and the “Son of Man” (104.1), also demonstrates this meaning.
Until the nineteenth century, Western biblical scholars tended to take for granted that the emergence of the early church was based on “one holy catholic and apostolic” faith and that the canon of Scripture was essentially the result of a continuous and intentionally advancing original movement from which others deviated. However challenged that movement may have been by detractors without and heretics within, on this view it proceeded organically from Christ to the apostles, to the fourfold apostolic gospel and the New Testament read in light of the apostolic rule of faith (regula fidei). This in turn became crystallized in agreed forms of worship and confession in the Trinitarian creeds of Nicaea and Chalcedon.
I am not myself averse to all aspects of this traditional picture. No doubt it may be said to oversimplify or distort. But this very excess also functions to some extent like a political cartoon, usefully capturing salient features precisely by its clarifying selectivity and exaggeration of a few defining attributes out of the mass of conflicting data.
At the same time, even mainstream accounts of Christian origins are today rightly more nuanced about the ecclesial diversity of the first two centuries. And even among those who (like myself) would wish to retain an account of creedal Christianity’s organic connection to the faith of the apostles, most accept the eloquent evidence for a rather more complex picture. In that sense the metaphor of the cartoon may usefully be balanced by that of a pointillist master painting, which is best appreciated from just the right amount of sympathetic distance rather than by overinterpreting its constituent points of detail.
Even a brief encounter with first- and second-century sources shows that the reception and circulation of early Christian writings about Jesus remained remarkably fluid and elusive during that period. This is true even for some of the canonical texts: it is, for example, difficult to know quite how many second- or even third-century Christians could have had regular access to written copies of Paul’s Letters, Acts, or indeed the Gospel of Mark: only a few small fragments survive from that period, all of them from Egypt (for environmental reasons, as explained below p. 10; see Hurtado 2013 for statistics).
We do know that the second century was extraordinarily generative and fertile in religious and literary terms; one widely (if perhaps somewhat credulously) cited calculation suggests that our surviving sources from that period represent approximately 15 percent of the known Christian literary output (so, e.g., Markschies 2002, 98; 2015, 21; he likes to refer to the second century as Christianity’s “laboratory”: Markschies 2003, 120; 2012g, 34 and elsewhere).
And yet it remains the case that by the mid-second century, gospel accounts in the names of Matthew, Mark, Luke, and John were increasingly emerging as the accepted fourfold narrative of the gospel of Jesus Christ. Before the century was out, this had become self-evident to someone like Irenaeus, closely familiar as he was with the practice of the churches both of Asia Minor and of Rome: in the face of multifarious sectarian alternatives, the catholic acceptance of the Four seemed to him as incontrovertible as the four winds of nature (Against Heresies 3.11.8). Writing a few decades later on the basis of both Alexandrian and Palestinian experience, Origen (ca. 185–254) famously quipped that “the Church has four gospels; heretics have many” (Homilies on Luke 1.2).
To be sure, few congregations even in urban settings will have owned copies of all four “canonical” gospels. And even in places that affirmed these four as authoritative, they were now often encountered collectively rather than discretely. The extant manuscript tradition and actual evidence of use suggest that Christians at this time may often have physically experienced these texts not so much as four complete individual books, but in more episodic fashion through excerpts, informal or formal harmonies (most influentially Tatian’s Diatessaron)—or indeed through one particular gospel (most often Matthew or John) understood in light of such a harmony.
Despite some early attempts to establish definitive lists of all New Testament books, including the gospels, this “canonizing” effort did not achieve an agreed final form until the later fourth century. (Famously this is articulated in the Thirty-Ninth Festal Letter of Athanasius in 367 CE: for the text, see Grosheide 1948; a partial translation is offered in Metzger 1997, 312–13 [and discussion on 210–12]; cf. Brakke 2010b, with a new translation of the surviving letter, 57–66.)
But does this mean, as some scholars continue to assert, that no consensus about authoritative gospels existed until fourth-century authoritarian decrees imposed their will upon the previously unlimited flow of early Christian tradition and literature?
A recent inventory of pre-300 Christian literary sources includes a little over thirty gospel texts (Hurtado 2006, 209–21, updated online as Hurtado 2013; cf. Lührmann and Schlarb 2000, 22). It is in the nature of the evidence that statistical statements on this subject are necessarily somewhat tenuous. The sample size is tiny, and a small textual fragment in any case cannot prove the existence of the entire text of which it is a part. With some notable exceptions (mainly from Derveni, Dura Europos, Herculaneum, Nessana, and Petra as well as Qumran: cf. Leach and Tait 2000, 239; Tov 2003, 100–103), papyrus evidence is largely restricted to Egypt, where atmospheric conditions particularly favored its survival—and where many of the known extracanonical texts originated and thrived. (The only surviving noncanonical gospel-like text from outside Egypt is perhaps a Greek fragment of the Diatessaron from Dura Europos on parchment, that is, processed animal skin rather than papyrus: Yale P.Dura 10, formerly Dura Parchment 24; see below, p. 127.)
Except for the gospels of Thomas (P.Oxy. 1, 654, 655) and Mary (P.Ryl. 463; P.Oxy. 3525), no noncanonical gospel before 300 CE is extant in more than one copy. (P.Oxy. 2949 and 4009 are both sometimes assigned to the Gospel of Peter, but this seems unlikely in one and possibly both cases, for reasons discussed below.) No other gospel-like texts approach the manuscript dissemination of Matthew or John, nor for that matter the persistent breadth of attestation in extant early Christian literature of any of the four gospels that became canonical.
The statistics of extant manuscripts from the first three centuries coincide with those from literary sources in documenting Matthew and John as the most popular gospels by far, with Luke a relatively distant third. Mark was almost never copied at all: out of just over thirty known gospel papyri predating the year 300, at most three contain Mark (45; P.Oxy. 5073; and possibly 88; see Hurtado 2013 [addenda], citing Barker 2009; also cf. Head 2012, 114–15, who points out that 45 appears to punctuate and mark up the text of Mark, though not of the other gospels, for public reading).
While commentaries or scholia on Matthew, John, and Luke emerge in the second century, the first commentary on Mark appears only in the seventh, half a millennium later. (See, e.g., Wucherpfennig 2002 and Hill 2004 on John; Löhr 2003 on Luke; Cahill 1998 on Mark; also Kok 2015 on second-century reception of Mark.) In the preface to his commentary on Matthew, Jerome (345–420) could already claim to have benefited from extensive expositions by Theophilus of Antioch (late second century), Hippolytus (ca. 170–236), Origen (ca. 185–254), and numerous subsequent commentators in Greek and Latin (Jerome, Commentary on Matthew, Preface 4; cf. Scheck 2008, 19–20).
If one stops to think about it, the position of Mark’s Gospel is perhaps the most surprising. After two centuries of New Testament scholarship’s preoccupation with the priority of Mark, this gospel’s virtual absence from the earliest manuscript tradition rightly strikes us as peculiar. What might explain this? At a time when manuscripts were beyond the reach of most private citizens, and even most churches could not afford the luxury of a complete four-gospel codex (for whose production expensive parchment rather than papyrus turned out to be more viable), perhaps there was less need for Mark: 90 percent of it does appear in Matthew, who provides a more satisfying introduction and conclusion for a biographical narrative.
That said, the intensive use of Mark by Matthew and Luke is itself eloquent tribute to this earlier gospel’s importance for the Jesus tradition in the late first century. Moreover, several different second-century endings of Mark imply that somebody was engaging specifically with this gospel in the second century, and in the light of other gospel narratives. Distinctively Markan features of Tatian’s gospel harmony (the Diatessaron) are difficult to substantiate with confidence and may not require that Tatian had at his disposal a specifically Markan manuscript, although the evidence does suggest knowledge of Mark 6:5; 10:18; and elements of the longer ending, 16:9–20 (see, e.g., Head 1992b, 130, 137). Further discrete second-century evidence for the use of Mark as one of the authoritative four is suggested by the mid-century Epistle of the Apostles (see below, p. 220 and cf. Kelhoffer 2000, 155).
The use of Mark along with the other gospels is also implicit in the emergence of four-gospel syntheses at this time: Theophilus of Antioch (fl. ca. 169–183) is said to have compiled an early example (Jerome, Epistle 121.6; De viris illustribus 25; cf. von Campenhausen 1972, 174–75). He was followed some decades later by the fourfold gospel synopsis that Ammonius of Alexandria (ca. 175–242) constructed on the basis of Matthew—and which Eusebius (ca. 260–340) refined in the system of gospel parallels known as the Eusebian canons (see Eusebius’s Letter to Carpianus, lines 4–5 to dia tessarōn … euangelion; on Ammonius’s synopsis see further Crawford 2015b). In the fourth century Ambrose of Milan knew (and heartily disapproved!) of a number of such attempts at synthesis, as he did of apocryphal gospels in the names of Basilides, Thomas, Matthias, or the Twelve (Commentary on Luke 1.2; CSEL 32.11).
Justin Martyr (ca. 100–165) indicates that Christians read “the memoirs of the apostles” at their weekly meetings (First Apology 67). He remains notoriously inexact about quite what texts he includes in this category, but it is clear that they are the writings of “the apostles and their successors” (Dialogue 103.8)—that is, at least two of each, and on one reading precisely four, which would coincide nicely with the fact that Justin includes among them Matthew and (almost certainly) John as apostles, as well as Luke and Mark as apostolic students (for documentation, see Hengel 2000, 19–20 and nn.; more fully Hengel 2008b, 34–38; Stanton 2004; on Mark, see also Bockmuehl 2010, 84–86).
Despite some strongly suggestive passages especially in the First Apology, John’s Gospel admittedly appears less prominent in Justin—possibly because it was at first less widely used in the West (as forcefully argued by Watson [2013, 473–93], who however fails to engage in appropriate detail with Hill 2004, 316–42, 191–204). Justin is familiar with a few “extracanonical” but widely influential traditions like the birth of Jesus in a cave or the fire appearing in the Jordan at his baptism (Dialogue 78, 88). Notably, however, the only apostolic gospels Justin explicitly acknowledges are those that appear in the New Testament—and no noncanonical gospel is either cited or mentioned. Leaving aside Justin and his pupil Tatian, other second-century writers of different stripes seem notably less familiar with Mark and cite Luke infrequently while foregrounding Matthew and John. Although individual sayings (agrapha) are indeed sometimes quoted as “gospel” or as words of the Lord (see below, page 45), identified noncanonical gospels do not appear to exercise a public liturgical role as analogous written sources alongside Matthew, Mark, Luke, and John. This observation applies even in Valentinian gnostic sources like Ptolemy’s Letter to Flora, as Martin Hengel (2008b, 36–38) rightly notes, and is with very few exceptions further confirmed in the manuscript tradition. Important exceptions are the Infancy Gospel of James and the Epistle of the Apostles, which in some settings did exercise a relatively widespread liturgical role; but even these texts were not copied alongside the four gospels in ancient codices.
While both the terms “canon” (a rule or norm) and “New Testament” are used in the second century, the combination of these terms to designate a defined collection of writings appears only in the fourth century (cf. Markschies 2012g, 13–14; Nicklas 2012b).
As we shall see, however, this does not mean that any of the additional or alternative gospels ever achieved a comparable catholicity that might place them in competition with the four gospels, whether individually or as a fourfold whole. Conversely, even though Matthew and John were clearly more popular than Luke and especially Mark, none of the Four was ever seriously questioned as authoritative for the church.
Thus, the fourfold gospel status clearly emerged over the course of the second century and gradually gained in definition and exclusivity vis-à-vis some of the other permutations just described. This reality stands in contrast to occasional assertions that the distinctive status of the canonical gospels derived from a wholly unanticipated, “fictive” executive decision of fourth-century “theorizers” engaged in “suppressing or manipulating” others (Watson 2013, 454 and passim; contrast Watson 2016, 16–20). To be sure, in the late second century there is some limited evidence for the so-called Alogi, an anti-Montanist splinter group around a Roman presbyter who rejected the Gospel of John. But despite periodic assertions to the contrary the evidence is marginal at best, and recent scholarship has gone a long way toward demonstrating that there was no sustained opposition to the Gospel of John in the early church (see, e.g., Hill 2004 on Irenaeus, Against Heresies 3.2.9, and Epiphanius, Refutation of All Heresies 51.100).
Regional plurality and gradual convergence in the pattern of Christian gospel usage in no way detracts from the surprisingly early appearance of a widely acknowledged core of the fourfold gospel narrative, in both the East and the West. And as Christoph Markschies points out, signs of clear implicit reception and cross-referencing already between the four gospels themselves points to a first-century origin of this emerging core—even to the point that such material began to be quoted as “Scripture” alongside the Old Testament before the year 100 (Markschies 2012g, 26–27, 32–33, with reference to Matt. 10:10//Luke 10:7 in 1 Tim. 5:18; cf. Didache 13.2).
So far, so straightforward, one might think. Does this not offer us a clear view of the canonical nest to which we may now contrast the noncanonical cuckoo as the hostile newcomer (so Wright 2013, 358)? It might seem so. But that assumption would be a mistake. Just when all seems order and clarity, the situation turns out to be confusingly interesting!
The existence of gospels—indeed, numerous gospels—other than the Four was well known in antiquity, although for long periods of church history our surviving mainstream Christian literature considered them a fringe phenomenon (an impression that, by itself, could in theory reflect either de facto marginality or deliberate marginalization).
But then in the first half of the last century the inherited view of these texts was dramatically transformed by extensive manuscript discoveries in a landfill site in Upper Egypt at the ruined ancient city of Oxyrhynchus (“The Sharp-Nosed Fish”) beginning in 1896 and at Nag Hammadi (ancient Chenoboskion) in the late 1940s.
For our discussion in what follows, it will be worth noting that these manuscripts are mostly written on either papyrus or parchment. Manuscripts of both kinds could be part of codices (singular: codex), books formed by folding larger sheets and binding them together. Although not entirely unique to Christians, this medium of the codex, rather than the book roll or scroll, is widely recognized as the preferred Christian book technology for authoritative texts. On a papyrus page the front is typically called the recto while the back is called the verso. (See Leach and Tait 2000; Hurtado 2006, 84–86; and note the definitions of terms offered in the glossary of technical terms below, page 239).
Unlike most other newly found written sources from antiquity, the treasure trove of manuscripts that first came to light at Oxyrhynchus in 1896 proved to be so vast that first editions of its contents are still being published today. Bernard Grenfell (1869–1926) and Arthur Hunt (1871–1934), two young archaeologists from Oxford, joined work at an excavation that included the remains of an extensive ancient landfill site on the outskirts of the Egyptian village of Al-Bahnasa, 200 kilometers (about 125 miles) south of Cairo.
Before long this ancient rubbish deposit yielded tens of thousands of documents, of which even now only a minority have been published. At the time of writing, over eighty volumes of papyri are in print (most recently Gonis et al. 2016), representing about 5 percent of the total—though a good deal of the remainder is in very small fragments. Most of the texts are in Greek, although there is some material in Coptic, Demotic Egyptian, Latin, and (for the latest period) even Arabic. Among a wealth of classical literary as well as scientific and documentary sources, there were large numbers of biblical and other ancient Jewish and early Christian texts from antiquity all the way to the seventh or early eighth centuries, with the bulk of material clustered between the first and third centuries. (I will here for convenience designate all Oxyrhynchus fragments as they were officially published, even though subsequent scholarship has suggested not only that some extraneously acquired fragments also originated from this site, but also that some early published “Oxyrhynchus” finds may instead derive from the region of Fayûm, about 100 km [62 mi.] closer to Cairo; see Blumell 2012, 89–162, on the extensive epistolary networks and traffic patterns to and from Oxyrhynchus.)
All in all, the literary material from Oxyrhynchus appears to have comprised five major book collections, discarded at different periods during the lifetime of the site (Houston 2014, 130–79). Among the various Christian writings were about twenty examples of what are often somewhat imprecisely called “apocrypha,” that is, quasi-biblical texts with subjects related to the New Testament. In addition to substantial parts of what we now know as the Gospel of Thomas, there were fragments of gospels in the names of Mary, Peter, and James, along with three or four other previously unknown gospel fragments. In the early decades of the twentieth century, this gave rise to a lively discussion about the significance of this seeming profusion of new and unknown gospels. Many other questions remain unanswered—not the least of which is why so many biblical, nonbiblical, and other manuscripts were consigned to an apparent rubbish dump (provocatively explored by Luijendijk 2010; cf. more recently Houston 2014, 130–79; and more generally Blumell and Wayment 2015, on the Christian texts from this site).
Continued study of the Oxyrhynchus discoveries has had far-reaching consequences for our understanding of Greco-Roman culture in late antiquity, and of Christianity’s place within it. But a generation later came a further Egyptian discovery that was tiny by comparison but at least initially seemed even more dramatic for an understanding of early Christianity.
In December 1945, farmers near the village of Nag Hammadi, several hours’ drive farther up the Nile (540 km [335 mi.] by road from Cairo), discovered a collection of thirteen leather-bound ancient codices in a pottery jar concealed at the foot of a cliff. Reliable accounts of their discovery are hard to come by, and successive versions related by James M. Robinson and other editors seem to have gained in the telling and become notably contradictory (see Goodacre 2013). One of these volumes and part of a second one were destroyed before they could be studied, but the rest turned out to contain a wealth of mainly gnostic writings that were initially thought by some to have been part of the library of the nearby monastery founded by St. Pachomius (ca. 290–346), and possibly to have been discarded at a time of tightening canonical boundaries. The improbabilities surrounding both this Pachomian theory and the original discovery narratives have encouraged other, perhaps more likely explanations. Among these is the idea that the manuscripts were an eclectic collection of privately commissioned copies, buried as part of their owners’ grave goods or Christian “books of the dead” (thus Denzey Lewis and Ariel Blount 2014)—a point rendered plausible by their discovery on the site of a large ancient burial ground.
The surviving fourth-century manuscripts are in Coptic translation rather than the original Greek, but they nonetheless represent writings that in some cases were composed as early as the second century. Several are presented as New Testament apocrypha. For our purposes the most significant ones may be those with these titles:
These Coptic documents are now all available in accessible editions and English translations (see below, p. 52).
Because the controversial terms “Gnosticism” and “gnostic” are frequently applied to texts like those found at Nag Hammadi, this may be an appropriate point at which to offer a brief definition. While the term itself derives from the Greek word for “knowledge” (gnōsis), its significance here is in relation to a highly fluid and diverse set of religious groups in the early Christian centuries. “Knowledge” may function as a technical term as early as the later writings of the New Testament (famously in 1 Tim. 6:20, where this concept already appears related to a preoccupation with “myths” and genealogies, 1:4). The noun’s complete avoidance in the Johannine writings may also be significant.
Many key ideas about access to secret and otherworldly salvific knowledge for the few were anticipated in popular Middle Platonism and esoteric mysticism. In the most general sense they have a wide and almost timeless currency across diverse religious, philosophical, and even pseudoscientific manifestations, whether ancient, medieval, or modern. Specifically Christian gnostics experienced their heyday in the second, third, and fourth centuries. They developed sometimes elaborate mythologies influenced by Christian and Jewish scriptural texts (not least the opening chapters of Genesis) as well as by Platonic philosophical ideas about the origin and nature of humanity and the cosmos. Scholars often distinguish Valentinians, on the one hand, from the more elaborate mythologies of Sethians, “Barbelo gnostics,” and Ophites, on the other (e.g., Rasimus 2009); but there is still no consensus on these typologies. Certain gnostic ideas enjoyed a long and intriguing afterlife from late antiquity to the Middle Ages and beyond and are sometimes thought to have influenced such groups as the Manicheans and (much later) the Bogomils and Cathars.
As the name suggests, one of the key ideas uniting otherwise diverse and differentiated groups was the belief that the adherents were an elite gaining privileged access to knowledge of divinely revealed insights. This secret knowledge was understood to carry a saving significance in that it allowed the initiates to escape humanity’s fleshly condition, liberating their true divine spark from its imprisonment in a corrupt materiality that is the evil design of an inferior creator (the workman or demiurge).
In view of the complexity of the evidence, few generalizations about gnostics or Gnosticism are likely to prove universally serviceable. Indeed there has been a lively debate about whether the “gnostic” terminology serves any useful purpose at all, with some scholars casting doubt on whether these terms can ever be meaningful (M. A. Williams 1996; cf. King 2003b; and more cautiously Marjanen 2005). Certainly an important upshot of such debates is the recognition that there was no single movement we could call “Gnosticism.”
It is of course true that “gnostic” labels have often been deployed indiscriminately or polemically. For our discussion it matters that such terminology should not be invoked in order deliberately to load the critical dice when discussing the apocryphal gospels. The fact is that the terms “gnosis” and “gnostics” were indeed widely used in antiquity for certain philosophical schools or haereses, by Christian and non-Christian outsiders (e.g., Plotinus) as well as by insiders. Even mainstream writers like Clement of Alexandria and Origen implicitly acknowledged the validity of concerns for saving knowledge in their rather different adoption of the language of gnosis in the service of catholic theological ends. So the active use of the terminology is significant even if the term “gnostics” (gnōstikoi) is relatively rare as a self-designation.
“Gnosticism,” by contrast, is a potentially misleading modern analytical construct (invented by the seventeenth-century commentator Henry More), which for the sake of clarity we will avoid in this book. (See further M. Edwards 1989; 1990; Layton 1995, 338; Markschies 2003, 10–11; also McGuire 2010, 203–5 with nn5–7.)
On a related note, recent scholarship has also repeatedly called into question whether several of the leading second-century founders of such groups were themselves gnostic in any distinctive sense. Martin Hengel, perhaps the late twentieth century’s leading New Testament historian, acknowledged the profound influence such prominent teachers clearly exercised on subsequent developments, and he assigned the investigation of this problem to several of his doctoral students and others. Their publications include Winrich Löhr 1996 on Basilides (fl. 117–138), Christoph Markschies 1992 on Valentinus (d. ca. 165), Niclas Förster 1999 on Marcus (founder of the Marcosians, mid-2nd cent.), and Ansgar Wucherpfennig 2002 on the Valentinian commentator Heracleon. Similar work on Ptolemy, the author of Letter to Flora explaining the Valentinian approach to the Old Testament, remains desirable.
In a much-cited essay (2008a), Martin Hengel argues for the development of these gnostic movements around the year 100 CE out of Christian, Jewish apocalyptic, and Middle Platonic roots (cf. further Lahe 2012; Drecoll 2013). In particular, Hengel suggests that the attractiveness of gnostic ideas may have been in combining a deep disillusionment about apocalyptic eschatology in the wake of the catastrophic Jewish War with the educated Christian desire for a viable philosophy of religion in Greco-Roman intellectual culture (560–63, 589–92). Such a cultural aspiration may also animate Valentinian ethics: as an attempt to provide an intelligible Christian philosophical account of the good life, gnostic moral teaching stresses ideas like the escape from materialism to a spiritual transcendence, the control of emotions, and the rational elimination of excess—from preoccupation with sex to the union of alienated gender differences (see Dunderberg 2015; also Tite 2009).
This question of the origin and appeal of gnostic ideas is clearly a large and complex topic to which we cannot do full justice here. (For useful further reading, see Brakke 2010a; Logan 2006; Marjanen 2008; Markschies 2003; Pearson 2007; van den Broek 2013. The most comprehensive overview of Valentinianism remains Thomassen 2006.)
In Europe and North America, the third millennium of the Christian calendar began on a note of surprisingly widespread confusion about Christianity’s origins, in the media and even in the churches. That confusion was fueled in no small part by several cleverly marketed new (or newly reinterpreted) discoveries of ancient artifacts, including a supposed bone box (ossuary) of James the brother of Jesus; a supposed family tomb of Jesus in Jerusalem, which somehow came to be rebranded and republicized as such several decades after its discovery; and a manuscript containing the so-called Gospel of Judas, whose existence had been rumored ever since its clandestine discovery in 1978, and which was published in 2006 to much fanfare by none other than National Geographic magazine. In 2012 Karen L. King of Harvard Divinity School caused a considerable stir when she announced (and later published, King 2014) a papyrus fragment mentioning Mary Magdalene as the “wife” of Jesus, although this was subsequently exposed as a modern fake (see below, p. 187).
We will return to these texts. But even before these finds one could not fail to notice the extraordinary media circus surrounding The Da Vinci Code (D. Brown 2003), a blockbuster novel predicated on wholly fanciful theories about the repercussions of Mary Magdalene’s imagined marriage and children with Jesus. Swashbuckling tales of conspiracy and deception at the heart of religion or power retain a timeless potential to entertain the gullible while generating impressive streams of revenue for their promoters (not to mention for industries like Hollywood and tour guides from Saint-Sulpice in Paris to Rosslyn Chapel outside Edinburgh). More recently, comparable historical nonsense on stilts was in 2014 “discovered” to fresh media fanfare in the supposed “decoding” of a seventh-century manuscript paraphrasing Joseph and Asenath, an early Jewish or Christian apocryphal narrative about the conversion of Joseph’s pagan Egyptian wife. Contrary to appearances, this “lost gospel” supposedly encodes the secret of Jesus’ marriage to Mary Magdalene, who, like Asenath, gave birth to two sons (Jacobovici and Wilson 2014; cf. Gen. 41:45, 50; 46:20). And so it goes on. Given Christianity’s accelerating and partly self-inflicted decline in public influence and credibility throughout many Western societies, the loss of old certainties unsurprisingly yields the stage to more fanciful counternarratives about Christian origins.
It is not sensationalist or misleading to point out that the ancient church was indeed aware of the existence of a large number of other gospels or gospel-like texts. The gradual acceptance of settled canonical boundaries in turn entailed a more confident demarcation of documents that as a result had not become canonical. (See Watson 2013; the more problematic notion of literature that “became” apocryphal or canonical has been particularly stressed by Lührmann 2004; Lührmann and Schlarb 2000; but see also the critique in Nicklas 2011.)
A pivotal twentieth-century contributor to this conversation was Walter Bauer (1877–1960). Known to students of New Testament Greek above all as the originator of a definitive lexicon (Danker, Bauer, et al. 2000), he also became ideologically influential in the 1920s for his depiction of Jesus as a “syncretistically softened” anti-Judean Jew who had grown up in a hellenized Gentile setting “in considerable freedom from the Law,” disdainful of “levitical purity” and of the Temple—as a place of conflict rather than of worship (Bauer 1967, 102–3, 108). His sort of liberal Protestant Jesus was hardly original, but seemed before long to lend grist to the mill of “German Christian” New Testament scholars determined to discover a Jesus who was not Jewish at all (e.g., Grundmann 1940, esp. 175; cf. Ericksen 1985; Head 2004; Heschel 2008).
More significantly for our purposes, however, Bauer’s book Orthodoxy and Heresy in Earliest Christianity, first published in 1934 (English translation, Bauer 1971), became particularly formative in postwar scholarship through its idea that “heretical” beliefs were not historically a deviation from singular Christian “orthodoxy,” but rather were the dominant expression of a fundamentally diverse and plural faith from which “orthodoxy” emerged only at a later stage. This view, seemingly boosted by fresh discoveries like those at Oxyrhynchus and Nag Hammadi, commended itself to the study of gospel literature among the students of Helmut Koester and James M. Robinson, as well as among members of the late-twentieth-century North American Jesus Seminar (see Koester 1990; Koester and Robinson 1971).
In more popularizing scholarship influenced by the Bauer and Koester schools, conspiracy-minded interpretations have frequently asserted that these ancient texts were at first of the same status as, and at least in some cases earlier than, the canonical gospels, superior witnesses to the real essence of the Jesus movement, and freely proliferating in Christianity’s charismatic infancy until they came to be brutally suppressed by authoritarian churchmen, perhaps at the emperor Constantine’s beck and call.
Bauer’s scholarly reception by Koester, Robinson, and their students encouraged the emergence of a view that the gnostic gospels in particular offered access to the authentic original genius of the Christian message—a view that has energized writers ranging from scholars like Elaine Pagels (1979) to racy fiction writers like Dan Brown (2003). Related to this are attempts to date the canonical gospels exceptionally late while insisting that certain noncanonical sources, including the Gospel of Thomas, grant exceptionally early access to the teaching of Jesus. That is the approach of a writer like John Dominic Crossan (1991, 427–30), who asserts no fewer than fifteen sources of “independent attestation” predating the Gospel of Mark, or of the optimistically titled volume The Complete Gospels, which conveniently presents “for the first time anywhere all twenty of the known gospels from the early Christian era,” all of which are said to be “witnesses to early Jesus traditions” (Robert Miller 1994, cover, 3).
Certain observations are, however, important to bear in mind if we are to keep the second-century profusion and variety of Christian literature in perspective. One pertinent insight to be elaborated in the course of our discussion is this: while scholars from time to time postulate the existence of primitive texts like Q or early sources of Thomas, no extant alternative gospel forms or attestations predate the New Testament Four. Even a large and diverse collection of early Christian literature like that at Oxyrhynchus turns out to corroborate the popularity of the two mainstream apostolic Gospels of John and Matthew. These were evidently—there as elsewhere—the most widely read and copied.
As for the apocryphal gospels, at one level the overwhelmingly Egyptian evidence is what we would expect, given the extent to which the climate favored the survival of papyri. But there, to some extent, lies the rub for a good deal of the evidence on which Bauer’s hypothesis draws: can the discoveries at Oxyrhynchus or indeed those at Nag Hammadi (which followed the publication of his book) really grant us representative insights into the nature of early Christianity more generally, or are they perhaps eloquent first and foremost about themselves and their own context—namely, at the core of the singular Nag Hammadi collection and nearer the periphery of the vast Oxyrhynchus finds? These discoveries do illustrate the rich diversity of theological approaches in early Christian circles of Upper Egypt—but they cannot straightforwardly establish the priority and predominance of heterodoxy in quite the way that Walter Bauer assumed. (Blumell [2012, 318–25] offers a valuable if incomplete inventory of Christian evidence from Oxyrhynchus published up to 2010, which implies a proportion of canonical to noncanonical gospel fragments in purely numerical terms at around 2.5 to 1, depending on which centuries are included. More significantly, only the dominant Gospels of Matthew [15x] and John [13x] seem to be attested consistently in every century up to the time of canonization, and they appear respectively five times and four times as often as the three confirmed fragments of Thomas, their nearest noncanonical rival; there are two fragments for Mary and arguably just single attestations for the nine other noncanonical gospel-like texts. Canonical and noncanonical gospels are not found together within the same manuscripts.)
In other words, even for Egypt the manuscript finds may help underscore the serendipitous and marginal or subsidiary character of what was discovered. Some readers at Oxyrhynchus were evidently interested in apocryphal gospel literature—but as we just saw, never to the extent that any of these texts competed with the preferred Gospels of Matthew or John, even if Luke and Mark are admittedly rarer. Nag Hammadi offers fewer statistical clues, but we have here a dozen books about whose status and representative currency we can have no assurance on the basis of this single find. This point becomes more significant if one considers how very few of these texts generated multiple copies or translations, let alone commentaries.
All in all, these observations certainly do not invalidate Bauer’s thesis, but they do urge considerable caution. The appeal of authoritarian suppression theories casts a long shadow—not least for a Protestant romanticism that loves to lionize an imagined primitive charismatic anarchy being crushed by authoritarian institutions and orthodoxies. In such scenarios Hegel’s philosophy of history, Adolf von Harnack’s nineteenth-century rediscovery of Marcion, and a popular heroic mythology of Luther’s battle against the pope are never far from the surface.
We must certainly take on board the important questions about early Christian diversity that were raised by the Bauer thesis and the Egyptian discoveries of the twentieth century. But this cannot make the evidence from the sands of Egypt yield answers as straightforward as either traditionalist or skeptical accounts would have us believe.
The idea of a fourfold apostolic gospel of Matthew, Mark, Luke, and John emerged during the first half of the second century and continued to gain in strength until it formally prevailed in the fourth. But other gospels and gospel traditions richly proliferated in the later second and third centuries, many of them informed—either directly or more often indirectly—by the narrative outline of the Four. Far from these documents being eliminated as “heretical” departures from a clear, uninterrupted orthodox line from the start, many appear to have coexisted happily with the protocanonical tradition and even taken it for granted.
Mainstream church leaders did indeed voice opposition to such alternative accounts in either written or oral form, whether or not they knew them at first hand. Such conflict began not in the fourth century but in the second, if not earlier. Several New Testament documents already explicitly discount false or inaccurate renderings of the ministry or teaching of Jesus (see, e.g., Matt. 5:17, 19; Luke 1:3–4; John 1:8; 6:66; 21:23–24; also 1 John 4:2–3; 1 Cor. 15:14–18).
But the idea that the noncanonical gospels disappeared from view simply or primarily because they were formally silenced by church authorities founders on several contrary facts.
First, the supposedly suppressed documents were evidently known and read by some, but—judging from the manuscript evidence—appear never to have gained widespread popularity, circulation, or acceptance. The following table illustrates this well.
Precise dates and therefore absolute manuscript statistics are admittedly always debatable; but it is now factually incorrect to claim that there are no pre-300 manuscripts of Mark (so, e.g., Watson 2016, 3–4; but note 45, P.Oxy. 5073, and perhaps 88), let alone that “papyri dating from 100 to 300 CE are equally balanced between canonical and non-canonical gospels” (Burke 2013a, 29, citing only Koester 1980, who was, among other things, unaware of at least eight as yet unpublished papyri on John [90, 95, 106, 107, 108, 109, 119, 121], five on Matthew [101, 102, 103, 104, 110], one on Luke , and one or possibly two on Mark [P.Oxy. 5073; 88?], all predating the year 300).
|Protevangelium of James||1x|
|Gospel of Mary||2x|
|Gospel of Judas||1x|
|Various “Unknown” Gospels||5x|
|This table is adapted from Hurtado 2013, a fuller inventory than the list of papyri in the appendix to Nestle-Aland28 (Nestle et al. 2012, 792–99).|
Those gospels that went on to become canonical in the fourth century are also the ones that were most frequently read and copied before 300, as well as most frequently cited and commented upon—Mark being a partial exception on both counts. It is a particularly telling additional observation that the early-third-century manuscript 45, possibly our earliest unambiguous multiple-gospel codex, includes parts of four and only four gospels: Matthew, John, Luke, and Mark—in that order, plus Acts. (Skeat  suspected an even earlier example in a compilation of 4 + 64 + 67; this has not been widely accepted, but Gathercole [2012b, 218, 235] does suggest “the possibility (but no more)” that 4 formed part of a codex containing both Matthew and Luke.)
Even if one counts all extant manuscripts prior to the invention of the printing press, the copies of gospels composed before 300 remain in single-digit numbers except in the case of the four New Testament gospels—and the Infancy Gospel of James (along with later translations of public texts like the Diatessaron).
Significantly, too, there are no extant manuscripts from antiquity (whether before or after canonization) that combine canonical with apocryphal gospels. The manuscript record does suggest that some or many early Christians knew both sorts of texts, and in some places like Rhossus or Oxyrhynchus some of them clearly read both sorts (see Hurtado 2015). But their manuscripts evidently distinguished between them. And Christians did not copy or read them as equivalents side by side, contrary to a twentieth-century scholarly prejudice in the wake of Walter Bauer that continues to be popularized to this day (nicely illustrated by collections like Robert Miller 1994 or for that matter Taussig 2013).
|Gospel of Thomas||4x|
|Gospel of Philip||1x|
|Gospel of Mary||3x|
|Gospel of Judas||1x|
|Sophia of Jesus Christ (Wisdom of Jesus Christ)||1x|
|Gospel of Peter||1x|
|Infancy Gospel of James (Greek manuscripts only; numerous translations and derivatives exist)||Over 150x|
All this says a great deal about a process of dissemination and acceptance of popular authoritative texts that, especially in the early centuries, was far too diverse and widespread to be explicable in terms of structures of authoritarian imposition or censorship.
Official suppression is of course one possible explanation for this disparity of attestation—and for particular cases this possibility cannot be categorically excluded. But occasional efforts to blacklist various documents (for example, by Irenaeus of Lyons and Serapion of Antioch in the second century or by the Gelasian Decree in the late fifth or early sixth) appear for the most part to have had little effect. The effective replacement of the Diatessaron with the fourfold gospel in Syrian churches of the fifth and sixth centuries demonstrates that some “opposed” texts did disappear; but others, like the Infancy Gospel of James and various Pilate cycles, continued to go from strength to strength.
Until the sixth century—and perhaps until considerably later—the church simply did not have the power to make such texts go away. Of course it could merrily anathematize and in certain places sporadically take or threaten action; but it could not, it seems, successfully enforce. With the notable exception of para-canonical texts like the Diatessaron and the Infancy Gospel of James, the absence of noncanonical gospel literature from the Eastern and Western churches’ public liturgical reading may be simply that—the nonappearance of texts that failed to attract a sufficient communal readership to establish themselves as universal Christian “classics,” that is, popularly received Christian texts that could garner consensus and stand the test of time and faith. As has been the case for popular religious literature and music through the ages, eventual success or failure was above all a function of their power to engage their subject matter credibly and authentically in the service of the faithful—and, at least partly as a result, their ability to weather occasional periods of local, popular, or official opprobrium.
An interesting early medieval confirmation of this view is provided by the discoveries in the genizah (storeroom for disused manuscripts) of the Old (Ben Ezra) Synagogue in Fustat near Cairo. These have long been noted to include palimpsests (scraped and recycled manuscripts) of the New Testament texts of Matthew, John, Acts, and 1 Peter. In addition to Greek Old Testament translations of Aquila and of Origen’s Hexapla, the genizah also contained translations of New Testament texts and lectionaries into languages including Arabic, Judeo-Arabic, and Syriac (on which, see Niessen 2009; some may have been used in part for liturgical purposes). But aside from the well-known medieval Toledot Yeshu fragments (on which see below, p. 129), not a single apocryphal gospel has turned up among the Christian texts and palimpsests of the Cairo Genizah. (On a more speculative note, Piovanelli (2011, 92–96) nevertheless suggests close links between the Toledot Yeshu and the Gospel according to the Hebrews as well as the Gospel of Judas.) This absence of Christian apocrypha seems particularly notable since the Cairo Genizah’s somewhat catholic selection of texts included a number of Jewish noncanonical or “apocryphal” writings like Hebrew Sirach, Aramaic Levi (a source for the Testament of Levi), and the Damascus Document—whether these attest a surviving ancient manuscript tradition or were later rediscovered, like the cache of Hebrew manuscripts found near Jericho at the end of the eighth century (mentioned in a famous letter by Patriarch Timothy I [780–823]; for text and translation, see Reeves 1999, 174–77).
As a deliberately brief and accessible guide to this complex and newly reinvigorated field of study, this book does not intend to break new ground or push the envelope on basic historical-critical questions of the authorship, date, and setting of the noncanonical gospels. The aim is to develop the argument in the context of a fairly middle-of-the-road approach to most critical debates, rather than to advance the field in this respect.
Five main emphases constitute the basic argument of this book.
1. First, the aim is to provide an introduction that is both accessible and nonsensationalist while offering a sympathetic account of these writings in relation to what became the New Testament. This involves taking the texts seriously on their own terms and in relation to a centrist range of assessments within mainstream critical scholarship. It will also relate them to their place within the reception history and formation of what was to become the canonical fourfold gospel.
2. This approach also favors the conviction that it is legitimate and instructive to read these texts alongside the New Testament gospels. In doing so we will find that their status can be usefully understood as epiphenomenal and supplementary to that gospel tradition. In relation to its narrative structure as well as its status as public and “apostolic,” all noncanonical gospels presume that New Testament tradition’s existence; many of them presuppose its substance or even its wording. In some sense, therefore, the apocryphal gospels occupy what at least in retrospect can be described as a “para-canonical” perspective—whether their intent is to supplement and reaffirm, to replace, or to subvert the four gospels that became canonical.
As I will suggest, this is routinely the narrative perspective they adopt, and sometimes their explicit self-understanding, whether or not they identify themselves (like the Gospel of Thomas, the Gospel of Judas, the Apocryphon of John, and others) as “apocryphal” or “secret”—and therefore whether or not they intend any bid for public ecclesial status at all. This para-canonical identity pertains even for texts that are formally or chronologically nearer the New Testament canon than others. (On this point it is helpful to consult the articulation by Luke Timothy Johnson  of the respective “canonical” settings of John and Thomas, in contrast to the alternative account of that relationship offered by writers like Elaine Pagels . See also below, p. 42, for an attempt to define the slippery term “apocryphal.”)
3. We will repeatedly find our attention drawn to one simple and obvious but easily overlooked feature of our source material, which in turn encourages this para-canonical way of looking at the texts. Of the dozens of noncanonical gospel-like documents from antiquity, whether familiar from ancient citations or only through manuscript discoveries, not a single surviving text offers an alternative narrative account of the kind provided in the four New Testament gospels. That is to say, none of them trace what Jesus did and said and suffered from his baptism through his public ministry to his crucifixion and resurrection. Apparent exceptions to this rule are either closely dependent on the text of the four gospels, like the second-century synthesis known as the Diatessaron, or else belong in this form to a much later period, like the Jewish antigospel Toledot Yeshu or the Muslim Gospel of Barnabas.
It is true that quite a number of fragmentary texts might in theory have provided such an alternative narrative—including, for example, the Gospel of Peter or Papyrus Egerton 2. But none is extant—or even attested in the ancient literature. Nor did any alternative gospel-style narrative accounts of Jesus’ mission and ministry from birth or baptism to death or resurrection experience significant attestation or circulation in antiquity. The canonical Four are thus notably distinctive in this regard; indeed their Markan outline appears in one way or another to have been the narrative reference grid for at least the large majority of noncanonical gospels.
4. Quite how that relationship between noncanonical and canonical gospels works intertextually will be a matter for repeated reflection in the following chapters. Most typically, we will encounter gospel-like texts showing a marked editorial distance from the New Testament gospels, while nevertheless revealing their own (and their readers’) presupposed consciousness of the narrative framework and even the wording of those protocanonical gospels to a greater or lesser extent.
We will thus see that traditional scholarly notions of literary dependence, when narrowly understood in terms of scribes working with written texts, are rarely serviceable for this relationship between the fourfold gospel and the others. In trying to describe this clearly epiphenomenal but often somewhat loosely or indirectly articulated connection, it seems in many cases preferable to think in terms of antecedence and influence rather than a relationship of direct dependence on a written text. This accounts for the frequent presence of shared themes or phrases while also explaining the considerable literary freedom and independence which some of the noncanonical texts manifest at the same time. (Others, like Foster [2010a, 116–17], prefer to maintain, e.g., for the Gospel of Peter, a somewhat extended notion of “literary dependence” that might include “drawing upon a literary work from memory.” This is a conceptually helpful clarification, but it still presupposes the mediated antecedent to be a distinct written text rather than, say, an informal harmony or conflation.)
5. Finally, another occasionally useful frame of reference can be the concept of “social memory,” which considers the social, cultural, ritual, and religious dimensions of how communities remember their past and understand their identity (see Dignas, Smith, and Price 2012; Fentress and Wickham 1992; the relevance for the gospel tradition is explored by Dunn 2007; Kelber 2002; Kirk and Thatcher 2005; Le Donne 2009). In a surprising number of cases the protocanonical pattern of attributing gospels to key apostles or their immediate disciples also characterizes the noncanonical gospels; and at least until the second century it remains in theory possible that such associations are informed in part by appeal to the often contested living memory of these apostolic figures or their students, as they were for Ignatius, Justin, and Irenaeus (see Bockmuehl 2010 and 2012b on Simon Peter).
There is always something unquestionably exciting and intriguing about public announcements that an ancient text about Jesus has come to light—or perhaps even just a papyrus fragment of such a document. What if it contains genuine sayings or stories previously unknown about him? What if it reveals mysterious or secret truths about Jesus of Nazareth or his followers, authentic insights into the earliest Jesus movement?
But it is in the nature of this material that much of what we are dealing with is either highly fragmentary, lost, or perhaps even entirely hypothetical. This fragmentary nature of our texts is self-evident for papyri like those found at Oxyrhynchus and elsewhere, but it is also clear from a number of apparently well-known documents that are repeatedly mentioned in early Christian writings, but whose actual text is only ever cited in passing or quoted in occasional short snippets in the church fathers. So how do we study this confusing wealth of disparate material outside what is presented to us so tidily in the four complete narrative gospels of the New Testament?
To arrive at a proper reading we must first find a serviceable way to describe what we are dealing with. How many texts are there? One reason the noncanonical gospels often seem particularly formidable and bewildering to the nonexpert is the sheer difficulty of even establishing how many sources we are talking about. It is one thing to discover that in antiquity more than the four biblical gospels were known, or perhaps that one of them was a text known as the Gospel of Thomas and another called the Gospel of Judas. What can be more confusing is to open one of several excellent recent volumes of translations to discover quite how many of these “gospel” documents there are, surviving either in their entirety or more often as fragments. One ballpark figure often cited is that there were approximately forty “other gospels” (e.g., Tuckett 2005; cf. Ehrman and Pleše 2011, viii); but the 1,500-page German work of Markschies and Schröter 2012 contains around twice that number (depending, inevitably, on how one counts). Ancient sources certainly cite or report dozens of other noncanonical gospels or gospel-related texts. For some of these we have fleeting descriptions or quotes, but no trace survives of many others.
It is easy to feel dismay or paralysis at the complicated mass of this material, much of which is either lost or fragmentary, and about whose original size and shape we can only speculate. The following two lists will help to illustrate the broad scope of the material, and the problems of taxonomy, by means of one ancient and one modern inventory of “apocryphal” gospels.
Below is a list of prohibited texts in the so-called Gelasian Decree (Decretum Gelasianum), a Latin document of uncertain provenance and authority transmitted under the names of popes including Damasus I (366–384) and Gelasius I (492–496) but thought to have been compiled more unofficially in the sixth century. The list below (culled from Klauck 2003, 3–5) includes only the gospel-related texts, in the order in which they appear. The text of the decree explicitly identifies each item as “apocryphal” (and therefore rejected).
The inclusion of one or two of these items is admittedly doubtful and reflects the extent to which the term “apocryphal” was applied fairly liberally to suspect documents. Faltonia Betitia Proba’s Cento, for example, however disagreeable to the author of the decree, is certainly not apocryphal in any sense either of secrecy or of supplementation or competition with the canonical gospels. In fact, it is a poetic composition that seeks to recapitulate them through the educated and aesthetically refined medium of Virgilian verse, which in turn facilitated a wealth of fresh allegorical associations (see Sandnes 2011, 141–80).
While the final category, amulets, perhaps was included in this catalog somewhat whimsically (but in keeping with general early Christian opposition to magic), it does hint at a relevant point for our purposes. Recent text-critical research has drawn attention to the neglected category of noncontinuous biblical texts. Among these are amulets containing or echoing biblical quotations, including, for example, P.Oxy. 5073, which is now by at least a century the earliest attestation of the text of Mark 1:1–2 (Head 2013, 439–43). As a number of these noncontinuous texts appear to redeploy gospel quotations in amulets or other unofficial formularies, they clearly constitute an interesting interstitial category between the customary ecclesial forms of biblical manuscripts and lectionaries, on the one hand, and popular praxis and belief, on the other. It is significant for our purposes that virtually all known examples of such talismanic gospel texts consist of authoritative canonical rather than apocryphal gospel excerpts—usually incipits (opening lines of works or passages) or the Lord’s Prayer. (The main exception is the story of Jesus and Abgar, on which see below, p. 121; de Bruyn 2015, 156–60, 173–74.) Greek examples are conveniently cataloged in de Bruyn and Dijkstra 2011, nos. 4, 8, 19, 21, 22, 26, 36, 38, 44, 45, 50, 59, 70, 77, 84, 95, 105, 117, 122, 134, 146, 148, 156, 157, 182; there are many others in Coptic. See further de Bruyn 2010; Sanzo 2014; also Kraus 2004; 2007 on problems of classification and definition.
This section is adapted from the large collection edited by Markschies and Schröter (2012). For ease of reference it follows their sequencing of the material (see, e.g., pp. ix–xii), lightly adapting the structure for easier representation.
It is not hard to find oneself multiply confused or overwhelmed by conflicting catalogs like these! For one thing, they suggest that the question of an overall document count may be the least of our problems. Much of what is known from ancient sources is not extant, while a good deal of what is extant is “unknown”—that is to say, we have no way of linking it either with other surviving texts or fragments or with sources mentioned in antiquity. There are of course occasional exceptions to this state of affairs. Once in a while a lost ancient text may indeed come to light (e.g., the Gospel of Judas); careful scholarly study may suddenly stumble upon a demonstration that one previously unconnected fragment of papyrus belongs with another, well-known text (e.g., P.Köln 255 with P.Eger. 2). But these are happy exceptions in what in many other respects remains a frequently perplexing state of affairs.
Then again, one cannot leaf through more than a few of the documents in the major collections without stopping to ask oneself in what sense some of these items can really be said to represent gospels, even on a broad nontechnical definition of that term. The Oxford English Dictionary, for example, characterizes apocryphal gospels as “certain ancient lives of Christ of a legendary character.” Should we really count ancient texts that show no interest in the life of Christ, whether legendary or otherwise? If not, our inventory immediately becomes very much shorter. What is more, no two scholars’ lists or taxonomies seem to agree: it is remarkably difficult even to describe, let alone to categorize, what we are dealing with.
It will help to begin our task by asking what defines an “apocryphal” gospel and questioning a few conventions that are likely to obscure rather than to clarify.
The designation of texts as “apocryphal” often carries the negative connotation of an implied value judgment between apparently normative texts and others that are deemed extraneous and quite possibly suspect. The ancient church’s usage of the term in relation to Jewish or Christian writings was overwhelmingly pejorative (for the Western church, see the documentation in Gallagher 2014). For this reason some interpreters prefer to speak only of “early -Christian” texts without singling out some as canonical and others as inferior: the very notion that some gospels are canonical is on that view a late and somewhat arbitrary fourth-century imposition upon texts which until that point had happily coexisted and cross-fertilized each other (thus Watson 2013).
Yet one of the more intriguing aspects of the texts we are studying is precisely the question of their historic place, and sometimes even their literary presentation, as in some sense “hidden,” apokrypha. This notion has had a variety of meanings in different contexts.
The Old Testament already implies the possibility of hidden secret knowledge that can be revealed only by God rather than by human inquiry: the God of Israel alone is the author and dispenser of wisdom (Prov. 1:7; 2:6; 20:27; cf. Job 12:22; Amos 3:7). A classically influential text in this respect was Deuteronomy 29:28 (29:29 in the Vulgate and most English translations), famously supplied in Masoretic manuscripts (and still in the standard modern critical edition, Biblia Hebraica Stuttgartensia) with ten scribal dots marked across particular words to warn against dangerous speculation: “The secret things belong to the Lord our God, but the revealed things belong to us and to our children forever, to observe all the words of this law.” The “secrets” in this case may well denote the future, as the preceding context implies (so Fishbane 1985, 540), but for Wisdom literature like Sirach 3:22 (cf. 20:30) or Tobit 12:7 there are other secrets of wisdom that should be kept concealed from the uninitiated. (See further Bockmuehl 1990, 66–68.)
Matters are once again different for the Christian texts related to the gospels, with which we are here concerned. Certainly it is true that anathemas or assertions of their “apocryphal” status (e.g., in documents like the Gelasian Decree, cf. above, p. 32) are denials of legitimacy or authority. This derogatory usage seems at least implicitly to be found as early as Hegesippus in the second century (cited in Eusebius, Ecclesiastical History 4.22.8: apocrypha composed by heretics) and Origen in the third (e.g., Commentary on John 2.31.188; Commentary on Matthew 10.18; Epistle to Afri-canus [Migne 1857–86, 11:65, 80]).
That said, some of the noncanonical gospels do indeed make explicit claims to contain material that is hidden, secret, or indeed—in that specific sense—apocryphal. Most obvious among these is the Gospel of Thomas, which famously begins with the words, “These are the secret words which the living Jesus spoke, and which Didymus Judas Thomas wrote” (incipit). A similar claim opens the Gospel of Judas, which claims to present “the secret discourse of revelation that Jesus spoke with Judas Iscariot” (Codex Tchacos 33). By implication the Lord’s instruction to Mary Magdalene in the Gospel of Mary is similarly identified as secret by Peter within that text (Gospel of Mary 17; see Tuckett 2007, 127, 188). The gnostic library of Nag Hammadi contains several other examples.
So it is interesting that the term “apocryphal” crops up as a deliberate self-designation in the opening statement or framing narrative of (frequently gnostic) texts that propose their material by means of the literary fiction of an alternative, “hidden” tradition about Jesus. Such hiddenness may be intended to emphasize that this Jesus is not part of the mainstream gospel tradition on which the subsequent text draws and which it seeks to interpret, supplement, or occasionally to subvert. It is thus hidden in the sense of being unfamiliar or unknown to the mainstream public gospel tradition, but also—indeed already in the Gospel of Thomas—in the sense of conveying something deliberately concealed because it is intrinsically difficult to understand and requires insider knowledge for its explanation. Compare Thomas 1: “And he [Jesus? Thomas?] said: the one who finds the interpretation of these words will not taste death.” In texts like Thomas that manifest gnosticizing sympathies, then, the terminology of apokryphos may thus be deliberately adopted and endorsed to convey difficult saving knowledge for the spiritual elite, while attaching to this a claim (implicit or indeed explicit) of antiquity and authority. Both Hippolytus and especially Irenaeus take it for granted that what sets Valentinians and other gnostic “heretics” apart from catholic Christianity is that they emphasize the secrecy of their writings and their meetings (e.g., Hippolytus, Refutation of All Heresies 1, preface; 6.1, 4, 36, 37; 9.10; 10.8; Irenaeus, Against Heresies 1, preface; 3.2–4 and passim).
Jesus’ private teaching of his disciples is as such, of course, a theme already familiar from the Gospel of Mark (most famously in 4:12); but the difference here is that the secret is evidently something to be publicly disclosed after the resurrection and publicly intended for all who respond in faith (contrast Mark 9:9; 14:9; cf. 16:15; Matt. 10:27).
In this sense, it turns out that the contrast between “apocryphal” gospels and mainstream sacred Scripture is at one level quite in keeping with the self-definition of such texts. Writings like Thomas that stake an explicit claim to secrecy appear deliberately to position themselves in competition with liturgically public, nonapocryphal Jesus tradition. Indeed their claimed superiority actually presupposes the prior givenness of nonsecret, accessible writings that already carry some sort of public authoritative character. And leaving aside the contested question of whether isolated sayings in (say) Thomas might indeed convey earlier strands of Jesus tradition, overall this self-identification acknowledges its own distance from the antecedent gospel literature.
In this sense, as H. Förster (2013, 144–45) also points out, while the New Testament gospels were indeed authoritative texts that in time “became canonical,” it is rather more problematic to assume that other gospel-like texts, after originally occupying the same ground as the Four, had then necessarily “become” apocryphal only by an act of formal exclusion. This is sometimes claimed (e.g., by Lührmann and Schlarb 2000; cf. also the general argument of Watson 2013, e.g., 606). While such a scenario is of course conceivable and worth considering for any given case, in most of the more prominent instances (including Thomas, Mary, and Judas) a more accurate description would envisage texts that remained apocryphal, in their quite self-conscious and deliberate competition with those that became canonical.
It is perhaps also important that many other extracanonical texts do not claim to be in any sense hidden or secret rivals to the protocanonical gospels. Certainly this is true of second-century writings counted among the Apostolic Fathers or the Apologists, but it also pertains to certain infancy gospels, Jewish Christian gospels, and the gospel harmony known as the Diatessaron.
It is an interesting, if slippery, corroboration of this point that none of the New Testament writings one might deem pseudonymous makes any analogous claim to preserve a secret, alternative connection to Jesus or to the church’s apostolic origins. All these texts claim on the contrary to stand within the same public, “catholic and apostolic” tradition.
Importantly, moreover, there are no known gospel texts, either extant or otherwise attested, that became apocryphal after having once been widely normative or authoritative—let alone canonical. The Synoptic Sayings Source Q, if it existed, could potentially be an exception; but quite apart from any intrinsic questions about Q it matters that no assertion of normativity for this hypothetical document can be found in antiquity (see further discussion on p. 89 below). Another exception to prove the rule is the widely popular Diatessaron, a harmony of the protocanonical gospels that eventually gave way to the discrete canonical Four. But no known individual gospel, inside or outside the canon, began in a normative ecclesial mainstream and then became unequivocally apocryphal.
This observation tends to lend further credence to the impression that the apocryphal gospels are instead—often indirectly and in part—epiphenomenal on the gospel tradition that became canonical. In other words, contrary to the impression frequently conveyed in some popular media (and occasionally reinforced by scholarly constructs), the fourfold gospel is not the endpoint of centuries of complete uncertainty about which of many gospels might be normative. Even while debate continued about certain minor epistles and Revelation, no such sustained doubt ever affected Matthew, Mark, Luke, or John either in the early manuscript tradition or in preserved second-century discussions like those of Serapion, Irenaeus, Justin, and Papias. We have previously cited in this connection Theophilus of Antioch, the Epistle of the Apostles, and other second-century sources. But it is worth adding here the much-debated but probably late-second-century list of authoritative writings known as the Muratorian Canon, which survives only in a fragmentary translation into a barbaric pidgin Latin. This puzzling document, too, nevertheless references precisely four gospels: the text begins in midsentence with a conclusion on Mark before introducing Luke as the third gospel and John as the fourth. (Its certainty about the four gospels seems particularly telling in view of the omission of New Testament documents like Hebrews, 1–2 Peter, and James along with the apparent inclusion of the Book of Wisdom, the Apocalypse of Peter, and possibly the Shepherd of Hermas.)
On the subject of terminology, common usage does not necessarily distinguish between “noncanonical” and “apocryphal” gospels. The second of these terms is more widely known and often preferred, and is therefore retained in the title of this book for pragmatic reasons; its continuation has similarly been advocated by Ehrman and Pleše (2011, vii) and Markschies (1998; 2012g, 18–21).
That said, there are reasons to be cautious and circumspect about this terminology and to note the advantages of the more neutral term “noncanonical.” First, while “apocryphal” in its Greek etymology derives from the word apokryphon, meaning simply “hidden,” in time it acquired a more ambivalent significance. “Apocryphal” writings might then be hidden either in the sense of communicating a knowledge reserved only for the few, or else perhaps as concealing secretive or conspiratorial knowledge as opposed to (and perhaps subversive of) the received public teaching of the church and its Scriptures.
Both of these meanings might function either as a positive claim on the part of a book’s advocates or as a negative value judgment about a text whose opponents deem it to be damaging, defamatory, and perhaps wholly fictitious. Neither of these overtones seems appropriate to the noncanonical gospels. Many are not obviously “hidden” at all, whether in an elitist or a subversive sense. Conversely, as we saw earlier, a handful of these texts actually claim hiddenness or secrecy for themselves as a positive quality—beginning arguably with the incipit of the Gospel of Thomas. So the term “apocryphal” might seem to skew the discussion, one way or another, from the start.
As applied to books often called “apocryphal” by Protestants and “deuterocanonical” by Roman Catholics in relation to the OldTestament, the significance of the terminology is rather different, and indeed almost equivalent to “noncanonical.” In the narrowest sense, the “apocryphal” label is here a Protestant designation implying that books contained in the Greek (Septuagint) or Latin (Vulgate) but not in the Hebrew Old Testament carry no canonical authority. In this case, their “apocryphal” status pertains not to any secrecy but to a perceived lack of authority—though the deuterocanonical books are in fact liturgically used in Roman Catholic and (with slight variations) Orthodox Churches, as well as by some Anglicans.
By contrast, none of the so-called apocryphal gospels appear in the canon of any major Christian tradition (though see below on the Epistle of the Apostles, p. 215).
Given the lack of any agreed-upon inventory, even the frequently encountered definite article “The Apocryphal Gospels” turns out to be misleading. It seems to imply the existence of an identifiable set of texts. But in reality it may be impossible to create a definitive list. Not only are we confronted with numerous fragments that may or may not be part of larger documents, but the textual traditions themselves often appear to be highly unstable and volatile compounds of which each new manuscript, and each new translation, may in fact be the creation of a new or secondary apocryphon. The so-called infancy gospels are a particularly accessible case in point (see Voicu 2011, 408–11), but various “gnostic” and “Jewish Christian” traditions present comparable challenges.
This point about the volatility of our texts becomes even more problematic if we recognize that comparable documents continued to be produced throughout antiquity into the Middle Ages and beyond. Does any and every literary retelling of the life of Jesus qualify, even once the New Testament’s canonical boundaries are firmly drawn?
For this reason, too, neither forty nor some other figure can offer a definitive or exhaustive total count of “the” apocryphal gospels.
Another problem of taxonomy in inventories like the above is that so much of what we have is either highly fragmentary or indeed known to us only from brief citations or allusions in other ancient texts. In many cases at Oxyrhynchus, in patristic citations, or elsewhere, the surviving fragments do not amount to enough material to give us confidence about the shape of the original document. For any given text, are we dealing perhaps simply with a single narrative episode or set of sayings, or even with a longer but still strictly limited set of passion or resurrection material? Or must we think of it as either a part or the whole of an integral composition, whether a continuous narrative account of the ministry of Jesus or a complete collection of sayings? What if the fragments we have are in fact all there ever was? Since in many instances the extant fragment may be the only one of its kind, these supposedly “fragmentary” gospels often leave us with more questions than answers.
Among the relevant examples are several Jewish Christian gospels, which are mentioned numerous times in antiquity but are known only from brief snippets of text. Some of these papyri and patristic citations are de facto fragmentary in their extant state. But in many cases it is impossible to be certain about the overall nature and shape of the document concerned—for example, whether it was only a sayings collection or also a narrative, and whether that narrative covered the entire ministry of Jesus or only one aspect of it. Were some of them perhaps only ever intended to supplement the retelling of an existing (canonical or harmonized) gospel outline rather than to constitute part of an entire alternative account? As we will see, there are reasons to think this may have been the case for the so-called Jewish Christian gospels. Perhaps the best-known example of such an “intruded” episode is the story of the adulterous woman that eventually attached itself to the end of John 7 (7:53–8:11) in the third century, perhaps to answer the Pharisees’ challenge of 7:15 (but in some manuscripts it follows 7:36; 7:44; or Luke 21:38; see further Keith 2009). That said, the actual number of such “intrusions” in the textual tradition of the gospels remains remarkably small.
Or there are the discourse collections—including short fragments, more extensive texts like the gospels of Thomas or Philip, and, for that matter, the sayings source Q, if it ever existed. And there are episodic narrative texts, including somewhat more extensive ones like the infancy gospels of James and Thomas and passion accounts like the gospels of Peter and (much later) Nicodemus.
But it concentrates the mind to consider that we do not have a single surviving alternative ancient narrative account of Jesus’ ministry ranging from the baptism of Jesus to his death and resurrection. Is it possible that this narrative structure, which is held in common by the four canonical gospels, was by the noncanonical writers either ignored or else (and more typically) used as a rough-and-ready scaffold in which to insert supplementary or substituted episodes of “rewritten gospel” (somewhat in analogy to the “rewritten Bible” or “rewritten Scripture” technique attested in the Dead Sea Scrolls and other ancient Jewish and Christian texts)? We will need to revisit these questions (see, e.g., p. 87).
A related question concerns the so-called agrapha or “unwritten” sayings of Jesus, which surface in a wide range of early Christian writings, in certain variant New Testament manuscripts (e.g., Codex Bezae, or D in the standard text-critical designation), and even in the mainstream text of the New Testament itself (esp. at Acts 20:35, where Luke has Paul quote a saying of Jesus that appears nowhere in his gospel).
The problematic nature of agrapha as originally conceived has become increasingly clear in a number of respects. Most obvious is the paradox that this material survives by definition in literary sources and is therefore not unwritten in any meaningful sense. Further, the distinction between agrapha and fragmentary gospels is sometimes a matter of degree rather than of kind: some isolated sayings might in theory be excerpted from larger sayings collections (like the Gospel of Thomas) while, conversely, some of the collections may be anthologies of such individual sayings. There is also the less compelling objection to any definition of agrapha based on a distinction between canonical and noncanonical sayings, which Ehrman and Pleše (2011, 351) in a somewhat curious dichotomy deem “a decision that involves theological rather than historical judgments.”
Relevant inventories therefore vary considerably, but involve at a minimum a number of New Testament sayings not recorded in the gospels (e.g., Acts 20:35, “It is more blessed to give than to receive”) and undesignated variants in the textual tradition (e.g., in Codex D: Mark 9:49, “Every sacrifice will be salted with salt,” and Luke 6:4, where Jesus says to a man working on the Sabbath, “Man, if you know what you are doing, you are blessed; but if you do not know, you are cursed and a transgressor of the law”). Among the many other sayings, the most notable include several in 2 Clement, such as a brief dialogue with Peter in which Jesus encourages the disciples to be as sheep even in the midst of ravenous wolves (5.2–4; cf. P.Oxy. 4009 below, p. 148) as well as an enigmatic statement that the kingdom will come “when the two will be one and the outside like the inside, and the male with the female will be neither male nor female” (12.2–6, also echoed in Gospel of Thomas 22; also 2 Clem. 4.5).
Clement of Alexandria also cites a number of logia (“sayings”; singular: logion) like “Ask for the big things, and the small things will be given to you as well” (Stromateis 1.24.158) and “My mystery is for me and the children of my house” (5.10.63, also in Pseudo-Clementine Homilies 19.19–20, and several other writers); Origen quotes “Be clever bankers” (Commentary on John, 19.7.2).
Some of these isolated sayings could in theory feature among the oldest parts of the Jesus tradition. But the majority seem more clearly the derivative product of reported speech, paraphrase, expansion, or quotation from memory. They are sometimes placed on the lips of Jesus in edifying or homiletical settings or indeed as devotionally experienced, rather than necessarily intended as recording authentic pre-Easter sayings.
Illustrations of this point abound from the earliest to the latest examples in Christian antiquity. The memory of what “he said to me” in prayer is already part of the apostle Paul’s spirituality: 2 Corinthians 12:8–9. In the fifth or sixth century, the recently published P.Monts.Roca IV 59 (inv. no. 996), in an apparently homiletical or meditative context, includes a saying of the Lord that might be rendered as “it has been kept to pronounce sweet words” (see Torallas Tovar and Worp 2014, 164–67).
It is significant that, as Jens Schröter has demonstrated, the gospel tradition’s “recollection” of Jesus is from the start articulated within an integrated complex of sayings and narrative. Individual agrapha and short dialogue units are in that connection much more likely to “emerge secondarily to the already existing gospels” (Schröter 2013b, 130–32, 262). This is an important corrective to formerly widespread views that imagined the gospels growing instead “from bare sayings” to full narrative lives (so again recently Hägg 2012, 148–86 and blurb).
Recent scholarship has tended to restrict the number of texts being considered under this heading—in particular by omitting clearly derivative or misattributed material, sayings of the pre- or post-incarnate Jesus, and non-Christian sources.
The most recent German edition (Markschies and Schröter 2012) for the first time considers these sayings under the three categories of (1) noncanonical sayings of the (earthly) Jesus (Hofius 2012; on this somewhat minimalist reading just seven early sayings are not derivative); (2) Jesus logia from Nag Hammadi texts other than apocryphal gospels (Plisch 2012b: five units from Testimony of Truth [NHC IX, 3] and Interpretation of Knowledge [NHC XI, 1]); and (3) sayings of Jesus in Arabic and Islamic literature (Eissler 2012: ten logia from the Qur’an and thirty-one traditions from post-Qur’anic sources of the seventh and later centuries).
While Otfried Hofius (2012; following his teacher Jeremias ; cf. Hofius 1991) focuses on potentially “authentic” sayings of the “historical Jesus,” recent scholarship has been more open to the development of such material from the perspective of the reception history of both canonical and noncanonical gospel traditions. As Elliott (1993, 26) rightly notes in critique of Hofius’s minimalism, notions of “authenticity” and “originality” are not normally thought pertinent to the selection of apocryphal gospels. That said, Elliott himself opts for a relatively sparse account (26–30), while Ehrman and Pleše (2011, 351–67) argue for an economical but more “representative” identification of agrapha as “sayings allegedly spoken by the historical Jesus that are recorded in documents other than the surviving gospels (canonical or non-canonical).” (Contrast further the more expansive collections of Morrice 1997 and Stroker 1989.)
Another category problem is presented by documents whose existence is asserted by either ancient or modern authors, but which may be either fictive or hypothetical. Most famous among the latter is the so-called Q source of sayings material shared by Matthew and Luke, which is still widely affirmed in Synoptic Gospel scholarship. We will return to this in a separate section (p. 89).
Several other hypothetical “lost” gospels have been postulated from time to time. Among these are first editions of the existing canonical gospels (Proto-Mark, Proto-Luke, etc.), as well as miscellaneous other narrative or sayings sources adduced to explain certain parts of the gospel tradition. John Dominic Crossan, for example, proposes a text called the “Cross Gospel” that he believes to underlie the canonical passion narratives as well as the noncanonical Gospel of Peter (Crossan 1988, 1991, 2007; cf. p. 137 below), while Francis Watson (2013) more recently has substituted a source he calls SC (Sayings Collection), which in his view predates the canonical gospels and is most conservatively preserved in the Gospel of Thomas.
None of these other hypothetical sources have been found compelling by a significant number of scholars, and not one is a document whose real existence, let alone whose ancient identity as a gospel, has been demonstrated.
Even so, this is not to deny that some gospels were indeed lost in antiquity: the list above from the Gelasian Decree contains a number of plausible examples, and Markschies (2012c) discusses others (cf. above, table 2, p. 27), several of which our discussion references in passing. All in all, however, we have no reason to think that lost gospels constitute a literary category in their own right. Rather than being unique, such texts are far more likely to belong to one or another familiar genre. Even many of the apocryphal gospels we will discuss survive only in a single copy and were “lost” until they were rediscovered.
How then should we structure the available evidence, using what we know to be an inevitably partial inventory of (so-called or self-styled) gospels, along with numerous other ancient texts and fragments that present gospel-like narrative or teaching about Jesus of Nazareth?
A quick comparison of the tables of contents in recent editions and textbooks illustrates the reality of the problem by showing a variety of often incompatible organizing principles. Some, like Christopher Tuckett, Hans-Josef Klauck, Paul Foster, and Christoph Markschies, recognize the difficulty of consistent classification and try to address it by variously mixing historical, literary, and linguistic categories with narrative or even geographic ones (e.g., for Nag Hammadi). This unfortunately leaves us with an unwieldy hybrid. Markschies and Schröter (2012) have compiled the most comprehensive collection but are left with an even more impenetrable taxonomy, the rationale for which is inadequately elaborated in an otherwise magisterial and wide-ranging introduction.
A number of serviceable schemes exist, including some that sort the texts by theological orientation and others by literary or geographic criteria. In the end, a degree of eclecticism is probably inevitable in choosing which particular documents to focus on. Here I will adopt a fourfold taxonomy patterned loosely on elements of the New Testament gospels’ narrative typology, developing insights drawn from Christopher Tuckett and Paul Foster. Tuckett (2005, 243–48) suggests four categories:
As Tuckett points out, the fragmentary nature of much of our material inevitably leaves our judgments about these categories provisional, and certain texts arguably belong to more than one category. But perhaps this is not the best we can do. There is a certain awkwardness in juxtaposing essentially literary categories of narrative and sayings genres with essentially narrative ones relating to the infancy and resurrection of Jesus. In terms of literary genre, Tuckett’s third category could arguably fold into the first, and the fourth into the second. Some editors in fact simply distinguish two headings, sayings and narratives (e.g., Cameron 1982, 7).
Paul Foster (2009) produces similar categories but introduces an additional twist:
Once again there is a certain clash of literary and biographical categories. Clustering the Nag Hammadi texts together under a separate heading of their own (an archaeological or perhaps geographical one) has a certain neatness about it, and at least in this case contextualizes like with like. And yet, as Foster acknowledges, these texts are vastly different from one another in form and substance, so it seems better not to prejudge the question of whether in origin, intent, or function they belong together. In any case the remainder of Foster’s outline seems tethered to a biographical sequence, especially when one realizes that the last genre is almost exclusively cast in a post-resurrection setting.
Genre distinctions are quite often somewhat crude and arbitrary. On the other hand, the attempt to offer finer distinctions very quickly turns unwieldy. Compare, for example, Hans-Josef Klauck’s taxonomy of no fewer than twelve categories, deployed in his introduction to the apocryphal gospels (2003):
Large collections like that of Markschies and Schröter (2012) resort to even more complicated and cumbersome lists (above, table 2, p. 27; cf. previously Schneemelcher 1991–92, vol. 1). Complex mixtures of literary and biographical categories are deployed by Elliott (1993) and Robert Miller (1994), while Ehrman’s 2003 book gives no account of order or taxonomy at all (though he remedies this in Ehrman and Pleše 2011; 2014).
We clearly need to start somewhere. Tuckett’s attractive breakdown of narrative, sayings, infancy, and resurrection quite reasonably points in a direction that is largely guided by the explicit content and setting of these documents themselves, without unduly prejudging questions of context, literary criticism, or interpretation.
Further along this line, I suggest, the most promising and least prejudicial taxonomy of these documents is therefore quite reasonably narratival rather than literary-analytical. In other words, it seems advisable to map the extracanonical sources onto the basic structure of the New Testament’s narrative gospels, in relation to which they most often position themselves. Among recent writers this decision is, for example, similarly adopted with minor variations by Ehrman and Pleše (2011), who follow the threefold division of (1) infancy gospels, (2) ministry gospels, and (3) passion, resurrection, and post-resurrection gospels.
This is also the approach we will take in the present volume. It seems best to structure the documents under four broadly biographical headings:
Similar or related finds from an important geographic location (e.g., Nag Hammadi) will also fit this scheme without undue difficulty—sometimes because interest in one or another of these headings predominates. Fragmentary gospels on papyrus, too, tend to be identifiable along these lines. The scheme arguably remains serviceable even where the writers apparently do not know, or do not accept, a biographical account of Jesus’ life: the focus in such cases is often on one aspect of the Jesus tradition that implicitly functions in a way that in other sources forms part of a narrative whole—be it the instruction of the risen Jesus or the parable-like sayings of his earthly teaching.
Given the lively interest and controversy the noncanonical gospels have generated since at least the nineteenth century, it is puzzling that for quite a long time these texts nevertheless remained relatively difficult for the general public to access. Until the late twentieth century, English translations were often partial, expensive, and not always up to date (e.g., James 1924; the two editions of Hennecke, Schneemelcher, and Wilson 1963; Schneemelcher 1991–92). The mid-1990s produced a couple of more serviceable but still partial translations, Elliott 1993 as well as Robert Miller 1994. Elliott 2006 additionally contained a useful “synopsis” specifically of the infancy gospels. At the time of writing the most accessible and reasonably priced collection and translation of noncanonical gospels, in both paper and e-book versions, is Ehrman and Pleše 2014. (This is also available in a useful four-language edition with facing pages in Greek, Latin, or Coptic [Ehrman and Pleše 2011]. The 2014 edition updates their translation of the Gospel of Judas and additionally includes the Unknown Berlin Gospel [Gospel of the Savior] and the Discourse on the Cross [the Stauros Text from Qasr El-Wizz].)
For the gospel-like texts from Nag Hammadi, several editions still advise readers to consult the exhaustive but expensive and unwieldy five-volume Coptic Gnostic Library (Robinson 2000). A more recent and handier, if still costly, bilingual Coptic-German edition is Nagel 2014, which contains specifically gospels and Acts material from Nag Hammadi. The most accessible English translation is that of Meyer 2007.
The most complete translation, with an outstanding monograph-length critical overview by Christoph Markschies as well as individual introductions to the texts, is at present available in German only (Markschies and Schröter 2012). Its translation into English seems highly desirable but also liable to prove a complex undertaking in view not only of the volume of material but also of the often contested and rapidly shifting lines of scholarly debate.
In what follows I will take as my base texts Ehrman and Pleše 2011/2014 wherever possible, as theirs is a widely available set of translations. Students of the original languages are strongly encouraged to consult the 2011 multilingual volume. Since the available space does not permit us to do full justice to all thirty-seven of this edition’s “gospels,” we will foreground the most important and supplement this with brief treatments of texts not included by Ehrman and Pleše, particularly a number of dialogue gospels from Nag Hammadi (drawing on the translation of Meyer 2007). For each document, I offer a brief historical introduction and survey of the content, concluding with an analysis in relation to some or all of the five interpretive emphases outlined above (see pp. 28–31).
Ehrman, Bart D., and Zlatko Pleše, eds. 2011. The Apocryphal Gospels: Texts and Translations. New York: Oxford University Press.
———, eds. 2014. The Other Gospels: Accounts of Jesus from Outside the New Testament. New York: Oxford University Press.
Meyer, Marvin W., ed. 2007. The Nag Hammadi Scriptures: The International Edition. New York: HarperOne.
Foster, Paul. 2009. The Apocryphal Gospels: A Very Short Introduction. Oxford: Oxford University Press.
Hengel, Martin. 2000. The Four Gospels and the One Gospel of Jesus Christ: An Investigation of the Collection and Origin of the Canonical Gospels. Trans. J. Bowden. London: SCM Press.
Hurtado, Larry W. 2006. The Earliest Christian Artifacts: Manuscripts and Christian Origins. Grand Rapids: Eerdmans.
———. 2013. “Christian Literary Texts in Manuscripts of Second & Third Centuries.” http://larryhurtado.files.wordpress.com/2010/07/second-third-century-christian-texts1.pdf.
Markschies, Christoph. 2015. Christian Theology and Its Institutions in the Early Roman Empire: Prolegomena to a History of Early Christian Theology. Translated by W. Coppins. Baylor-Mohr Siebeck Studies in Early Christianity. Waco, TX: Baylor University Press.
Nicklas, Tobias. 2011. “‘Apokryph Gewordene Schriften’? Gedanken zum Apokryphenbegriff bei Grosskirchlichen Autoren und in einigen ‘Gnostischen’ Texten.” In “In Search of Truth”: Augustine, Manichaeism and Other Gnosticism; Studies for Johannes van Oort at Sixty, edited by J. A. van der Berg et al., 547–65. Leiden: Brill.
Tuckett, Christopher. 2005. “Forty Other Gospels.” In The Written Gospel, edited by M. Bockmuehl and D. A. Hagner, 238–53. Cambridge: Cambridge University Press.
Watson, Francis. 2016. The Fourfold Gospel: A Theological Reading of the New Testament Portraits of Jesus. Grand Rapids: Baker Academic.