When Were the Gospels Written, and Which Ones Remained Outside the Canon?

Gospel and canon: two different questions

In the first and second centuries the word “gospel” did not always denote the same literary format. Mark, Matthew, Luke, and John are narratives of Jesus’ activity, death, and resurrection; the Gospel of Thomas is chiefly a sayings collection; other “gospels” focus only on infancy, passion, or post-resurrection revelation. The title “gospel” therefore does not by itself make a text a direct competitor to the four canonical narrative works.

“Canon” is also a later technical category. Early Christians more often spoke of writings as accepted, disputed, useful, or rejected. Oxford’s survey of Christian apocrypha notes that “canonical” became a clear classificatory term only in the fourth century, even though particular collections—especially the four Gospels and Pauline letters—had become normative much earlier.

When were the four canonical Gospels written?

Mark: roughly the mid-60s to shortly after 70 CE Among contemporary scholars Mark most commonly has the status of the earliest canonical Gospel. Oxford Bibliographies summarizes the consensus as composition in the mid-60s or shortly after 70 CE; many locate it near the Jewish-Roman war and the destruction of the temple in 70, although no precise year can be proved.

Markan priority is not only a matter of dating: many scholars hold that Matthew and Luke used Mark literarily. This pushes Matthew and Luke later than Mark, but the Synoptic Problem is not mathematically solved and minority alternatives remain.

Matthew: roughly the 80s–90s Matthew is commonly dated around 85–90 CE and is widely understood as a revised and expanded use of Mark together with additional traditions. Its final form therefore reflects both memory of Jesus and the concerns of communities several decades later.

The traditional name “Matthew” associates the Gospel with the apostle and tax collector. The narrative itself, however, does not identify its author by name. Historians therefore distinguish between a text that is internally anonymous and early ecclesial tradition about its authorship.

Luke: roughly 85–95 CE, with some later proposals Oxford Bibliographies summarizes a common dating of Luke to 85–95 CE, while noting that some scholars argue for earlier dates or for a date into the second century. Luke and Acts were written by the same author, who does not identify himself by personal name in either work.

Late-second-century tradition identifies the author with Luke, a companion of Paul. That tradition is historically important, but it must be distinguished from what the text itself directly establishes. Dating Luke is therefore not a publication date but a range shaped by judgments about its relationship to Mark, Josephus, Paul, and early church history.

John: late first century or early second century John is the hardest canonical Gospel to compress into a single year. Traditional and many modern reconstructions place it around 90–100 CE, while contemporary debate allows composition or final redaction into the early second century. A 2026 Oxford overview accordingly gives the four Gospels a broad horizon between the fall of Jerusalem in 70 and the beginning of the Bar Kokhba revolt in 132.

John differs sharply from the Synoptics in chronology, discourse style, and Christology. Scholars disagree over whether it emerged from one authorial act or several editorial stages. John therefore especially illustrates why approximate dating should not be turned into a falsely precise date.

How do the four canonical Gospels complement one another?

If we use the currently dominant model of Markan priority as a working starting point, we can compare which material appears only in Matthew and Luke and which material is distinctive to each Gospel. The table must not be read as proof of a simple chain “Mark → Matthew → Luke → John”: Matthew and Luke are both probably later than Mark, but their relative order and direct literary dependence remain debated. John is even more independent and should not be treated merely as the final supplement to the Synoptics.

The most important “middle layer” is material Matthew and Luke share that Mark lacks. In the classic Two-Source Hypothesis, much of this is assigned to the hypothetical source Q or to a shared tradition; Q is not an extant manuscript and remains a scholarly reconstruction rather than a material artifact.

The table also shows why “supplementing” needs qualification. A later evangelist does not simply add material: Markan episodes can be omitted, moved, shortened, expanded or theologically reshaped. John is even less like the final chapter of the same book—its chronology, discourses and selection of events are different enough that its relationship to the Synoptics must be treated separately.

If Mark is used as the earliest Synoptic baseline, the safest summary is: Matthew and Luke preserve much of Mark’s narrative axis but each expands and reshapes it with additional shared and distinctive material; John represents a much more independent literary and traditional shaping of Jesus’ story.

Authorship and manuscripts: what do we actually know?

Did Matthew, Mark, Luke, and John actually write the Gospels? The safest formulation is two-sided. The narrative bodies of the works do not provide conventional authorial self-identification, so much critical scholarship treats them as formally anonymous writings that later received traditional attributions. Papias, preserved by Eusebius, already connects Mark with Peter and speaks of Matthew’s “oracles” in a Hebrew language in the early second century; Irenaeus around 180 explicitly attributes the four Gospels to Matthew, Mark, Luke, and John.

Yet even here there is no total consensus. Simon Gathercole, for example, has argued that the attributions or titles may have been very early or even original. The article therefore avoids the slogan “the names were simply invented.” A more accurate boundary is: the attributions are early and strikingly stable, but their historical identification with first-century figures is not equally demonstrable in all four cases.

A manuscript is not the same thing as the date of composition No autograph of any canonical Gospel survives. The earliest manuscript fragments are copies, often appreciably later than the supposed composition. The famous Rylands fragment of John, P52, is a tiny part of an early copy, and its palaeographic dating is not precise enough to settle the Gospel’s date on its own.

Larger codices are more important for the history of collections. P75, copied in the early third century, contains Luke and John together; the Vatican Library describes it as very early material evidence for their sequence within a Gospel collection. P45, from the third century, contained all four Gospels and Acts. Manuscripts therefore demonstrate circulation and collection, not the day a lost original was composed.

Gospels outside the canon: which ones and how early?

Which other gospels circulated? The list is long, and the texts are not all of the same kind. Famous examples include the Gospel of Thomas, Gospel of Peter, Gospel of Mary, Gospel of Judas, Infancy Gospel of Thomas, Protevangelium of James, and fragmentarily preserved Jewish-Christian gospels such as the Gospels of the Hebrews, Nazarenes, and Ebionites. Marcion also used a gospel closely related to Luke.

It is therefore misleading to place them all in one folder called “secret gospels.” Thomas is a collection of 114 sayings; Peter is chiefly a passion-resurrection narrative; Mary preserves dialogues and revelation; infancy gospels fill gaps in stories of Jesus’ childhood. Their historical value is real but heterogeneous.

Thomas, Peter, Mary, and Judas: early, but not equally early The Gospel of Thomas is the most debated case. It survives complete in Coptic at Nag Hammadi and in three older Greek fragments from Oxyrhynchus. Its final date and relationship to the Synoptics remain disputed; individual traditions may be early, but there is no consensus that the complete extant Thomas predates the canonical Gospels.

Most scholars place the Gospel of Peter in the second half of the second century. The Gospels of Mary and Judas likewise witness to the diversity of second-century Christianity; the surviving Judas codex dates to the third or fourth century, while a Greek form was already known to Irenaeus. These distinctions matter: the age of a manuscript, the age of a translation, and the age of an original composition are different questions.

Why four — and were the others simply banned?

Why four? The four-Gospel collection is older than the final New Testament Around 180 CE, Irenaeus not only uses Matthew, Mark, Luke, and John but explicitly defends the claim that the Gospels are four. This is the clearest early defense of a fourfold Gospel. Yet the collection did not suddenly begin with him: Tatian’s Diatessaron, other second-century readers, and scholarship on the Epistula Apostolorum suggest that these four had already become an important set decades earlier, probably by the mid-second century at the latest.

This means the story is not “Nicaea in 325 chose the four Gospels.” A four-Gospel collection was widely established before the Council of Nicaea. Nicaea primarily addressed a Christological controversy, not a vote over which Gospels belonged in the Bible.

Did the Church simply ban the other gospels? Ecclesiastical authority and conflicts over orthodoxy certainly shaped which writings gained normative status and which were labelled disputed or heretical. In the fourth century Eusebius explicitly distinguishes accepted, disputed, and rejected books and mentions Gospels of Peter, Thomas, and others. Athanasius in 367 warns even more sharply against apocryphal books.

But the model of a single decree followed by systematic destruction of every alternative is too simple. Oxford’s survey notes that non-canonical works continued to circulate, be read, and be copied; some were supplementary narratives rather than direct competitors for public liturgical scripture. A text could disappear through limited use, language shift, material accident, theological marginalization, active rejection, or several of these together.

When was the New Testament canon actually closed?

Here we must distinguish the four-Gospel core from the entire New Testament. The four Gospels were strongly established by the second century. Other books later included in the New Testament—Hebrews, James, 2 Peter, 2–3 John, Jude, and Revelation—remained disputed in some regions for much longer.

Athanasius’ 39th Festal Letter of 367 is the first surviving text to list exactly the modern 27-book New Testament. It is an important milestone, not a magical moment when the Bible suddenly came into existence. Regional variation and different liturgical collections continued afterward.

What can we reasonably conclude?

Well supported: Mark is probably the earliest canonical Gospel; Matthew and Luke are later; John is usually placed in the late first or early second century; multiple other gospel traditions existed in the second century; and a four-Gospel collection is visible well before the final stabilization of the 27-book New Testament.

Not well supported: that all four canonical Gospels are direct eyewitness transcripts; that all non-canonical gospels are older or historically more authentic; that the Council of Nicaea selected the four Gospels; or that exclusion from the canon by itself proves a text contained truth authorities wanted to hide.

The history of the canon really is a history of selection, authority, use, memory, and theological conflict. Precisely for that reason it should be reconstructed from dates, manuscripts, and citations rather than from an inverted dogma—whether ecclesiastical or anti-ecclesiastical.