Somewhere in the transmission of the Jesus tradition, two incompatible claims about access to God ended up inside the same canon. One tradition, preserved independently in Luke and echoed in Matthew’s own inherited material, condemns the very act of holding a key that controls who gets in. The other, found nowhere but in a single passage of Matthew, has Jesus hand that same kind of key to one man. I want to trace both halves of this problem separately, because they call for different kinds of evidence. The first is a question about redaction: did Matthew inherit the Petrine keys scene, or did he build it? The second is a question about reception: does any strand of early Christian writing outside the canonical four know this scene at all, and if the Gospel of Thomas is any guide, what does its silence tell us?

Part One: Matthew’s Insertion

Mark 8:27-30 gives the fullest early account of the scene at Caesarea Philippi. Jesus asks who people say he is, the disciples report the rumors, Peter answers that he is the Christ, and Jesus orders them to tell no one. That is the whole exchange. Luke 9:18-21 follows this almost exactly, same location implied, same question, same confession, same charge to silence, and without keys, without a rock, without anything resembling an institutional commission.

Matthew reproduces this scene closely, matching Mark’s wording and sequence to a degree that leaves his dependence on Mark beyond dispute, and then, at the precise moment where Mark and Luke move on to the passion prediction, Matthew stops and inserts three verses found nowhere else in the tradition: Peter as rock, the church built on him, the gates of Hades unable to prevail, the keys of the kingdom, and the power to bind and loose. Then Matthew rejoins Mark’s sequence, moving into the same passion prediction that follows in Mark and Luke. The insertion rests inside a scene otherwise lifted almost verbatim from a source Matthew is visibly following elsewhere, which is what makes it identifiable as an insertion rather than as material Matthew received along with the rest of the episode.

What follows six verses later makes the insertion look even more like a seam than a unity. The same Peter who has just been named the rock on which the church will stand is called Satan and a stumbling block for objecting to the passion prediction, using language Matthew elsewhere reserves for real adversaries of Jesus’s mission. A single author composing one coherent scene about Peter’s authority does not usually write his way into that kind of reversal within the space of six verses. It reads instead like two separate memories of Peter, one exalting him and one rebuking him, joined by an editor who did not smooth the transition between them.

The community setting behind Matthew’s Gospel supplies a plausible reason for wanting the exalting half. Matthew is generally dated to the 80s of the first century, written for a community still working out what it meant to be Jewish and Christian after the Temple’s destruction in 70 CE, in the same general period when rabbinic Judaism was reorganizing itself around Yavneh and Christian communities were separating from the synagogue as distinct bodies needing their own procedures for governance and discipline. Matthew is the only Gospel that uses the word ekklesia, and he uses it exactly twice: once here, at 16:18, founding the church on Peter, and again at 18:17, in a passage that walks a congregation through how to handle a member who refuses correction, repeating the same binding-and-loosing formula, this time given to the gathered community rather than to Peter alone. A Galilean teacher speaking of “my church” years before any such institution existed is difficult to read as anything other than a later community’s authority structure written back into Jesus’s mouth. The borrowing from Isaiah 22:22, where Eliakim receives the key of the house of David as a hereditary court office, reinforces the same reading: Matthew reaches for an image of institutional stewardship at exactly the point where his community needed one, and Mark and Luke, working from sources that did not carry this concern, never received it.

Reading Matthew this way asks for nothing more dramatic than a working evangelist with Mark’s confession scene in hand, plus whatever sayings material about withheld and interior access to the kingdom reached him through Q or his own special source, building an institutional layer on top of that material which his fellow evangelists, drawing on the same or adjacent traditions, simply did not include. The most economical explanation for a passage attested nowhere outside one Gospel, resting at a visible seam in that Gospel’s own source, echoing that same Gospel’s own institutional vocabulary elsewhere, and undercut six verses later by a rebuke, is that it is Matthew’s addition rather than Jesus’s saying.

The case for that reading gets support not just from the seam itself but from what the surrounding sayings tradition, preserved independently in Luke and inherited by Matthew in his own Sermon on the Mount, says about how a person actually reaches the kingdom in the first place. Luke frames the charge against the legal experts in the vocabulary of entry:

“Woe to you lawyers! You have taken away the key of knowledge. You did not enter yourselves, and you hindered those who were entering” (Luke 11:52).

The verb “enter,” eiserchomai, shows up twice in that single verse, and the sin Luke names is blocking an act every person is otherwise expected to perform without anyone standing in for them. Matthew’s own Sermon on the Mount carries that same logic without ever reaching for the vocabulary of keys:

“But when you pray, go into your room and shut the door and pray to your Father who is in secret. And your Father who sees in secret will reward you” (Matthew 6:6).

Prayer here runs directly between the person and the Father, without a priest handling the transaction or an office standing witness to it. “But seek first the kingdom of God and his righteousness, and all these things will be added to you” (Matthew 6:33) puts the weight on seeking, something the individual does, rather than on receiving something handed down through an office. Luke rounds out the picture from the other Gospel: “the kingdom of God is in the midst of you” (Luke 17:21), a phrase that could just as legitimately be rendered “within you,” since the underlying Greek, entos hymon, really does support either reading, and either way puts the kingdom somewhere no key was ever going to reach, because it was never behind a door in the first place.

Read against that material, Matthew 16:19 isn’t just a passage lacking outside attestation; it runs directly against something the rest of the tradition, Matthew’s own Sermon on the Mount included, spends real energy establishing. “I will give you the keys of the kingdom of heaven, and whatever you bind on earth shall be bound in heaven, and whatever you loose on earth shall be loosed in heaven” reintroduces the very arrangement, one figure controlling access on everyone else’s behalf, that Luke 11:52 names as a monopoly and that Matthew 6:6 makes unnecessary by clearing every intermediary out of prayer entirely. That’s the deeper reason I read the passage as an insertion rather than one saying among equals. Its problem isn’t only that nobody else attests it; it works against the theological grammar of entering, seeking, and shutting the door that Matthew and Luke otherwise share, reaching for the same imagery of doors, rooms, and keys and arriving at the opposite conclusion from everywhere else that imagery shows up.

Part Two: What Thomas Does Not Have

If Matthew 16:17-19 is a later construction, we should expect its absence from traditions that developed independently of Matthew’s own community, and its presence, in some form, in whatever traditions preserve the material Matthew was working against. The Gospel of Thomas offers a useful test on both counts, with the caveat stated at the outset rather than left for a reader to discover: Thomas is usually dated to the late first or the second century, later than any canonical Gospel, and its relationship to the synoptics remains disputed among scholars who study it. I am treating it here as historical evidence for what an early reading community preserved and omitted, not as an eyewitness source equal in weight to Mark.

Logion 39 gives the clearest parallel to the anti-gatekeeping tradition: “The Pharisees and the scribes have taken the keys of knowledge and hidden them. They have not entered, nor have they allowed those who wish to enter to do so.”[1] This is closer to Luke 11:52 than Matthew’s own parallel at 23:13 is, since Thomas keeps both the key and the language of knowledge together, where Matthew drops the key image entirely and speaks only of shutting the kingdom. Logion 3 gives a parallel to the interiorizing strand: the kingdom is located inside the hearer and outside the hearer at once, and the saying warns against those who would claim it as a locatable, external thing. That tracks closely with Luke 17:21’s placement of the kingdom within rather than in observable form.

Across all one hundred fourteen sayings, nothing in Thomas parallels Matthew’s Petrine commission. Peter, or anyone else, is never given a set of keys to the kingdom. The image of a church built on a rock, secure against the gates of Hades, is missing entirely, along with any mention of an individual granted binding and loosing power. If anything, Thomas is cool toward Peter in particular: logion 13 has Peter offer a comparison of Jesus to a righteous angel, a comparison the text frames as falling short of the deeper answer Thomas himself gives, and this suspicion of Peter’s authority appears again, more sharply, in the Gospel of Mary, where Peter resists the teaching Mary Magdalene received directly from Jesus. Thomas goes further still in logion 12, where the disciples ask who will lead them once Jesus is gone, and the answer given is James the Just rather than Peter. That saying does not touch the gatekeeping question directly, since it concerns who succeeds Jesus rather than whether succession itself should work through a single controlling office. But it confirms that whatever tradition stands behind Thomas had no investment in Peter as the founding office-holder Matthew makes him, and looked toward a different figure entirely for what came after.

The weight this absence can bear depends on how Thomas relates to the synoptic tradition, and scholars remain divided on the point. Some, including Stevan Davies[2] and April DeConick[3], have argued for a core of material in Thomas that developed independently of Mark, Matthew, and Luke, drawing on early sayings collections that predate the narrative Gospels. If that is correct, then Thomas becomes a second and independent witness: a tradition that preserved the anti-gatekeeping and interiorizing sayings while never once encountering, or never once transmitting, anything resembling Matthew’s institutional scene. That is exactly the asymmetry the source-critical case in Part One would predict, since it mirrors the pattern already visible inside the canonical Gospels themselves, where the anti-gatekeeping material reaches both Matthew and Luke through more than one line of transmission while the Petrine keys material reaches no one but Matthew.

Other scholars think Thomas knew Luke’s sayings tradition, or Q, directly, and reworked it for its own purposes. Under that reading, Thomas’s silence proves less on its own, since a text built from Luke’s material would naturally lack anything unique to Matthew regardless of any argument being made about it. What that reading still gives us is reception rather than independent attestation: a community handling this material in the second century preserved and sharpened the sayings that condemn gatekeeping, and never picked up the passage that installs a gatekeeper, even though nothing about working from Luke or Q would have prevented them from inventing such a scene themselves if the idea of an authorized successor had struck them as something to pass along. The silence, read this way, tells us that the Petrine commission was separable from that shared material, and recognizable as belonging to one particular stream, so that a community reworking Luke’s or Q’s sayings felt no pull toward inventing anything like it.

Either reading supports the same conclusion from a different angle. Thomas is not organized around questions of church office or succession, and that indifference should be weighed on its own terms rather than treated as if it settles the matter by itself. The stronger case rests on the positive material alongside the silence: logion 39 and logion 3 track the Lukan tradition against Matthew’s, logion 13 shows a community unimpressed by Peter’s grasp of who Jesus is, and across the whole collection nothing answers to the scene at Caesarea Philippi that Matthew alone preserves. Taken together with the internal seam visible in Matthew’s own text and the institutional pressures facing his community after 70 CE, the pattern points toward a single conclusion: the Jesus who condemns the monopolizing of the key of knowledge and the Jesus who hands the keys of the kingdom to one man belong to two different layers of tradition, and only one of them has any claim to reach back before the Gospels that record it.

There’s a real irony once you lay the evidence out this way, and I don’t think it should pass without comment. The saying the church kept, canonized, and used to found its highest office survives nowhere outside the one Gospel written for the community that needed it, while the saying condemning that same kind of office survives independently in Luke, in whatever source stands behind Luke’s version, and in a Gospel the church eventually excluded from its own canon and spent centuries trying to suppress. The passage with the weaker evidentiary base became the foundation stone of the institution, and its rival, carrying the broader and more independent support, ended up pushed to the margins by the very body that saying had warned against in the first place. An institution consolidating its authority after the Temple’s fall had every reason to keep a saying that grounds a single office, and comparatively little reason to preserve one that names such an office as the sin of the scribes, which is more or less what happened: the one Jesus saying with no attestation anywhere else in the tradition became the passage Matthew made foundational, while the passage questioning whether any single figure should hold the key at all outlived its own exclusion anyway, surfacing independently in Luke, in whatever lies behind Luke’s source, and in a Gospel of Thomas that names not Peter but James as the one to turn to once Jesus was gone.

---

[1] Gospel of Thomas quotations (logia 3, 12, 13, and 39) follow Thomas O. Lambdin’s translation, in James M. Robinson, ed., The Nag Hammadi Library in English, rev. ed. (San Francisco: HarperSanFrancisco, 1990). The critical Coptic edition is Bentley Layton, ed., Nag Hammadi Codex II,2-7, Nag Hammadi Studies 20 (Leiden: Brill, 1989).

[2] Stevan L. Davies, The Gospel of Thomas and Christian Wisdom (New York: Seabury Press, 1983); Stevan Davies, “Thomas: The Fourth Synoptic Gospel,” The Biblical Archaeologist 46, no. 1 (1983).

[3] April D. DeConick, Recovering the Original Gospel of Thomas: A History of the Gospel and Its Growth, Library of New Testament Studies 286 (London: T&T Clark, 2005); April D. DeConick, The Original Gospel of Thomas in Translation: With a Commentary and New English Translation of the Complete Gospel, Library of New Testament Studies 287 (London: T&T Clark, 2006).