Matthew's Fulfillment Formula: Creative Writing or Historical Reporting?

There is a phrase that echoes through the Gospel of Matthew like a drumbeat, a literary refrain so distinctive that scholars have given it a technical name: the fulfillment formula. "That it might be fulfilled which was spoken through the prophet..." (World English Bible). Matthew uses this formula about ten times — some scholars count as many as fourteen, depending on which borderline cases they include — each time connecting an event in Jesus' life to a passage from the Hebrew Bible. The theological message is unmistakable — Jesus is the culmination of Israel's scriptures, the one to whom the prophets pointed. But here is the thing: when you actually go back and read the Old Testament passages Matthew is quoting, something uncomfortable happens. The original contexts frequently have nothing to do with a future messiah. In several cases, Matthew appears to be ripping verses out of their historical and literary settings, bending them into shapes they were never meant to take. In at least one case, the verse he cites does not exist at all.

This is not a fringe observation. It is mainstream biblical scholarship, and Catholic and evangelical commentators supply several of the admissions below. So let's walk through the most egregious examples of Matthew's fulfillment formula and ask the question that the evidence demands: is this historical reporting, or is it creative theological writing?

The Formula Itself

Krister Stendahl, in his landmark 1954 work The School of St. Matthew and Its Use of the Old Testament, argued that Matthew's fulfillment quotations represent a distinctive scribal practice. Stendahl observed that these quotations differ textually from both the Masoretic Text and the Septuagint, suggesting that they come from a tradition of independent translation and interpretation — what Stendahl called a Matthean "school" analogous to the pesher interpretation found at Qumran. The Dead Sea Scrolls community also read their scriptures as coded predictions about their own time and community. Matthew, Stendahl argued, is doing the same thing: reading backward from events to find scriptural warrants for them. (The "school" itself has not worn well — Robert Gundry contested the textual analysis behind it, and Luz, as we are about to see, doubts it — but it is the comparison with Qumran's way of reading scripture, not the school, that this essay relies on.)

Ulrich Luz, in his multi-volume commentary on Matthew, is sceptical of the "school". The evidence, he writes, "does not fit the picture of a scribe who would deliberately have produced only in his formula quotations a new form of the text," and he thinks most of the quotations reached Matthew already attached, in oral tradition, to the stories in which we now find them (Luz, Matthew 1-7, 159-60). But he also reports a "broad consensus that the fulfillment formula comes from the evangelist" (157). And nearly all the formula quotations sit in the narrative framework — they are editorial comments by the author, not words placed on Jesus' lips. This is a crucial distinction. Matthew is not claiming that Jesus said "I am fulfilling Hosea 11:1." Matthew is telling the reader, as narrator, that this event fulfilled prophecy. The theological interpretation is Matthew's own contribution. And as we will see, that interpretation is frequently strained to the breaking point.

"Out of Egypt I Called My Son" — Hosea 11:1

Matthew 2:15 tells us that Joseph took the infant Jesus to Egypt to escape Herod, and that the family remained there until Herod's death, "that it might be fulfilled which was spoken by the Lord through the prophet, saying, 'Out of Egypt I called my son.'" The quote comes from Hosea 11:1. Now, open Hosea 11 and read the verse in context. Here is what it actually says:

"When Israel was a child, I loved him, and out of Egypt I called My son." (Berean Standard Bible)

This is not a prophecy. It is not even future-tense. Hosea is looking backward, recalling the Exodus — God delivering the nation of Israel from slavery in Egypt. The "son" is Israel, collectively, as a people. The very next verse, Hosea 11:2, says "the more I called Israel, the farther they departed from Me. They sacrificed to the Baals" (BSB, which here follows the Greek's "I called"; the Hebrew, as in the WEB, reads "They called to them" — Israel's apostasy is the point either way). Unless one wants to argue that Jesus went on to sacrifice to Baals after returning from Egypt, it is blindingly obvious that Hosea is talking about Israel's history, not predicting a future messianic infant's travel itinerary.

Raymond Brown, in The Birth of the Messiah, is characteristically thorough on this point. Brown notes that Matthew is employing a typological reading — Jesus recapitulates Israel's history. Just as Israel was God's "son" who came out of Egypt, so Jesus, God's Son, comes out of Egypt. Brown acknowledges this is creative theological interpretation, not the original meaning of Hosea. We have no evidence that the passage was read messianically in pre-Christian Judaism — none. No Jewish interpreter before Matthew is known to have read Hosea 11:1 as a prediction about a future individual.

There is also a small but telling textual detail here. Matthew's wording — "Out of Egypt I called my son," singular — agrees with the Hebrew Masoretic Text against the Septuagint, which reads the plural "out of Egypt I called his children." Had Matthew simply been working from the standard Greek Bible of the Diaspora, the singular "son" that makes the verse usable as a one-to-one type of Jesus would not have been sitting there waiting for him. We will see the same thing with Matthew's Jeremiah quotation below: his scriptural text is not mechanically copied from any single version. He selects, adapts, and shapes his wording — and the version he reaches for here is precisely the one that yields the individual "son" his typology requires.

And here is what makes this even more interesting: the trip to Egypt only appears in Matthew. Luke's birth narrative has the family going from Bethlehem to Jerusalem for the purification rites and then straight back to Nazareth. There is no Egypt detour in Luke. The parsimonious explanation is that the Egypt sojourn is theological narrative rather than history — a Moses-and-Israel story told of the infant Jesus. Whether Hosea generated it is less clear. Matthew hangs the verse on the flight into Egypt — the family "departed into Egypt, and was there until the death of Herod" (2:14-15) — not on the return, which he narrates only later (2:21). That looks like a citation attached to a story that already existed, and Brown argues exactly that: the formula citations were added to a pre-Matthean narrative, which would read more smoothly without them. Either way, the "prophecy" is found in the story after the fact, not predicted before it.

"He Will Be Called a Nazarene" — The Verse That Does Not Exist

Matthew 2:23 is, in some ways, the most revealing of all the fulfillment quotations. After the return from Egypt, Matthew tells us the family settled in Nazareth: "So was fulfilled what was spoken through the prophets: 'He will be called a Nazarene.'" (BSB; the Greek hoti here can introduce either a quotation or a report, and the BSB reads it as a quotation where the WEB reads a report.)

Go ahead. Search the entire Old Testament. You will not find this quotation. It does not exist. No prophet says "he will be called a Nazarene." Scholars have spilled enormous amounts of ink trying to figure out what Matthew could possibly be referring to. The proposals are creative, I will give them that. Some suggest a wordplay on the Hebrew netser (branch) from Isaiah 11:1 — "A shoot will come out of the stock of Jesse, and a branch out of his roots will bear fruit." Others point to the Nazirite vow (Numbers 6), though Nazirite and Nazarene are etymologically distinct. Still others note Matthew's plural "prophets" and suggest he is summarizing a general prophetic theme rather than quoting a specific text.

Brown surveys these options at length in The Birth of the Messiah, and he does not conclude that the verse is unfindable. Matthew, he holds, "is citing known Scripture and has two definite prophetic passages in mind": Isaiah 4:3, where the remnant in Jerusalem "shall be called holy," and the Nazirite Samson of Judges — though Brown concedes that "This style of exegesis is complicated" (Brown 1993, 223-25, as quoted by August 2018). Davies and Allison likewise put the primary weight on Isaiah 4:3, with perhaps a secondary allusion to Isaiah 11:1. Luz reverses the emphasis: Isaiah 11:1 "clearly stands in the foreground," and Matthew wrote the vague plural "prophets" because he was "unable to identify the quotation handed down to him" (Luz, Matthew 1-7, 150). R.T. France, and more recently Jared August, read the verse not as a citation at all but as the flagging of a theme. August's evidence is the formula itself: of Matthew's fourteen fulfillment formulas, the ten that cite a specific passage all name a single "prophet" and add "saying," while 2:23 shares its plural "prophets" and its missing "saying" with formulas such as 26:56 — "that the Scriptures of the prophets might be fulfilled" — which cite no passage at all (August 2018, 70-71). Brown, writing earlier, rejected the non-citation reading because hoti normally introduces a direct citation (Brown 1993, 208), though Luz allows that it "could, as in 26:54, introduce an indirect quotation" (Luz, Matthew 1-7, 150); this essay follows Brown in quoting the verse from the BSB, which renders it as speech.

Whatever Matthew is doing, he is not quoting a verse: no text says this. The most defensible readings — a pun on netser and the Nazirite, or a summary of a prophetic theme — still have Matthew building a "prophecy" out of wordplay on a town no prophet ever mentions. By modern standards of citation that would be unacceptable. In a first-century Jewish literary context, it tells us something important about how Matthew understood the relationship between scripture and narrative — not as a matter of strict citation, but as a fluid, interpretive enterprise.

Rachel Weeping for Her Children — Jeremiah 31:15

Matthew 2:17-18 connects the massacre of the innocents — Herod's alleged slaughter of male infants in Bethlehem — to Jeremiah 31:15. The source verse reads, in the World English Bible: "A voice is heard in Ramah, lamentation and bitter weeping, Rachel weeping for her children. She refuses to be comforted for her children, because they are no more." (Matthew's Greek of this verse, at 2:18, follows the Septuagint in some words and the Hebrew in others: his nouns for weeping and mourning are the Septuagint's, but where the Septuagint's Rachel will not stop weeping, Matthew's, like the Hebrew's, refuses to be comforted. It is the same selectivity we saw with Hosea's "son.")

Two problems here, and both are significant. First, the Jeremiah passage. In its original context, Rachel's weeping is about the Babylonian exile. Jeremiah 31 is a chapter of consolation — the weeping is followed immediately by God's promise to bring the exiles home. "There is hope for your latter end," says Yahweh. "Your children will come again to their own territory" (Jer. 31:17). The trajectory of the passage is from sorrow to restoration. Matthew seizes the sorrow and ignores the restoration. He takes a verse about the exile and repurposes it as a prophecy about infanticide. The fit is, to put it mildly, forced.

Second, and this matters historically: most critical scholars doubt the massacre of the innocents actually happened. There is no independent corroboration — nothing in Josephus, who extensively documented Herod's brutalities and would have had every reason to include such an atrocity. The only other ancient mention, an anecdote in the Latin writer Macrobius in the early fifth century (Saturnalia 2.4.11), comes more than three centuries after Matthew and very likely depends on his story. Brown discusses this at length and concludes that the massacre narrative likely derives from the Moses typology — just as Pharaoh ordered the killing of Hebrew infants and Moses escaped, so Herod orders the killing of infants and Jesus escapes. Matthew is constructing a parallel between Jesus and Moses, and the Jeremiah quotation adds scriptural texture to a narratively constructed event.

This is not cynical debunking. This is simply reading the texts carefully. When an event has no independent attestation, follows a well-known literary template, and is connected to a scriptural citation that does not match its original context, the most parsimonious explanation is that we are reading theology rather than reportage — a story told on the Moses template, with a scripture laid over it.

The Wrong Prophet — Matthew 27:9-10

Matthew 27:9-10 is perhaps the most puzzling of Matthew's fulfillment quotations. After Judas returns the thirty pieces of silver and the priests use them to buy the potter's field, Matthew writes: "Then that which was spoken through Jeremiah the prophet was fulfilled, saying, 'They took the thirty pieces of silver, the price of him upon whom a price had been set, whom some of the children of Israel priced, and they gave them for the potter's field, as the Lord commanded me.'"

The problem is simple. This quotation does not come from Jeremiah. It is a loose paraphrase of Zechariah 11:12-13. Matthew attributes it to the wrong prophet. Ancient scribes noticed this — some manuscripts actually change "Jeremiah" to "Zechariah" or simply to "the prophet" without a name. Augustine knew copies that omit the prophet's name and set them aside, because the majority and the more ancient Greek copies read "Jeremiah." He thought the slip was Matthew's own — Jeremiah's name came to mind in place of Zechariah's — but held that the Spirit let it stand, because all the prophets speak by one Spirit (Harmony of the Gospels 3.7.29-30). Jerome reported that a member of the Nazarene sect had shown him the words in an apocryphal Jeremiah, yet still judged it more likely that Matthew had taken them from Zechariah (Commentary on Matthew, on 27:9-10). The seventeenth-century Hebraist John Lightfoot argued that "Jeremiah" names the whole volume of the prophets, "because he was placed the first in that rank" in the Jewish ordering of the books (Horae Hebraicae, on Luke 24:44; cf. on Matt 27:9). Modern commentators have proposed further solutions, to which we now turn.

Critical scholars divide here, and Maarten Menken, who has studied the quotation more closely than most, notes that on its questions "no consensus has emerged so far" (Menken 2002, 310). Luz is blunt: Matthew named the wrong prophet, "incorrectly naming Jeremiah instead of Zechariah," and Luz infers that "no copy of the Minor Prophets was in the library of the Matthean community" (Luz, Matthew 1-7, 157). Menken argues the opposite: the naming is deliberate. The quotation is built on Zechariah 11:13, but its "field" comes from Jeremiah's purchase of a field (Jer 32:6-15), and "the ascription of the quotation to Jeremiah serves to draw attention to the source" of that word (Menken 2002, 319). Many commentators hear Jeremiah's potter (Jer 18-19) as well. Mark does something similar at Mark 1:2, crediting "Isaiah the prophet" (BSB) for a blend of Malachi and Isaiah; the WEB, following the later Byzantine text, reads "in the prophets" there, the same kind of scribal correction we have just seen at Matthew 27:9. On Menken's reading nothing has been misidentified. But the "prophecy" is still an assemblage no reader could have found in either prophet, and Menken himself finds that the name "the potter's field" was already part of the story and was worked into the quotation with some "exegetical violence" (323, 326). The narrative shaped the quotation far more than the quotation shaped the narrative.

The Two-Donkey Problem — Matthew 21:4-5

And now we arrive at one of the most illustrative examples of Matthew's method. In the triumphal entry narrative, Zechariah 9:9 is the driving text: "Rejoice greatly, O Daughter of Zion! Shout in triumph, O Daughter of Jerusalem! See, your King comes to you, righteous and victorious, humble and riding on a donkey, on a colt, the foal of a donkey" (BSB).

This is standard Hebrew synonymous parallelism. "A donkey" and "a colt, the foal of a donkey" refer to the same animal described two ways. This is how Hebrew poetry works — the second line restates the first with variation. Every student of biblical Hebrew learns this in their first semester. (Matthew's own version at 21:5 also opens with a line from Isaiah 62:11 — "Say to the Daughter of Zion" (BSB, used here to match the BSB of Zechariah 9:9 quoted above) — and quietly drops Zechariah's "righteous and victorious": the same selecting and shaping we have seen throughout.)

Mark has one animal, a colt (Mark 11:2). Luke follows Mark, and John, who does quote Zechariah (John 12:14-15), also has one. But Matthew — and this is where it gets remarkable — has Jesus send for two animals: "a donkey tied, and a colt with her" (Matt. 21:2). Matthew then writes "he sat on them" (Matt. 21:7) — most naturally, on both animals, though many interpreters since Origen have taken "them" to be the garments (Luz, Matthew 21-28, 8, as quoted by Coppins 2012).

So what is Matthew doing with two animals? Here the scholarship genuinely divides, and honesty requires laying both sides out. My own reading — and I will defend it — is that Matthew read the parallelism over-literally, manufacturing a second animal from a single poetic line and producing the awkward image of a man riding two beasts at once. That reading has a respectable pedigree: S. Vernon McCasland, in his pointed 1961 article "Matthew Twists the Scriptures," charged Matthew with turning Mark's dignified scene into "something like a circus spectacle" by failing to grasp the poetic structure, and John P. Meier once entertained the idea that such a misreading might even hint at a Gentile author unfamiliar with Hebrew verse. But this is contested, and the opposing camp is substantial. Robert Gundry, David Instone-Brewer, and John Nolland argue that Matthew knew perfectly well that Zechariah meant one animal and deliberately read the two clauses as two animals — a hyper-literal exegetical move that, as Nolland notes, also turns up at Qumran, and — he argues, though Instone-Brewer disputes the rabbinic evidence — among the rabbis, some of whom likewise found two animals in Zechariah 9:9. It is important not to misrepresent Davies and Allison here: in their ICC commentary they do not treat the second animal as a clumsy mistake. They read it as a knowing, sophisticated piece of exegesis — observing that Matthew has actually increased the parallelism of Mark's account (21:7), and floating a Moses typology (the LXX of Exodus 4:19-20 has Moses' family travel on "asses," plural). For Davies and Allison, Matthew is a deft reader rewarding those who know their scripture, not an evangelist tripped up by Hebrew poetry. Their view belongs squarely in the "deliberate" camp, and I will not conscript it for the "misreading" thesis it actually opposes.

A third group reverses the direction altogether. Stendahl, Barnabas Lindars and France hold that Matthew knew, or inferred, that two animals were present, and found that Zechariah's wording made room for the second. Stendahl thought "the only reason for such a treatment of the O.T. text must be that Matthew knew a tradition that spoke about two asses"; Lindars reasoned that an unridden foal would not yet have been separated from its mother; France has Matthew aware that two animals were there and pleased that the oracle's wording "lent itself to the mention of the other" (Stendahl 1968, 200; Lindars 1961, 114; France 2007, 778-79; all as set out by Coppins 2012).

There is, as Wayne Coppins puts it at the outset of his study, "no consensus" on why Matthew has two animals (Coppins 2012, 277). But notice what the first two camps share, whether the reading was clumsy or clever: the scriptural text is doing the driving. Menken, in passing, reads the passage the same way — Matthew introduces the she-donkey "on account of the quotation" (Menken 2002, 310). Only the third camp denies it, and it has to explain why Mark, Luke and John know of only one animal. Coppins, for his part, does not think France, Stendahl and Lindars are correct, and neither do I. On the first two readings, Matthew is so committed to demonstrating precise scriptural correspondence that he will reshape the narrative to track the wording of the prophetic text — adding a second animal that exists in his story only because the wording of Zechariah seemed to call for it. The events are being molded to the citation, not the other way around.

What This Means

Let me be clear about what the evidence shows. Matthew is not reporting events and then discovering, to his amazement, that they happen to match ancient prophecies. Matthew is doing something quite different. He is working with Israel's scriptures and Israel's stories, and shaping his narrative — or the traditions he inherited — to create correspondences with them. Sometimes a scriptural story is the template: Moses' endangered infancy stands behind the flight to Egypt and the massacre of the infants. Sometimes the wording of a verse bends the story: two donkeys appear because Zechariah seemed to name two. Sometimes a story already in hand is fitted with a text that has been bent to match it, as with the potter's field. And when no verse says what he needs, a prophetic theme is compressed into words no prophet wrote and credited to "the prophets."

This is creative theological writing. It is midrashic storytelling in the best tradition of Jewish scriptural interpretation. It may even be, in its own terms, a sophisticated and profound theological claim about the relationship between Jesus and Israel's story. But it is not historical reporting. And it is certainly not evidence that ancient prophecies were supernaturally fulfilled.

Call that my opinion if you like. But I would challenge anyone to read Hosea 11:1 in context and tell me with a straight face that the prophet was predicting a first-century infant's trip to Egypt. I would challenge anyone to produce the Old Testament verse that says "he will be called a Nazarene." I would challenge anyone to find, in Zechariah or in Jeremiah, the prophecy Matthew says was fulfilled.

And there it is. Matthew's fulfillment formula is not evidence for prophecy fulfilled. It is evidence for a first-century author reading scripture the way his contemporaries at Qumran read scripture — as a coded text about his own community, to be decoded through creative interpretation. The difference is that scholars read the Qumran pesharim as evidence of a community's conviction that the prophets spoke of its own day — evidence of how the sect read, not of what the prophets foresaw. Matthew's are still, by millions of people, treated as straightforward historical proof of supernatural prediction. The double standard is breathtaking.

The choice, then, is not really between two options but three. Matthew may have been faithfully recording events that miraculously corresponded to prophecies — prophecies that, in their original context, are not prophecies and in one case do not even exist. Or he may have inherited stories already told in the shape of Moses and Israel and hung scriptures on them, as Brown argues for the infancy narrative and Menken for the potter's field. Or he may have composed more of the correspondences himself, as I have argued for the second donkey. Critical scholars divide between the second and third options, and passage by passage the answer differs. But the first requires multiple miracles plus prophecies that are not prophecies. The other two require only a first-century Jewish writer doing what first-century Jewish writers did. I know which side of that line I find more parsimonious.