Undesigned Coincidences in the Gospels: Do the Accounts Quietly Confirm Each Other?

On a hillside by the Sea of Galilee, a vast crowd has followed Jesus into open country, and there is nothing to eat. According to John’s Gospel, Jesus turns to one disciple in particular: “Lifting up his eyes, then, and seeing that a large crowd was coming toward him, Jesus said to Philip, ‘Where are we to buy bread, so that these people may eat?’”¹ It is a small thing, easy to read past. But it raises a question John never answers. Of the twelve, why Philip? He is not the group’s spokesman; that is usually Peter. He is not the treasurer; that is Judas. Why ask him where to find bread?

John does not say. But Luke, telling the same story, drops a geographical note that John omits: the feeding took place near the town of Bethsaida.² And John, far earlier and for an entirely unrelated reason — he is listing who first followed Jesus — mentions in passing that “Philip was from Bethsaida, the city of Andrew and Peter.”³ Lay the three details side by side and the puzzle dissolves. Jesus asked Philip because they were near Philip’s hometown, and Philip would know the local bakeries and markets. No single Gospel makes that point. Neither writer seems aware he is solving the other’s riddle. The detail in one casually, accidentally, answers a question left dangling in another.

That is what defenders call an undesigned coincidence, and the claim built on dozens of such cases is the subject of this essay: that the Gospels and Acts interlock in ways that look exactly like independent reports of real events — and nothing like a single fabricated story or a tidy literary harmonization.

Stating the Argument Carefully

An undesigned coincidence, in the technical sense, is a particular shape of fit between two texts. One account mentions a detail that, on its own, is puzzling or incomplete; a second account, written for unrelated purposes and showing no awareness of the first’s difficulty, contains an offhand detail that resolves it. The fit is too specific to be likely chance, yet too casual and unflagged to look planted. The classic statement of the inference is straightforward: when accounts dovetail this way without contrivance, the most economical explanation is that both writers were drawing, from different angles, on knowledge of the same real events.

It helps to state the reasoning formally. Premise 1: when independent accounts interlock at incidental, unintended points — one casually answering a question another leaves open — the best explanation is that both report real events that the writers actually knew. Premise 2: the Gospels and Acts (and Acts with Paul’s letters) contain many such interlocks. Conclusion: therefore these documents are best explained as well-informed, substantially independent testimony to real events, not as late legend or coordinated invention.

Two features of that structure must be marked at the outset, because everything honest about the argument depends on them. First, it is an inference to the best explanation, not a deductive proof. No premise guarantees its conclusion; the cases raise the probability of reliability, they do not establish it with certainty. Second, the force is cumulative. No single coincidence is meant to carry weight; any one of them might be chance or might have a mundane literary explanation. The argument lives or dies on the pattern — many cases, of the right shape, pointing the same direction. As we will see, the contemporary defender Lydia McGrew frames this explicitly in the language of probability: each coincidence is a piece of evidence whose likelihood is far higher if the sources are well-informed and independent than if they are not.⁴

How the Argument Developed

The argument has a distinguished and slightly forgotten pedigree. Its first systematic exponent was William Paley — better remembered today for the watchmaker design argument of his Natural Theology — in an earlier work, Horae Paulinae (1790), whose full title announces its method: Horae Paulinae; or, The Truth of the Scripture History of St. Paul Evinced, by a Comparison of the Epistles Which Bear His Name with the Acts of the Apostles, and with One Another.⁵ Paley worked entirely within the Pauline corpus and Acts. He noticed that Paul’s letters and the narrative of Acts repeatedly confirm one another at small, oblique points that no forger would have thought to coordinate — a travel plan mentioned in a letter that quietly matches a movement narrated in Acts, a name dropped in one that explains a relationship in the other. His conclusion was that the letters and Acts were neither forged from each other nor invented, but were genuine documents independently reflecting the same history.

It was the Cambridge divine John James Blunt who, half a century later, coined the phrase and extended the method across the whole Bible. His Undesigned Coincidences in the Writings Both of the Old and New Testament (an early form in 1833; the expanded edition in 1847) catalogued case after case, including an appendix matching the Gospels and Acts against the Jewish historian Josephus.⁶ For Blunt these incidental agreements were “an argument of their veracity” — internal marks that the writers told the truth.

The argument then went quiet for over a century, neglected as biblical scholarship turned to source and form criticism. Its modern revival is largely the work of the philosophers Timothy and Lydia McGrew, and above all Lydia McGrew’s Hidden in Plain View: Undesigned Coincidences in the Gospels and Acts (2017), which restated the case for a general readership while sharpening it with explicit probabilistic reasoning and direct engagement with the objection from literary dependence.⁷ It is the indispensable contemporary source, and the one a serious reader should consult firsthand.

Premise 1: Why Unintended Interlocking Points Toward Truth

The first premise is, in fact, the part skeptics tend to grant, because it is simply how we evaluate testimony everywhere else. Detectives, historians, and courts routinely treat accounts that agree at incidental, uncoordinated points as more credible than accounts that agree only on the headline. The reason is probabilistic. If two witnesses to a car accident independently mention, without prompting, that the streetlight was out — one because it explains why she did not see the other car, the other because he tripped on the curb in the dark — the convergence is striking precisely because neither was trying to corroborate the other. Coordinated liars produce agreement on the main story; what they rarely produce is a web of small, mutually explanatory details that none of them flagged.

The deeper point is about what fabrication tends to look like. A single author building a narrative, or a community polishing a legend, works from one storyline and smooths the seams; loose ends get tied off or dropped. What is hard to manufacture is a detail in document A that quietly answers a question raised — and left unanswered — in document B, with neither author appearing to notice. That requires either genuine knowledge of the underlying facts or a feat of cross-document coordination far more elaborate than the simpler hypothesis that the events happened and the writers knew about them. So the probability of the interlocking pattern is high if the reports are well-informed and substantially independent, and low if they are invented or derived from a single tidy source. That likelihood gap is the whole engine of the argument.

Premise 2: Are There Really Many Such Cases?

The strength of Premise 2 is in the examples, and the honest way to present them is to show the range — including which kinds are strong and which are weaker.

The Philip-and-Bethsaida case with which we began is among the cleanest, because the explaining detail (Philip’s hometown) sits in a completely different part of John, planted for an unrelated narrative reason chapters earlier, while the geographical anchor (Bethsaida) comes from a different Gospel altogether. A second often-cited case is the “green grass” of the same feeding. Mark alone notes that Jesus had the crowd sit down “on the green grass.”⁸ Around the Sea of Galilee the hills are brown and parched for most of the year; the grass is green only briefly, after the winter rains, in spring. And John alone, in his account of the same event, mentions that “the Passover, the feast of the Jews, was at hand” — and Passover falls in spring.⁹ Mark says nothing about the season; John says nothing about the grass; together they cohere. (As we will see, this particular case is also where the literary-dependence objection bites hardest, and it should be held more loosely than the Philip case for that reason.)

A third cluster concerns Herod’s court, and it shows the argument operating across the Gospels and into Acts. Luke records an odd episode in which some Pharisees warn Jesus that Herod Antipas wants to kill him, and Jesus replies, “Go and tell that fox…”¹⁰ How would Jesus’ circle have such specific intelligence about the tetrarch’s intentions and court? Luke himself supplies the thread elsewhere, again incidentally: among the women supporting Jesus was “Joanna, the wife of Chuza, Herod’s household manager.”¹¹ And the book of Acts, listing the leaders of the church at Antioch, names “Manaen, who had been brought up with Herod the tetrarch.”¹² The movement had inside connections to Herod’s household — which is exactly why it could report what went on there. None of the three texts points out the link; each was written for its own purpose. (Blunt’s classic version of this cluster runs through Matthew’s remark that Herod spoke about Jesus “to his servants” in Matthew 14:2, paired with Luke’s Joanna.¹³)

And then there is Paley’s original domain, which many defenders regard as the strongest of all: the Acts–Epistles coincidences. The traditional reasoning is that the interlocks there cannot be explained as one document tidying another, because the genuine letters were certainly not written with Acts in view, and — on what was for most of the twentieth century the majority view — the author of Acts did not write with Paul’s letters open before him. It is partly the cumulative force of such agreements that helped secure near-universal scholarly agreement on the authenticity of the major Pauline letters (Romans, 1–2 Corinthians, Galatians).¹⁴ But that “Acts is independent of the letters” premise can no longer be presented as settled. A significant revisionist line — Richard Pervo’s Dating Acts (2006) and the work of Joseph Tyson and the Acts Seminar — dates Acts to the early second century and argues that its author knew and mined the Pauline letters as a source.¹⁴ This matters enormously here, because it is precisely the hinge on which Carrier rests his objection, and we will need to take it head-on below. Even granting the contest, however, the Acts–Epistles cases remain among the argument’s strongest — not because dependence is impossible, but because, as we will see, the shape of these interlocks is not what literary borrowing predicts.

The Strongest Objections, Stated at Full Strength

This argument has fewer dedicated academic critics than, say, the resurrection debate — partly because it has been under-engaged in mainstream scholarship since its revival. But the objections that exist are serious, and the most forceful comes from the mythicist historian Richard Carrier, whose essay bears the blunt title “There Are No Undesigned Coincidences: The Bible’s Authors Are Simply Changing Up Their Sources.”¹⁵

Carrier’s central charge is that the argument quietly assumes the texts are independent, when the dominant view in New Testament studies is that they are not. On the standard two-source hypothesis, Mark was written first, and Matthew and Luke both used Mark (and a shared sayings source). If that is right, then a “coincidence” between Mark and Luke may be nothing more mysterious than Luke copying and lightly editing his source. Apparent dovetailing, on this account, is an artifact of redaction: a later author adjusting, expanding, or recombining material he inherited, which will naturally produce details that “fit” because they descend from a common text. On this view the whole method commits a kind of anachronism: it was forged in an era when the Gospels were thought to be four independent witnesses, and it has not absorbed what source criticism established.

It is important to state Carrier’s position precisely, because it is sharper than it is often reported. He does not grant that the Acts–Epistles cases are the safe tier and then quarrel only about the Gospels. He argues the reverse. The only coincidences with any chance of being undesigned, he writes, are the “internal” ones — Gospel to Gospel — and even those he attributes to shared sources; the Acts–Epistles correspondences he dismisses outright, on the ground that “the author of Acts obviously used those Epistles as a source text.”¹⁵ In other words, Carrier closes off the very escape route the argument’s defenders have leaned on hardest: if Acts mined the letters, then the Acts–Epistles interlocks are deliberate borrowing dressed up as independent memory, and the strongest tier collapses into redaction along with the rest. He even names the Philip-and-Bethsaida case specifically — reading it not as two independent witnesses but as John harmonizing his narrative to Luke.¹⁵ Stated this way, the objection deserves the last word before any reply: on Carrier’s telling there is no tier the dependence move cannot reach.

A second objection, pressed by various skeptical commentators and statisticians, is about selection bias and the base rate of coincidence. Search any two long, overlapping texts for details in one that “explain” details in the other, and you will find some by sheer chance — the human mind is a relentless pattern-finder. Defenders, the charge runs, parade the hits and never count the misses: they do not tabulate how many puzzles in Gospel A find no answer in Gospel B, nor how often the accounts flatly conflict. Without a denominator, a pile of impressive-looking fits tells us little. This is the “remembering the hits and forgetting the misses” problem, and it is a genuine methodological worry.¹⁶

The sharpest form of the literary objection, and the one a defender must answer last, returns to the green grass. Suppose John did know Mark’s Gospel — a position held by a growing number of scholars. Then John could simply have read Mark’s “green grass,” inferred that the event fell in the rainy spring season, and added his own note that Passover — a spring festival — was near. The “coincidence” would then be no coincidence at all but a chain of inference by a later author from an earlier text: designed, not undesigned.¹⁷ This is the objection at its most uncomfortable, because it does not deny the fit; it offers a fully natural, non-evidential account of how the fit arose.

A Reply — and What It Can and Cannot Carry

Take the literary-dependence objection first, because it is the most important and partly correct. The honest concession is large: if a coincidence holds only between texts in a known dependence relation — especially Mark and a Gospel that used Mark — then redaction is a live alternative explanation, and such cases carry less weight. McGrew herself accepts this discipline; she treats different cases as having different strengths, and even sets aside examples she judges too weak to lean on.¹⁸ A defender who waves away Markan priority is not arguing in good faith. The serious reply is more targeted: the objection does not touch the cases where it cannot reach.

Consider the shape of genuine redaction. When a later author notices a connection in his source, the natural editorial move is to state it — to clarify, to smooth, to make the link explicit for the reader. That is what tidying looks like. But the strong undesigned coincidences do the opposite: the explaining detail sits in a different document, dropped for an unrelated reason, with neither writer flagging the link. A copyist who spotted that Philip was from Bethsaida and that the feeding happened near Bethsaida would have written, “Jesus asked Philip, because Philip was a local.” The fact that no Gospel says this — that the reader has to assemble it — is precisely what redaction does not predict.

This is also the heart of the answer to Carrier on the Acts–Epistles cases, and it must be made on the merits rather than by waving the dependence question away. Grant Carrier his premise for the sake of argument: suppose the author of Acts had read Paul’s letters. The trouble is that Acts does not behave like a document built out of them. It never verbatim-quotes the letters — not once — and at the points where a borrower would most want to lean on his source, Acts diverges from it. The sharpest example is the very one critics press as a contradiction: Galatians has Paul insist that after his conversion he “did not go up to Jerusalem” but went away into Arabia, and recount only two Jerusalem visits, while Acts narrates a prompt visit to Jerusalem and a different sequence of trips.¹⁵ A writer mining Galatians to manufacture verisimilitude would have had reason to smooth exactly this rather than leave his flagship narrative crosswise to his own alleged source; this tells against borrowing. The point should not be overstated, however: a redactor with a theological agenda can diverge from a source deliberately, so the divergences are suggestive rather than decisive. The real weight falls on the shape of the agreements: the points where Acts and the letters do converge have the offhand, question-and-answer character of genuine interlocks — a detail in the letter quietly answering an oddity in Acts, neither author noting the link — which is not what deliberate harmonization produces. Harmonizers connect the dots; these texts leave them unconnected. So the Acts–Epistles material survives Carrier’s objection not because dependence has been ruled out a priori — it has not — but because the pattern fits independent knowledge better than it fits a redactor reworking the letters.

The same answer disposes of Carrier’s reading of Philip/Bethsaida as John harmonizing to Luke. If John were adjusting his story to match Luke’s placement of the feeding near Bethsaida, the place to do it would be in the feeding scene itself, with a cue the reader could follow. Instead the load-bearing detail — that Philip was “from Bethsaida” — sits in John 1, five chapters earlier, inside a disciple-list, planted for the unrelated purpose of introducing Andrew and Peter, with no narrative signal pointing forward to the feeding at all. That is the opposite of how a harmonizing redactor works. A harmonizer signals his link; he does not bury the one detail that makes the connection in a context where no reader would think to look for it.

Two further points limit the damage even on the Gospel-to-Gospel front. First, the strongest cross-Gospel cases (like Philip/Bethsaida) draw the explaining detail from a Gospel not in the relevant source relation, so simple copying does not account for them. Second, the green-grass case, by contrast, should be conceded as weak if John used Mark; it is exactly the kind of case the objection legitimately deflates.

The selection-bias objection deserves real respect and a real concession: defenders have not, on the whole, done the rigorous work of counting misses, and the argument would be stronger if they had. The probabilistic framing helps but does not fully answer it. What can be said is twofold. First, the genuine cases are not mere “fits” but question-and-answer shaped — one text raises a specific oddity (why Philip?) that it never resolves, and another casually resolves it. That asymmetric structure is considerably rarer than a loose thematic resemblance, and harder to generate by chance trawling. Second, the response to base-rate worries is the same as the response to any cumulative case: the claim is not that one fit is decisive but that the number and specificity of well-shaped cases, concentrated in documents purporting to report the same events, is what an honest-and-informed hypothesis predicts and a fabrication hypothesis does not. Still, the candid verdict is that quantifying the pattern remains the argument’s least developed flank.

Now the green-grass objection in its sharpest form should be granted its due before the reply — and here the reply is mostly to retreat, not to fight. If John knew Mark, the green-grass/Passover link may indeed be a designed inference rather than an undesigned coincidence, and it should be dropped from the load-bearing cases. Carrier’s instinct on this kind of example is sound. But notice what the move costs him and what it does not. The green-grass move turns on a writer drawing an easy inference from a detail his source highlights; it has no purchase on cases where the explanatory detail is buried in an unrelated context for an unrelated reason, and — as argued above — the Acts–Epistles cases resist the parallel borrowing story for a separate reason, namely that Acts diverges from the letters rather than smoothing toward them. An author inferring the season from “green grass” is plausible; an author scattering Philip’s hometown into a disciple-list in chapter one so that it will quietly rationalize a question he raises in chapter six — while never connecting the two — is a far less economical hypothesis than that he simply knew Philip was from Bethsaida. The reply, in short, is to surrender the weak cases cleanly and rest the argument on the strong ones — on which the dependence move, even when granted its premises, predicts the wrong shape.

The Honest State of the Scholarship

It must be said plainly that this is not, at present, a front-line debate in academic New Testament studies, and that fact cuts in more than one direction. The argument has enthusiastic defenders — Lydia McGrew foremost, with endorsements from scholars such as Craig Keener and the apologist J. Warner Wallace¹⁹ — and energetic popular critics, of whom Carrier is the most prominent. But it has not generated a large peer-reviewed literature pro or con. McGrew’s most rigorous statement of the underlying probability is a 2018 article in the philosophy journal Erkenntnis, “Undesigned Coincidences and Coherence for an Hypothesis” — a treatment of the formal logic of varied, dovetailing evidence rather than a verse-by-verse defense.²⁰ The mainstream guild has largely not engaged the revived argument, and a fair observer should treat that silence as neither endorsement nor refutation. The live, unsettled questions are exactly the ones above: how much the consensus of literary dependence erodes the cross-Gospel cases, and whether the pattern survives a properly disciplined accounting of hits and misses. These are real disputes among serious people, and the matter is genuinely contested — it should be presented as a promising and suggestive line of evidence under live debate, not as a settled result.

Conclusion: What This Does and Does Not Establish

Undesigned coincidences are best understood as a contributor to verisimilitude — the quality of reading like truthful report from the inside — and, where the strongest cases hold, as positive evidence that the Gospels and Acts rest on well-informed, substantially independent sources rather than on late legend spun at a distance or on a single coordinated fabrication. That is a real and valuable conclusion. It is also a bounded one, and the bounds must be stated as firmly as the strength.

The argument does not prove the Gospels inerrant; honest, well-placed witnesses can still err in detail, and the existence of apparent discrepancies among the accounts is not in dispute (indeed, surface tensions are themselves evidence the writers were not colluding). It does not, by itself, establish that any miracle occurred; coincidences mostly touch the mundane scaffolding of the narratives — geography, names, seasons, court gossip — and the credibility they build must be carried over to the extraordinary claims by separate argument. And after the appropriate concessions to literary dependence and selection bias, the case rests most securely on the Acts–Epistles material and the best-shaped cross-Gospel cases, not on every example ever paraded.

What it does accomplish is to make one thesis harder to maintain: that the Gospels are detached, late, free-floating legend. Texts that interlock at the level of why-Jesus-asked-Philip read like the residue of real knowledge of real events and people. That is corroboration, not proof — one strand in a larger cable that includes how early the Gospels were written, whether they rest on eyewitness testimony, and whether their incidental details (such as personal names) fit first-century Palestine. Taken together those strands build a cumulative case for reliable, well-sourced testimony. Whether what that testimony reports is true — up to and including Easter morning — is the larger conversation these quiet coincidences help open, but cannot by themselves close.