HistoricalChristian.Faith
HistoricalChristian.Faith
Related doctrines: Hard passages — see also Who Killed Goliath? (2 Samuel 21:19) · How Tall Was Goliath? (six or four cubits) · Solomon's Stalls: 40,000 or 4,000? (1 Kings 4:26 vs. 2 Chronicles 9:25) · Ahaziah's Age: 22 or 42? (2 Kings 8:26 vs. 2 Chronicles 22:2) · The Bronze Sea: 2,000 or 3,000 Baths? (1 Kings 7:26 vs. 2 Chronicles 4:5) · What Jesus Wrote on the Ground (John 8:6) · The Two Genealogies of Jesus (Matthew 1 vs. Luke 3) · The Death of Judas (Matthew 27 vs. Acts 1) · Baptized for the Dead (1 Corinthians 15:29)
Four times the Gospels name a prophet — or “the prophets” — and quote words that the named prophet did not write, or that no single passage contains. Mark 1:2–3 opens “as it is written in Isaiah the prophet,” then quotes Malachi 3:1 (fused with Exodus 23:20) before it reaches Isaiah 40:3. Matthew 27:9 credits “Jeremy the prophet” with an oracle whose words are Zechariah 11's. Matthew 2:23 reports “He shall be called a Nazarene,” “spoken by the prophets,” yet no verse reads so. And Matthew 13:35 cites “the prophet” for a line that belongs to Asaph's Psalm 78 — where a strong early line of manuscripts, harder still, names “Isaiah.” The charge is the same each time: the Evangelist has named the wrong prophet, a plain error in the text.1
This page gathers the readers who felt the difficulty and answered it, earliest first — not the sources for any one crux (Matthew's “Jeremy” is weighed in full with Judas's biographical timeline) but the chain of interpreters who recognized a single phenomenon and explained it. The arc runs one way. The “correction” begins in the manuscripts almost at once — Irenaeus already reads Mark's opening as “the prophets,” plural — and the fathers answer with a shared principle: the citation is a deliberate composite, and the Evangelist names the senior or earlier prophet, giving (in Jerome's phrase) “the sense, not the words.” Augustine adds a text-critical defence of the harder attribution; the medievals systematize the answer; and modern criticism vindicates the harder readings as original while recovering “composite citation” as a recognized ancient convention rather than a blunder. Whether any given case is deliberate design or a genuine slip is weighed adversarially under Arguments.
The earliest datable trace of the difficulty is not an explanation but a “correction” already in place. Quoting Mark's prologue, Irenaeus reads it in the plural: “Mark, the interpreter and follower of Peter, does thus commence his Gospel narrative: The beginning of the Gospel of Jesus Christ… as it is written in the prophets, Behold, I send My messenger before Thy face.” His second-century text already carries the general “in the prophets” rather than the singular “in Isaiah” — the shape the later manuscript tradition would take, and the first sign that scribes felt the tension between the named prophet and the composite quotation.2
The first writer known to stop over the difficulty and name it is Origen. Reaching Matthew 27:9 in his Commentary on Matthew, he says plainly that the quotation credited to “Jeremiah” “is not found to have been prophesied by Jeremiah anywhere in his books, whether those which are read in the churches or those which are recited among the Jews,” and — inviting correction (“if anyone can know, let him show where it is written”) — offers two conjectures: “either it is an error of the Scripture and Jeremiah is put for Zechariah, or… there is some secret writing of Jeremiah in which it is written.” He is thus the source of both later escape-routes — the scribal slip and the lost book — though he presses the weaker (defending the “secret writing” by analogy with Paul’s use of non-canonical sources) and never reaches the deliberate-composite reading the tradition would settle on.3
Eusebius, the great early systematizer of Gospel-discrepancy questions, takes the text-critical escape-route on Matthew 27:9 and presses it further than anyone before him. Expounding Zechariah 11 in his Proof of the Gospel, he reaches the thirty-pieces oracle Matthew credits to “Jeremiah” and finds it “not… in the prophecy of Jeremiah”; the reader must therefore “consider whether… they have been removed through any evil intention, or whether there has been an error in copying, through the mistake of some careless transcriber of the Holy Gospels, who wrote Jeremiah instead of Zechariah.” He extends the conjecture from the prophet’s name to the wording, supposing the same scribe replaced “they cast them into the house of the Lord, into the furnace” with “they bought… the field of the potter.” It is the copyist theory the reference works credit to him — but it undoes the harder, original name rather than the later corrections, and the deliberate-composite reading the tradition would settle on is still ahead of him.4
The anonymous Roman commentator known as Ambrosiaster is the first to state the crux as a formal question and give the classic answer. His Questions on the Old and New Testaments asks outright why, though the words stand “in the prophet Malachi… the Evangelist Mark claims that this is written in the prophet Isaiah.” His reply: Mark “could not have been unaware of what he was writing… Knowing that all sayings must be referred back to their original author, he attributed these words to Isaiah, who had first conveyed this meaning… so that he might unite the words of both prophets… under the name of the earlier prophet.” Naming the senior prophet for a blended citation is thus a deliberate convention, not a mistake.5
Preaching on Matthew 2:23, Chrysostom meets the harder version of the charge — a citation found in no extant prophet at all — with a different resort: the loss of prophetic books. “And what manner of prophet said this? Be not curious, nor overbusy. For many of the prophetic writings have been lost; and this one may see from the history of the Chronicles.” Where Ambrosiaster harmonizes a known composite, Chrysostom appeals to a source no longer preserved — the weakest of the ancient answers, but a candid admission that the words could not be located.6
Jerome is the pivot: he addresses all four loci and supplies the principle that governs the rest. On Mark 1:2 he is blunt — the quotation “is made up from two prophets, Malachi and Isaiah… on what ground then has Mark … set the words ‘As it is written in the prophet Isaiah’” when part is “not written in Isaiah at all, but in Malachi?” His answer is the method itself: the Evangelist “has not rendered his original word for word, but… has given the sense in different terms.” On Matthew 2:23 he notes that Matthew wrote “the prophets” in the plural precisely to show “that he did not take words from the Scriptures, but the meaning” (pointing to Isaiah 11:1's netzer, “branch”). On Matthew 27:9 he judges the Jeremiah oracle “more like a testimony taken from Zechariah.” And on Matthew 13:35 he diagnoses the “Isaiah” reading as an early scribal error for “Asaph the prophet”: the “first scribe did not understand Asaph, and thought it was a mistake… and corrected it to the more obvious name of Isaiah.”7
Augustine's Harmony of the Gospels gives the fullest single engagement with the charge — on Matthew's “Jeremiah” for Zechariah, the case the tradition treats as paradigmatic. He first weighs the text-critical escape and declines it: though “this ascription… is not contained in all the codices,” so that one might follow those that read simply “the prophet,” he is “not satisfied” with that, “because a majority of codices contain the name of Jeremiah, and… the more ancient Greek exemplars” have it — the harder reading is original, and there “was… an intelligible reason for erasing the name.” He then explains the name as providential: even if Matthew's memory first supplied “Jeremiah” for “Zechariah,” he would have corrected it “had he not reflected that… it was not without a purpose… directed by the Holy Spirit,” showing “the marvellous manner in which all the holy prophets, speaking in one spirit, continued in perfect unison.”8
Bede transfers Augustine's reasoning back to Mark and gives it its fullest harmonizing form. Of Mark's two testimonies, one “is only found in Isaiah, while the other is found in Malachi,” yet “the evangelist should not be thought to have erred or been deceitful… even if the words… from Malachi are not found in Isaiah, the sense of them is found in Isaiah.” He allows that “Isaiah occurred instead of Malachi” in Mark's mind “as happens,” but that Mark left it uncorrected because the name that surfaced was “governed by the Holy Spirit, not without reason” — Augustine's providential-memory argument, now settled Latin doctrine for the Mark passage.9
By the high Middle Ages the composite citation is stated as routine and the rival answers are catalogued. Theophylact of Ohrid notes without alarm that Mark's testimony “is taken from two prophets — from Malachi… and from Isaiah,” and for Matthew 2:23 offers both the lost-books resort and “an unwritten prophecy among the Jews.” Thomas Aquinas's Catena Aurea gathers the whole spectrum — including the Greek Pseudo-Chrysostom (Victor of Antioch), who states the two live options plainly: “it may be said that it is a mistake of the writer… Otherwise… he has compressed into one, two prophecies… by two prophets.” The deliberate-composite reading and the scribal-error reading are, by now, both on the table.10
Modern textual criticism vindicates the harder attributions as original: at Mark 1:2 the United Bible Societies' committee prints “in Isaiah the prophet” with an {A} (virtually certain) rating, reasoning that the composite Malachi–Isaiah quotation gave scribes the motive to soften the singular name to “in the prophets” — the lectio difficilior is therefore the source, not the “correction.” Modern exegesis, in turn, has recovered “composite citation” as a recognized convention of ancient authors (Jewish, Greco-Roman, and Christian alike), not a Gospel blunder — the two-volume survey of Sean Adams and Seth Ehorn distinguishing combined, conflated, and condensed citations across the corpus, with a dedicated study of Matthew. The one case still argued as a mislabelling (Matthew 13:35's “Isaiah” for Asaph) is the case the manuscripts and Jerome alike treat as a scribal slip — the exception that proves the fathers were reading the rest as design.11
The contested cruxes — each cutting across the timeline rather than sitting at one date — are weighed adversarially on their own pages. The settled floor beneath them is that “composite citation” is a real ancient convention and that the harder attributions are the original text; what is genuinely disputed is whether a given case is deliberate design or a genuine slip.
All four contested cruxes turn on the same move: a fulfilment formula that names one prophet — or “the prophets” — for words the named prophet did not (singly) write. Read together they show a range: a demonstrable two-text composite headed with one name (Mark), a words-from-one-prophet/imagery-from-another composite headed with the other (Matthew 27:9), and a summary of a whole theme that sits in no single verse at all (Matthew 2:23) — and then the honest counter-example (Matthew 13:35), where the quotation is from a single psalm, no composite is in play, and the best reading is not design but a scribal slip. It is the case that keeps the doctrine from over-reading every instance as deliberate.
Why Mark names Isaiah (Mark 1:2–3) — the headline case of the doctrine. Mark opens “as it is written in Isaiah the prophet,” but the first line quoted is Malachi 3:1 (its wording drawn toward Exodus 23:20) and only the second is Isaiah 40:3; best read as a deliberate composite headed with the name of the major, governing prophet — the source of the very “voice in the wilderness” line (Isaiah 40:3) the citation is built to reach. This is sound: the composite is undeniable, the “name the senior prophet” convention is documented (Matthew does the same at 27:9), and text-criticism confirms the harder “in Isaiah” as the original wording — the plural “in the prophets” a later softening, rated {A} by Metzger’s committee — so the bare charge of a blunder is the weaker verdict, and it fails either way, since on the minority plural reading there is no misattribution at all. What stays genuinely open is only why Isaiah in particular.12
Why Matthew says “the prophets” for the Nazarene saying (Matthew 2:23) — the hardest form of the crux, since “He shall be called a Nazarene” sits in no verse at all. Best read not as a mislabelled quotation but as a deliberate summary citation: Matthew’s plural “prophets” (with the participle “saying” dropped — his consistent marker for a thematic rather than verse-specific fulfilment) gathers up the Old Testament expectation of a lowly, despised Messiah that Nazareth’s obscurity answers. This is sound: the plural is grammatically deliberate, the summary needs no single verse (so the “no such text” charge dissolves), and it fits the same convention as the sibling cases; the weakest resort — an appeal to a lost prophetic book (Chrysostom, Theophylact) — is made needless by it. What stays open is only whether a proof-text (Isaiah 11:1’s netzer the strongest) also colours the wording.13
Why Matthew names Jeremiah (Matthew 27:9–10) — the paradigm case, argued in full with Judas's biographical timeline. The quoted words come mainly from Zechariah 11, yet Matthew credits “Jeremy”; best read as a deliberate composite — Zechariah's words wrapped in Jeremiah's imagery of potter, purchased field, and a place renamed for innocent blood, and named for the governing prophet. This is sound: coherent, consonant with Matthew's demonstrable habit with fulfilment citations, and the mainstream explanation, so the bare charge of a blunder is not the reasonable verdict; what remains underdetermined is only why Jeremiah in particular.14
Why Matthew 13:35 names Isaiah for Asaph’s psalm (Matthew 13:35) — the honest counter-example in the set. The quoted words are wholly from Asaph’s Psalm 78:2, a single text, so — unlike the other three — there is no composite to head with a name, and the “name the senior prophet” convention cannot apply. Best read not as design but as a genuine mislabelling: on the harder early text (ℵ*, Θ, f1, f13, 33) that names “Isaiah” it is a slip, most plausibly (with Jerome) an early scribe’s substitution for the unfamiliar “Asaph the prophet”; on the majority text (“the prophet”) there is no error at all, since Asaph is himself named a “seer” who “prophesied.” This is sound: the interpretation holds across both readings of a genuinely {C}-rated text, and on neither is this a designed composite — it is the case that keeps the doctrine from treating its explanation of the other three as a universal solvent. What stays open is only the text itself, which the United Bible Societies’ committee rates {C}.15
The testimonia-source mechanism — a question that cuts across all four cruxes rather than sitting at one. Where the timeline asks which prophet a composite is headed with, this asks where the conflation came from: did the Evangelist fuse the texts himself, or inherit them already fused — and already labelled — from a pre-formed collection of Christian proof-texts (a “testimony book,” the classic hypothesis of J. Rendel Harris)? This is plausible: the testimonia genre is now documented fact rather than conjecture — Qumran’s pre-Christian 4QTestimonia is a physical sheet of grouped, excerpted proof-texts, one block of which is itself a pre-formed composite — and written testimony-sources are attested near the New Testament (Justin, per Skarsaune). What holds it short of proven is that no surviving collection contains Mark’s or Matthew’s composite, so the mechanism is inferred for these verses, not read off a document; that Harris’s maximal “one fixed Testimony Book” form is set aside (C. H. Dodd: such books were “the result, not the presupposition” of early Christian exegesis); and that the mechanism relocates the deliberate composite upstream to a compiler rather than dissolving it, and — since the Qumran sheet labels nothing — explains the conflation better than the wrong prophet, which still needs the “name the senior prophet” convention or a slip.16
The epistolary comparanda — the second question that cuts across the whole doctrine rather than sitting at one crux. Where the four Gospel cases each pair a conflation with a single (allegedly wrong) prophet’s name, the epistles display the bare act of fusing texts — unattributed and unchallenged — all through: Romans 3:10–18 strings six texts (Psalms 14, 5, 140, 10, 36 and Isaiah 59) under one “as it is written”; Romans 9:33 welds the two Isaiah “stone” texts (28:16 + 8:14) into one testimonium; Romans 11:8 splices Deuteronomy 29:4 with Isaiah 29:10; and Hebrews 1 builds its case from a strung catena of seven psalms and prophets. This is sound: the compositeness of the Pauline cases is uncontested, was recognized as a chain by their earliest careful readers (Origen tracks the Romans 3 catena source by source; Aquinas names the two-Isaiah seam at 9:33 and the two-source splice at 11:8), and is the stock example of the technique in modern citation scholarship — so the doctrine’s “settled floor” is here corroborated from outside the four disputed cruxes, and with no misattribution in play to muddy it (strip the name away and combining texts draws no charge of error). The scope stays honest: these prove the conflation was ordinary craft, not that the wrong prophet is thereby explained (that still needs the “name the senior prophet” convention or, at Matthew 13:35, a slip) — and Hebrews’ chain is a florilegium of separately-introduced citations, comparanda for the clustering habit rather than the welding. Their text-form is a separate question, adjudicated verse by verse on Which text did the New Testament quote?17
The four loci and their Old Testament sources: Mark 1:2–3 (“in Isaiah the prophet”), quoting Malachi 3:1 (with Exodus 23:20) then Isaiah 40:3; Matthew 27:9–10 (“Jeremy the prophet”), the words substantially Zechariah 11:12–13; Matthew 2:23 (“spoken by the prophets… a Nazarene”), located in no verse verbatim (cf. Isaiah 11:1, netzer); Matthew 13:35 (“by the prophet”), quoting Psalm 78:2, a psalm of Asaph. ↩
Irenaeus, Against Heresies III.10.5 (ANF vol. 1; local text Against Heresies Book 3), quoting Mark's prologue as “as it is written in the prophets” — a second-century witness to the plural (“corrected”) reading. On the manuscript split, see note 9. ↩
Origen, Commentary on Matthew, Commentariorum Series §117 (on Matthew 27:3–10): the quotation ascribed to “Jeremiah” “is not found to have been prophesied by Jeremiah anywhere in his books… I suspect… that either it is an error of the Scripture and Jeremiah is put for Zechariah, or that there is some secret writing of Jeremiah in which it is written.” This later portion of the commentary (Matthew 22–27) survives only in an old Latin translation, the Commentariorum Series — the Greek is lost, and the Ante-Nicene Fathers volume renders only the surviving Greek books (through Matthew 22) — so the words are quoted from the fuller Latin series. See the detail page, Origen on Matthew’s “Jeremiah” for Zechariah’s words. ↩
Eusebius of Caesarea, Proof of the Gospel (Demonstratio Evangelica), Book 10 (W. J. Ferrar’s translation; also online at tertullian.org) — on Matthew 27:9: the oracle is “not found in the prophecy of Jeremiah,” so either the words “have been removed through any evil intention, or… there has been an error in copying… who wrote Jeremiah instead of Zechariah… and instead of ‘And they cast them into the house of the Lord, into the furnace,’ wrote in error, ‘And they bought with them the field of the potter.’” See the detail page, Eusebius of Caesarea on Matthew’s “Jeremiah”. On why “Jeremiah” is nonetheless the harder, original reading (so that the copyist conjecture reverses the true direction of the variation), see note 11 and the Arguments section. ↩
Ambrosiaster, Questions on the Old and New Testaments, Question 57 (“Why Mark attributes Malachi's words to Isaiah”). Long transmitted under Augustine's name (whence its appearance as “Augustine” / “Pseudo-Augustine, Quaest. nov. et vet. Test. lvii” in Aquinas's Catena Aurea), it is Ambrosiaster's — one witness, not several. ↩
John Chrysostom, Homilies on Matthew, Homily 9 (on Matthew 2:23; NPNF1 vol. 10; local text Homily 9 on Matthew) — the appeal to lost prophetic writings. ↩
Jerome on all four loci: on Mark 1:2, Letter 57 to Pammachius §9 (“made up from two prophets, Malachi and Isaiah… given the sense in different terms”; local text Letter 57); on Matthew 2:23, 13:35, and 27:9, his Commentary on Matthew (the plural “prophets” = “the meaning, not the words” and Isaiah 11:1's netzer; the “Jeremy” oracle “more like a testimony taken from Zechariah”; the “Isaiah” of 13:35 a scribal error for “Asaph the prophet”). Cf. Lives of Illustrious Men 3 (the Evangelists follow the Hebrew, not the Septuagint). ↩
Augustine, Harmony of the Gospels III.7.28–31 (NPNF1 vol. 6; local text Book 3) — on Matthew 27:9's “Jeremiah”: the harder reading is original (“a majority of codices… the more ancient Greek exemplars”), and the name is providential (“directed by the Holy Spirit”). Where Augustine himself quotes Mark 1:2 (Harmony II.6.18) he reads “Isaiah” and passes the composite by in silence; it is his Matthew 27:9 discussion that the later tradition applies to Mark. ↩
Bede, Commentary on Mark, on 1:2–3 (“the evangelist should not be thought to have erred… the sense of them is found in Isaiah… governed by the Holy Spirit, not without reason”) — Augustine's providential-memory argument transferred to the Mark passage. ↩
Theophylact of Ohrid, Commentary on Mark 1 (the composite of Malachi and Isaiah) and Commentary on Matthew (on 2:23, the lost books and “an unwritten prophecy among the Jews”); Thomas Aquinas, Catena Aurea, Mark 1, gathering the guesses — including Pseudo-Chrysostom / Victor of Antioch (“it may be… a mistake of the writer… Otherwise… he has compressed into one, two prophecies”). ↩
On Mark 1:2's text: Bruce M. Metzger, A Textual Commentary on the Greek New Testament, 2nd ed. (United Bible Societies, 1994), 62 (“in Isaiah the prophet,” rated {A}: the composite citation motivated the scribal change to the plural “in the prophets”), with the manuscript catalogue at “Mark 1:2 — In Isaiah, or In the Prophets?” (noting James Snapp's minority dissent). The composite-citation convention: Sean A. Adams and Seth M. Ehorn, eds., Composite Citations in Antiquity, 2 vols. (Bloomsbury T&T Clark, 2016, 2018) — vol. 2, New Testament Uses, incl. Maarten J. J. Menken on Matthew, distinguishing combined, conflated, and condensed citations. On Matthew 2:23 as a summary (non-)citation, “‘He shall be called a Nazarene’ — the Non-Citation of Matthew 2:23,” Tyndale Bulletin; on Matthew 27:9's variants, the standard reference article. ↩
See the argument page Why does Mark say “in Isaiah the prophet” for a Malachi–Isaiah blend? (assessment: sound), reading Mark 1:2–3 as a deliberate composite of Malachi 3:1 (in an Exodus 23:20 frame) and Isaiah 40:3, headed with the name of the major prophet; with the text-critical finding (Metzger, {A}) that “in Isaiah the prophet” is the original reading and “in the prophets” a later softening. ↩
See the argument page Why does Matthew say “spoken by the prophets” for a saying found in no verse? (assessment: sound), reading Matthew 2:23 as a deliberate summary citation — marked by Matthew’s plural noun and dropped participle — of the Old Testament theme of a lowly, despised Messiah, with Isaiah 11:1’s netzer the strongest of the debated single-proof-text candidates and a lost prophetic book (Chrysostom, Theophylact) the weakest. ↩
See the argument page Why does Matthew credit the thirty-pieces oracle to “Jeremy the prophet”? (assessment: sound), hosted with Judas's biographical timeline, reading Matthew 27:9 as a deliberate composite of Zechariah 11 and Jeremiah 18–19/32. ↩
See the argument page Is Matthew 13:35’s “Isaiah” a scribal slip for Asaph, not a composite citation? (assessment: sound), reading Matthew 13:35 as the counter-example in the cluster: a single-source quotation of Asaph’s Psalm 78:2 (superscription “Maschil of Asaph”), so no composite is in play — a mislabelling (an early scribe’s substitution for “Asaph,” per Jerome’s Commentary on Matthew) on the harder early reading “Isaiah,” and no error at all on the majority “the prophet,” since Asaph is a prophet on Scripture’s own terms (2 Chronicles 29:30). The text-critical decision is rated {C} by Metzger’s committee. ↩
See the argument page Were the composite citations drawn from a pre-formed testimonia collection? (assessment: plausible), weighing the Testimony Book / Testimonia hypothesis of J. Rendel Harris (Testimonies, 1916–20) — that the conflation, and sometimes the single wrong attribution, was inherited from a pre-formed proof-text collection rather than fused by the Evangelist — against C. H. Dodd’s critique (According to the Scriptures, 1952: such books were “the result, not the presupposition” of early Christian exegesis) and the Qumran evidence that the genre existed pre-Christian (4QTestimonia, 4Q175). A mechanism upstream of the deliberate-composite reading, not a rival to it. ↩
See the argument page Do Paul’s and Hebrews’ unattributed conflated catenae show composite citation was a workaday convention? (assessment: sound), reading the strung proof-texts of Romans 3:10–18 (a six-text combined citation), the two-Isaiah stone of Romans 9:33 (Isaiah 28:16 + 8:14), the Deuteronomy–Isaiah splice of Romans 11:8, and the catena that opens Hebrews 1 as comparanda for the composite-citation convention — unattributed, so members of neither the “Naming the Wrong Prophet” set nor the text-form catalog of Which text did the New Testament quote? The compositeness is recognized in Origen’s and Aquinas’s Romans commentaries and catalogued as the technique’s stock examples by Adams and Ehorn, Composite Citations in Antiquity, and Stanley, Paul and the Language of Scripture. A companion to the testimonia-source argument: both reach the same limit — the comparanda evidence the conflation, not the wrong prophet. ↩