An earlier tradition of using Matthew against Marcionites is demonstrable. Its incorporation into Tertullian’s Against Marcion is a serious hypothesis, especially for the arguments about fulfilling the law and the Father who sends sun and rain. What remains unproven is a separately recoverable document, originally organized around Matthew, underlying the whole of Tertullian’s Gospel commentary.
The strongest evidence comes from surviving Irenaeus, not simply from anomalous quotations in Tertullian. Irenaeus, Against Heresies IV.34.1–2, explicitly directs his argument principally against Marcion’s followers and quotes Matthew 5:17–18. IV.13.1 likewise names Marcionites while discussing the Sermon on the Mount. Thus the earlier use of Matthew in this controversy does not depend on reconstructing Irenaeus’s projected separate work against Marcion.
1. What Williams noticed
David Salter Williams examined the problem in “On Tertullian’s Text of Luke,” Second Century 8 (1991):193–199. His proposal was cautious: “in at least one case,” a harmonistic reading in Tertullian’s own Luke might explain a charge that Marcion removed material otherwise found in Matthew. His example concerned Luke 6:35–36 and Matthew 5:45; he described the investigation as “exploratory and programmatic.” These formulations are verified through Roth’s quotations of pp. 196 and 199, in his 2022 review, p. 97 n. 21. Williams’s complete article was not recovered for independent examination.
That proposal concerns a harmonized Gospel manuscript. The proposal investigated here concerns the reuse of an earlier argument. A commentator could inherit a Matthew-based defense without possessing a Luke manuscript containing those Matthean sentences.
Roth’s 2008 article, “Matthean Texts and Tertullian’s Accusations in Adversus Marcionem,” offers another explanation: most accusations concern rejection of material belonging to the church’s fourfold Gospel, conceived as one Gospel. He treats the sword/division case as a memory failure. His published abstract and his full 2022 restatement were examined; the entire 2008 article was not recovered.
Peter Head’s 1993 discussion, pp. 316–317 n. 40, identifies another possible route: Matthew could be debated in the Antitheses, without belonging to Marcion’s Gospel. He attributes that explanation to Harnack and Blackman. None of these alternatives automatically excludes earlier argumentative sources.
2. The main passages, distinguished by function
AM denotes Tertullian’s Adversus Marcionem; AH denotes Irenaeus’s Adversus haereses. English renderings are supplied here. The table is a targeted audit of diagnostic cases, not a count of every Matthean allusion.
| Passage and Gospel material | Tertullian’s wording or procedure | Significance for an earlier layer |
|---|---|---|
| IV.7.4; IV.9.10, 15; IV.12.14; IV.36.6; V.14.14 Matthew 5:17: fulfillment of law and prophets. | Hoc enim Marcion ut additum erasit (IV.7.4): “For Marcion erased this as an addition.” Later he reasons that Jesus must have said it because his actions fulfilled it. | The saying becomes a recurring rule for interpreting episodes scattered through the Lukan narrative and even Romans. This looks like a reusable doctrinal argument. Its repetition does not establish a source document, but is more substantial than an isolated quotation slip. |
| IV.7.5 Matthew 15:24, 26: Israel’s lost sheep; children’s bread and dogs. | Aufer etiam illud de evangelio… “Remove that too from the Gospel…” | The imperative issues a sarcastic challenge. It should not automatically be converted into a past-tense report that Marcion had excised these verses from Luke. Matthew’s Israel-centered Jesus provides an external argument against Marcion’s Christology. |
| II.17.1; IV.17.6–7; IV.36.3 Matthew 5:45: sun and rain. | … Marcion de evangelio eradere ausus est (II.17.1): “Marcion dared to erase [this testimony] from the Gospel.” Euge, Marcion, satis ingeniose detraxisti illi pluvias et soles… (IV.17.6): “Well done, Marcion: very cleverly you removed his rains and suns…” | This is the strongest recurring cluster. A defense of the Creator’s goodness appears in Book II, is applied to Luke 6:35 in Book IV, and returns with the rich ruler. The same prooftext works independently of a running commentary on Luke. |
| IV.29.13–14 Matthew 10:34 / Luke 12:51: sword / division. | Machaeram quidem scriptum est. Sed Marcion emendat. “‘Sword,’ indeed, is what is written. But Marcion corrects it.” | A Matthean reading is treated as the norm when Tertullian discusses a Lukan passage. It is a clear problem for his comparison. Memory, a harmonized source, and inherited accusation remain distinguishable explanations. |
| IV.14.1, 13 Luke 6:20 / Matthew 5:3: kingdom of God / heavens. | The first quotation reads regnum dei; the later composite quotation reads regnum caelorum. | A shift within one discussion shows that Matthean phrasing can enter Tertullian’s exposition. Because it occurs in a scriptural combination with Isaiah, it cannot by itself establish a separate literary source. |
| IV.26.11; IV.28.2 Luke 11:15 / Matthew 12:24: Beelzebul accusation. | The later reference uses Hic non expellit daemonia nisi in Beelzebub, the Matthean formulation, in retrospectively discussing the earlier scene. | Another useful control for flexible synoptic recollection. A theory attributing every Matthean formulation to a single earlier document would have to explain these local shifts. |
| IV.19.6–10 Matthew 12:48 / Mark 3:33; compare Luke 8:21. | Quae mihi mater, et qui mihi fratres? “Who is my mother, and who are my brothers?” Tertullian calls this an argument used by those contesting Jesus’ birth. | The interrogative form belongs to Matthew/Mark rather than ordinary Luke. This is a portable anti-docetic argument, not uniquely Matthean material. On the Flesh of Christ 7 explicitly reuses Tertullian’s reply to Marcion. |
| IV.34.2–6 Matthew 19:3–9 and 5:32: divorce. | Non enim recepisti illud quoque evangelium… “For you have not accepted that other Gospel as well…” Then: Nam et in evangelio Matthaei…, “For also in Matthew’s Gospel…” | A decisive control: Tertullian can knowingly distinguish an external Gospel from the material Marcion accepts. He explicitly returns to refuting Marcion from accepted material in IV.34.3. All Matthean use cannot be classified as unconscious confusion. |
| III.13.6 Matthew 2:11: Magi. | Redde evangelio veritatis quae posterior detraxisti… Maneant autem orientales illi Magi… “Restore to the Gospel of truth what you removed later… Let those eastern Magi remain…” | Matthew’s infancy material already serves the prophecy argument before the running Gospel examination. Compare Justin, Dialogue 78, on the Magi and Isaiah’s Damascus/Samaria prophecy: a demonstrable older apologetic setting, here directed against Trypho. |
| IV.21.1 Mission instructions; Matthew 10:10 and Luke 10:7. | Dignus operarius mercede sua. Haec Marcion deleat, dum sensui salva sint. “The worker is worthy of his wage. Let Marcion delete these things, provided they remain safe in meaning.” | A control against inflating the Matthean inventory: the wage saying also occurs elsewhere in Luke. The jussive “let him delete” is not, by itself, a documentary report of an actual deletion. |
3. The strongest earlier evidence: Irenaeus explicitly argues against Marcionites from Matthew
AH IV.34.1–2 supplies a direct answer to the historical question. Irenaeus announces a response principally to the followers of Marcion, urges comparison of the apostolic Gospel with the prophets, and quotes Matthew 5:17–18:
Ne putetis, quoniam veni dissolvere Legem aut prophetas: non veni dissolvere, sed adimplere.
“Do not think that I came to dissolve the law or the prophets: I came not to dissolve, but to fulfill.”
The earlier argument is already in place: Jesus fulfills the prophets, so the God of the prophets cannot be alien to him. Tertullian repeatedly applies this conclusion to episodes in his Gospel commentary. Irenaeus also joins the fulfillment passage to Pauline testimony in Romans 3:21. This is a surviving example of Matthew functioning in an anti-Marcionite argument, not evidence requiring the hypothetical contents of a lost book.
AH IV.13.1 makes the controversy’s terminology even clearer. After citing Matthew’s sayings on adultery, anger and oaths, Irenaeus rejects the Marcionite interpretation:
Omnia enim haec non contrarietatem et dissolutionem praeteritorum continent, sicut qui a Marcione sunt vociferantur; sed plenitudinem et extensionem…
“For all these contain no contradiction and dissolution of the earlier provisions, as Marcion’s followers loudly maintain, but fulfillment and extension…”
The opposition is between contrarietas and plenitudo/ extensio: contradiction and overthrow, or fulfillment and enlargement. Tertullian’s treatment of the corresponding ethical material follows the same conceptual pattern:
| Irenaeus | Tertullian | Assessment |
|---|---|---|
| IV.13.1: non contrarietatem et dissolutionem… sed plenitudinem et extensionem. No contradiction and dissolution, but fulfillment and extension. | IV.16.2: supplementa consentanea disciplinae creatoris. Additions consonant with the Creator’s teaching. IV.16.6: non adversario sed adiutore praecepto. A supporting command, rather than an opposing one. | The same answer to the claim that stronger ethical demands imply an opposing God. Conceptual continuity is clear; the wording does not prove a direct translation of this particular Irenaean passage. |
| IV.13.3: love of enemies, generosity to those taking possessions, and the Golden Rule belong to the fulfillment of the law. | IV.16.1–17: the corresponding ethical instructions express or extend the Creator’s earlier teaching. | A related cluster of arguments. Some order follows shared Gospel material, so sequence alone cannot identify a source. |
| IV.13.3: Patri, qui solem suum oriri facit super malos et bonos, et pluit super justos et injustos. The Father makes his sun rise on evil and good and sends rain on just and unjust. | II.17.1; IV.17.6–7; IV.36.3: sun and rain identify the Father’s goodness with the Creator’s action. | A particularly relevant earlier prooftext, occurring in the same broad defense of continuity between Christ and the Creator. |
The limitation is instructive: Irenaeus’s discussion itself combines Matthew and Luke. It supports an earlier collection or tradition of arguments using Matthew; it does not supply a purely Matthean commentary from which every Lukan element must have been added later.
4. Matthew 5:17 behaves as an interpretive rule, not simply a collated verse
Three formulations expose the structure of Tertullian’s argument.
First, IV.7.4 makes the direct accusation: Hoc enim Marcion ut additum erasit, “Marcion erased this as an addition.” That is what Tertullian alleges. It does not independently show how he reached the allegation.
Second, IV.9.15 explicitly gives an inference from conduct:
Constat ergo dixisse illum, quia et fecit.
“It is therefore established that he said it, because he also did it.”
Jesus tells the healed leper to perform what the law requires. Tertullian concludes that Jesus’ actions secure the truth of the Matthean statement, and that Marcion removed Jesus’ words rather than the church inserting them. The argument’s stated warrant is theological consistency between deed and saying. A modern reconstruction must not silently replace that warrant with an unreported manuscript collation.
Third, V.14.14 can make the point irrespective of the disputed saying:
Dixerit Christus an non, Ego non veni legem dissolvere sed implere…
“Whether Christ said or did not say, ‘I came not to dissolve the law but to fulfill it’…”
The surrounding argument is that law and Gospel mutually support one another. This is a concession for argument’s sake, not a confession that Tertullian does not know the words of Matthew. But it confirms that his principal aim is to establish doctrinal continuity. Matthew 5:17 serves that aim across changing scriptural contexts.
5. Sun and rain: the best candidate for a reusable argumentative unit
Matthew 5:45 identifies the Father through his sun and rain. Luke 6:35 instead identifies him as kind to the ungrateful and evil. In AM IV.17.6 Tertullian joins the Lukan premise to an accusation that Marcion removed the Matthean phenomena. He then reasons that an unknown God who had given no earlier benefits could not complain of ingratitude; the Creator, who had sustained even idolaters, could.
The same Matthean testimony already appears in II.17.1. There Tertullian says that even though Marcion removed it from the Gospel, the world itself remains inscribed with the Creator’s goodness. The proof is available from nature as well as from Scripture. It reappears in IV.36.3 while Jesus’ declaration about the one good God is discussed.
Even this repeated prooftext has variable wording. II.17.1 pairs rain with “good and evil” and sun with “just and unjust”; IV.36.3 pairs rain with “just and unjust” and sun with “good and evil.” The stable argumentative function does not guarantee a fixed quotation form. An earlier source could be reused with changes, but these formulations cannot simply be treated as identical transcriptions from one exemplar.
This portability is significant. The argument is not generated solely by the immediate wording of Luke 6:35: its decisive illustration comes from Matthew and the shared created world. Its existence in Irenaeus gives the proposed earlier tradition a surviving point of comparison. What the repetition cannot decide is whether Tertullian took it directly from Irenaeus, another predecessor, a common collection, or his own earlier composition.
6. Controls against reconstructing too much
Tertullian knows when he is using another Gospel. In IV.34.2–6 he openly appeals to Matthew’s divorce material, acknowledges that Marcion does not receive it, and distinguishes the task of arguing from what Marcion does accept. This supports the possibility of deliberate comparison across Gospels and strengthens Roth’s objection to treating every Matthean reference as a lapse.
Some fluctuations look local. The movement from regnum dei to regnum caelorum within IV.14, and the changed Beelzebul formulation, show why ordinary synoptic mixing must remain an alternative. Roth discusses these citation habits in his 2015 book, pp. 90–91. Such variation need not identify an earlier source.
Some material has a different earlier target. Justin’s Dialogue 78 supplies an older comparison for the Magi/Damascus/Samaria interpretation used in AM III.13.6. Its surviving setting is argument with Trypho. The existence of older Matthean apologetic material does not establish that it originated against Marcion.
Tertullian’s own reuse can be explicit. On the Flesh of Christ 7.1 tells Apelles to hear the reply already given to Marcion in the work appealing to Marcion’s Gospel. The mother-and-brothers argument is therefore demonstrably reusable. That self-reference points back to Against Marcion; it cannot be used to prove that the argument entered it from an earlier version of On the Flesh.
The sword/division accusation remains problematic. Matthew has “sword”; ordinary Luke has “division.” A late Lukan manuscript attesting “sword” cannot by itself establish what Tertullian’s second- or third-century copy contained. Conversely, an unusual or harmonized earlier copy cannot simply be ruled out. A shared earlier version of the same mistaken accusation would be much more diagnostic of inherited polemic than the wording alone.
7. What can be reconstructed?
The evidence permits a working model with two overlapping elements:
- An older thematic defense: Christ fulfills the Creator’s law; stronger ethical commands extend it; the Father’s sun and rain identify him as the Creator. Irenaeus securely attests this tradition in an explicitly anti-Marcionite setting.
- The present sequence of AM IV: episodes from the Gospel associated with Luke become occasions for applying that defense. Matthean prooftexts recur, and their rejection is sometimes expressed as deletion from “the Gospel.”
This is a model of how reuse could work, not a recovered sequence of editorial stages. Tertullian could have combined the materials himself; an intermediary could already have combined them. His references to the four Gospels also allow him to accuse Marcion of rejecting the church’s Gospel as a whole. The evidence does not establish that an original Matthew-based refutation preceded the identification of Marcion’s text with Luke.
The next decisive evidence would be a distinctive sequence or shared error connecting a specific earlier witness to Tertullian’s accusations. General agreement that Jesus fulfills the law establishes continuity of controversy more readily than the identity of an immediate exemplar. Surviving Irenaeus is already a source candidate; the separate projected Against Marcion need not be invoked to explain every agreement.
A defensible conclusion: Tertullian’s account incorporates a Matthew-based defense of the Creator whose principal arguments are attested before him in Irenaeus’s controversy with Marcionites. That earlier argumentative tradition helps explain the Matthean material embedded in the present Luke-ordered commentary. It does not yet establish the boundaries or authorship of one underlying written stratum.
References and access
- Tertullian, Adversus Marcionem, Latin Books II, III, IV, V. Relevant passages were checked in Latin.
- Irenaeus, Against Heresies IV.13.1–3 and IV.34.1–2. Latin excerpts were checked against the surviving ancient Latin version in the Harvey edition; the links provide the public-domain English translation. Shared Latin terminology does not by itself show that Tertullian used that Latin translation rather than Greek Irenaeus.
- David Salter Williams, “On Tertullian’s Text of Luke,” Second Century 8 (1991):193–199. Relevant wording and scope verified through Roth 2022, p. 97 n. 21; full article not independently recovered.
- Dieter T. Roth, “Matthean Texts and Tertullian’s Accusations in Adversus Marcionem”, Journal of Theological Studies 59 (2008):580–597. Published abstract consulted; argument also checked through his 2022 restatement.
- Dieter T. Roth, review of Matthias Klinghardt’s Das älteste Evangelium und die Entstehung der kanonischen Evangelien, TC: A Journal of Biblical Textual Criticism 27 (2022):93–103, especially pp. 96–97 and notes 21–24. Full text consulted.
- Dieter T. Roth, The Text of Marcion’s Gospel (Leiden: Brill, 2015), pp. 87, 90–91, especially §4.3.4. Full relevant discussion consulted. The earlier dissertation is openly available; its pagination differs.
- Peter Head, “The Foreign God and the Sudden Christ: Theology and Christology in Marcion’s Gospel Redaction”, Tyndale Bulletin 44.2 (1993):307–321, pp. 316–317 n. 40. Relevant full discussion consulted.
- Tertullian, On the Flesh of Christ 7.1–5, Latin; English translation.
- Justin, Dialogue with Trypho 78, especially the Magi and the Damascus/Samaria prophecy.
Research checked 7 November 2026. The distinction between an attested earlier argument and a hypothesized written source is maintained throughout.
