Tertullian can repeat a biblical phrase closely, but he does not consistently reproduce Paul or the Gospel sayings in one fixed Latin form. His wording varies across works as well as within Adversus Marcionem. Sometimes he substitutes vocabulary; sometimes he changes the construction; sometimes he openly turns a saying into an explanation or a rhetorical question.
Table 1 compared his citations inside Adversus Marcionem. Table 2 compared his testimony with Epiphanius. Table 3 supplies another control: Tertullian against Tertullian across different works. Its first section compares Adversus Marcionem with other treatises. Its second compares other works with one another, where the comparison does not depend on assigning a reading to Marcion’s canon.
The table contains 24 selected comparisons across nine other works: 18 with lexical, structural or grammatical differences; 2 explicit recastings; 2 stable quoted cores; and 2 requiring special caution. These examples establish that his citation wording is not uniformly fixed. They were selected to illustrate the kinds of agreement and variation, so their proportions are not a frequency estimate for his entire corpus.
What is being compared
The object of comparison is the received Latin citation form. A Latin difference may reflect translation, rhetorical adaptation, memory, different biblical exemplars or the transmission of Tertullian’s own works. Each case needs to be assessed before it becomes evidence for a different underlying Greek reading. Romans 8:11 makes the transmission problem particularly visible.
Every Latin excerpt below has an English translation. Different Latin expressions sometimes produce the same natural English sentence. Ellipses remove intervening material; square brackets supply words needed to translate an excerpt intelligibly. Stable means that the displayed core is stable, not that an entire verse or Gospel has been independently verified word for word.
Table 3 summary — The 24 selected comparisons
Classification
Comparisons
What is being counted
Verbal variation
18
12 lexical or structural differences and 6 smaller grammatical or word order differences.
Explicit recasting
2
Commands or declarations deliberately turned into questions or narrative exposition.
Stable quoted core
2
The displayed clause repeats with the same words apart from punctuation or a removed he says.
Special caution
2
Romans 8:11 has transmission complications; 1 Corinthians 7:39 involves an argument about opponents’ readings.
Table 3A
Adversus Marcionem compared with other works
Table 3A — Sixteen comparisons across Adversus Marcionem and other works
qui suscitavit Iesum a mortuis suscitabit et mortalia corpora vestra
English: He who raised Jesus from the dead will also raise your mortal bodies.
Special caution Give life becomes raise, and Christ becomes Jesus in the printed texts. However, the resurrection treatise’s transmission includes both raise and give life readings. This is an important comparison requiring apparatus control, not an unqualified demonstration of authorial change.
English: Has God not made foolish the wisdom of this age?
Verbal variation World becomes this age, while the verb for making foolish remains. The Latin difference is clear. The demonstrative and the Greek equivalents require separate assessment before assigning a specific underlying Greek reading.
English: Do you not know that your bodies are members of Christ?
Verbal variation One negative verb becomes two words, and the second wording leaves is or are implicit. The English meaning is the same; this is grammatical flexibility, not a reversal about the body.
Verbal variation The printed forms differ in the case ending after in. The distinctive contracted time idiom survives. A small Latin grammatical difference should not be promoted into a different Pauline doctrine.
Caro et sanguis regnum dei hereditate possidere non possunt
English: Flesh and blood cannot possess the kingdom of God by inheritance.
Verbal variation A future statement becomes an inability construction, with inheritance expressed explicitly. Both contexts retain Tertullian’s defence of bodily resurrection through transformation.
Verbal variation Two Latin adjectives describe the same clay vessel image. This is lexical variation without a necessary difference in the Greek text or the argument.
English: We do not wish to be unclothed, but rather to be clothed over.
Verbal variation The participle becomes a finite verb and rather is added as Tertullian embeds the clause in his explanation. The clothing image and intended transformation remain.
Ut unusquisque ... reportet quae per corpus secundum quae gessit, bonum sive malum
English: So that each may receive back the things through the body, according to what he has done, good or evil.
Verbal variation Receive becomes receive back; committed becomes done; the construction also expands. Tertullian himself discusses the difficult arrangement of the latter wording in De resurrectione 43.7–8. It should not be flattened into a supposedly certain Greek reconstruction.
ad caput, id est ad initium, reciprocare universa in Christo
English: To bring all things back to the head, that is, to the beginning, in Christ.
Verbal variation Tertullian varies his rendering and explanatory gloss of summing up all things. In Adversus Marcionem he explicitly comments on how the Greek word sounds, making his own work as a translator visible.
English: Who is my mother, and who are my brothers?
Stable quoted core The quoted core is identical apart from punctuation. This question belongs to the wider Synoptic parallels, not to the exact wording of canonical Luke 8:21. De carne Christi expressly refers back to the earlier argument against Marcion: stable reuse is not an independent manuscript witness.
English: Whoever is ashamed of me, I too shall be ashamed of him.
Verbal variation The construction and word order change, while the reciprocal shame saying keeps the same force. The ellipses remove Tertullian’s he says, not biblical words.
Aspicite ... quod ego sum, quia spiritus ossa non habet sicut me habentem videtis
English: Look ... that it is I, because a spirit does not have bones, as you see me having.
Verbal variation See becomes look, the identity construction changes, and the final verb construction is recast. De carne Christi narrates the hands and feet immediately before its shorter quotation. Their absence from the quoted words is not proof of their absence from a Gospel exemplar. Both quoted forms lack flesh in the bones clause.
English: Nor does anyone else know the Father except the Son.
Verbal variation The knowledge saying changes its negative and verbal construction. The shared Synoptic saying does not identify which Gospel, or which Greek exemplar, supplied each occurrence.
a quo petam ut accipiam? apud quem quaeram ut inveniam? ad quem pulsabo ut aperiatur mihi
English: From whom should I ask, so that I may receive? With whom should I seek, so that I may find? At whose door should I knock, so that it may be opened to me?
quaerite et inuenietis ... pulsate et aperietur uobis ... petite et accipietis
English: Seek and you will find ... knock and it will be opened to you ... ask and you will receive.
Explicit recasting Adversus Marcionem deliberately turns the commands into rhetorical questions. These are citation and exposition, not two formal transcriptions of a single Gospel verse. The excerpts from De praescriptione follow the order in that discussion.
English: Which Christ did not dissolve, but fulfilled.
Explicit recasting The Gospel’s first person declaration becomes Tertullian’s third person description of Christ. This is deliberate narrative adaptation, not a contradiction in the underlying proposition.
per baptismum ... ut, sicut Christus resurrexit a mortuis
English: Through baptism ... so that, just as Christ rose from the dead.
Verbal variation The form of baptism, the comparative connective and the resurrection verb vary. Both passages use the burial through baptism argument.
Si enim consepulti sumus simulacro mortis eius, sed et resurrectionis erimus
English: For if we have been buried together in the likeness of his death, we shall also be [so] in resurrection.
Verbal variation Planted together becomes buried together. This is more than a change of word order: the metaphor has shifted, with the latter wording echoing the preceding burial passage. Both still connect Christ’s death with resurrection.
dolere cum dolentibus et ita demum congaudere gaudentibus
English: To grieve with those who grieve and only then rejoice together with those who rejoice.
Verbal variation The mourning verb changes, the clauses reverse order, and only then adapts the saying to the fasting argument. This is an identifiable rhetorical use of Paul outside a Marcionite textual contrast.
English: Why do you boast as if you had not received?
Stable quoted core The quoted core is stable. The first ellipsis removes he says. This is a control showing that Tertullian can reproduce a short Pauline formulation closely.
English: But the person who commits adultery transgresses against his own body.
Verbal variation Fornication becomes adultery, and the verb for sin also changes. The terminology suits the respective discussions. The Latin difference does not by itself prove that Tertullian had two Greek readings.
Vinctus es enim uxori? ne quaesieris solutionem; solutus es ab uxore? ne quaesieris uxorem
English: For are you bound to a wife? Do not seek release. Are you released from a wife? Do not seek a wife.
Verbal variation Marriage and a bond replace wife and seeking a wife in one formulation. This is a substantial contextual recasting of the same Pauline saying. The first ellipsis removes he says.
Verbal variation Two Latin verbs express taking or marrying a wife. Only the wife-taking clause is compared; the separate virgin clause is not silently joined to it. The Latin idioms alone cannot decide between possible underlying Greek formulations.
si autem dormierit, libera est: cui volet nubat, tantum in domino
English: But if he has fallen asleep, she is free: let her marry whom she wishes, only in the Lord.
Special caution Death and sleep language occur in a discussion that also accuses opponents of altering two Greek syllables. Because Tertullian distinguishes quoted forms and polemical alternatives here, this cannot be scored as an ordinary, unproblematic change in his own wording.
Why Ephesians 1:10 is especially revealing
In Adversus Marcionem V.17.1, Tertullian uses a Latin verb for summing up, then explains it with alternatives meaning to bring back to the beginning or to review from the beginning. He also comments on the expression itself:
ut ita dixerim, sicut verbum illud in Graeco sonat
English: If I may put it this way, as that word sounds in Greek.
In De monogamia 5, he instead describes bringing everything back to the head, that is, the beginning. The comparison exposes his work as a translator and interpreter. The explanatory alternatives in Adversus Marcionem are his glosses, not three additional independent quotations of Paul.
This matters when reconstructing Marcion. An unusual Latin word in an anti-Marcionite argument cannot automatically be attributed to Marcion’s own text when Tertullian is visibly explaining how he is rendering the Greek.
Variation also occurs outside the Marcion comparison
Romans 6:5 supplies a clear example. In De resurrectione, Tertullian has being planted together in the likeness of Christ’s death. In De pudicitia, he has being buried together. The second expression echoes the burial language immediately preceding it. Both connect Christ’s death with resurrection, but the metaphor is not reproduced in a fixed form.
The comparison of 1 Corinthians 6:18 is equally instructive. The person committing fornication becomes the person committing adultery; the verb for sin changes as well. The wording suits the respective discussions of sexual conduct. That contextual difference must be examined before the Latin is used to reconstruct a supposed change in Paul’s Greek.
Romans 12:15 shows a more obvious rhetorical adjustment. In De idololatria, rejoicing precedes mourning. In De ieiunio, grieving comes first, and only then does rejoicing follow. The order has been made useful to the argument about fasting and shared observance. It would be misleading to present those formulations as two neutral transcripts from competing biblical codices.
The stable examples also matter
The mother and brothers question appears with the same words in Adversus Marcionem and De carne Christi. The latter work explicitly points back to the earlier anti-Marcionite argument. This demonstrates stable reuse of a formulation, although it does not provide an independent manuscript witness.
The short Pauline challenge about boasting as though one had not received also remains stable between De oratione and De pudicitia. Tertullian’s practice therefore includes both close repetition and rewording. The right question is what he is doing in the particular passage under examination.
Comparisons that need a different rule
Romans 8:11. The printed texts contrast giving life with raising. But the apparatus for De resurrectione includes both readings in its transmission. The visible difference cannot simply be assigned to Tertullian without considering the manuscripts and editions. His parallel in Adversus Praxean 28 also uses raising language; that adds evidence, but does not remove the transmission problem.
1 Corinthians 7:39. The remarriage discussion in De monogamia includes an allegation that opponents altered two Greek syllables. Quoted alternatives and Tertullian’s own preferred interpretation cannot all be counted as equivalent personal citations. This case remains separate from the ordinary verbal differences.
Explicitly attributed Marcionite readings. At Adversus Marcionem V.10.7, Tertullian contrasts the last Adam with a reading he assigns to Marcion:
dominum enim posuit novissimum pro novissimo Adam
English: For he put the Lord as the last in place of the last Adam.
That is a specific attribution of an alteration. It is a different kind of evidence from a synonym, an abbreviated quotation or a rhetorical reformulation. General flexibility in Tertullian’s citations does not justify explaining every explicit allegation away.
Different Synoptic versions. The confession saying before God in Adversus Marcionem IV.28.4 and the before my Father form in Scorpiace 9 overlap with Luke 12:8 and Matthew 10:32. Such a comparison must first allow for different Gospel parallels. It is not added to the table as a straightforward instance of Tertullian changing one identifiable verse.
Different lengths of citation. At Galatians 4:4, a citation ending with God’s sending of the Son cannot be treated as a denial of the birth clause quoted in De carne Christi 20.2. The absence of that clause from a shorter excerpt is insufficient, on its own, to establish an omission from a manuscript.
What Table 3 establishes
Tertullian is not consistently verbatim in citing either Paul or the Gospel sayings. The variation is not confined to passages in which modern interpreters are trying to distinguish his text from Marcion’s. It also occurs between his other works. At the same time, some clauses remain stable and many reworded passages preserve the same argument.
The practical consequence is that a cross-work difference does not identify its own cause. Before turning it into a reading of Marcion’s canon, we must distinguish Tertullian’s quotation from his exposition, identify the biblical clause being compared, and check whether the wording is secure in the transmission of his work. His testimony remains evidence, but its exact Latin cannot be treated as an automatic transcript of a Greek exemplar.
Sources
The Latin excerpts were checked against the printed texts linked in each row. For Adversus Marcionem, the text is The Latin Library’s reproduction credited to Ernest Evans (1972), transcribed by Roger Pearse (2002). For the other works, the linked Tertullian Project texts identify their editions and, where available, provide an apparatus. Chapter references are used where electronic editions differ in paragraph numbering. De resurrectione is also known as De resurrectione carnis or De resurrectione mortuorum; these titles do not count as separate works.
Benjamin Douglas Haupt, Tertullian’s Text of the New Testament Outside the Gospels (PhD thesis, University of Birmingham, 2019), and its associated citation dataset, helped locate the Pauline parallels. Haupt argues that Tertullian often translated from Greek into Latin as he cited. That interpretation helps explain the evidence, without settling the cause of every individual difference.
English translations of the displayed excerpts are supplied here. Differences of punctuation, capitalization, and the spelling of Latin u and v are not independently counted as variation.
Tertullian’s dependence on Irenaeus is demonstrable, extensive, and creatively managed. He preserves long sequences of unusual information while translating freely, inserting objections, changing the setting, omitting scriptural demonstrations, and moving whole blocks. This supplies a documented model for possible reuse of an earlier anti-Marcionite work. It does not, on its own, prove that a particular lost work supplied Against Marcion.
A preliminary correction is necessary: the surviving source compared here is Irenaeus’s Against Heresies (Adversus haereses), Book I, chapters 1–12, not a surviving Against Marcion I.1–12. The substantial continuous exposition in Tertullian chiefly follows I.1–7, with selected material from I.11–12 in its appendix. The omissions from I.8–10 are themselves important evidence.
Method. “Line by line” here means identifiable clauses and short argumentative units, cited by chapter and paragraph, rather than the physical lines of one edition. The tables follow all thirty-nine Tertullianic chapters and distinguish retention, translation, addition, omission, inference, and rearrangement. English descriptions and renderings below are new; short Greek and Latin anchors make the comparison checkable. AV means Tertullian, Adversus Valentinianos; references beginning I. mean Irenaeus, Against Heresies, Book I.
1. Chapters 1–6: Tertullian constructs a new frame
The continuous adaptation begins in chapter 7. The first six chapters establish the reader, the enemy, and the literary method. Some themes overlap Irenaeus’s preface, but these chapters should not be described as a continuous translation of Irenaeus I.1–12. “Addition” below means absent from the corresponding extant passage, not necessarily invented by Tertullian.
Irenaeus treats the system as concealed falsehood and an abuse of biblical language.
Etiam solummodo demonstrare destruere est: merely exposing it can destroy it. Serpent imagery and a nursery-story comparison make disclosure itself the attack.
Recasts the purpose of the dossier. The compiler’s voice supplies a new argument around inherited information.
Irenaeus already includes ridicule, including jokes about waters and a parody using vegetables.
Tertullian announces preliminary exposition, with puncturing attacks and laughter whenever the material deserves it.
States the adaptation’s tone. His originality lies in its particular language and staging, not in inventing anti-Valentinian satire.
2. Chapter 7: a close comparison of the opening, clause by clause
This is the clearest small-scale demonstration. Almost the whole doctrinal sequence comes from one Irenaean paragraph, I.1.1. Tertullian repeatedly interrupts that sequence and then resumes it. The Greek anchors identify the source; the Latin anchors identify what he does to it.
The system begins in invisible, ineffable heights: ἐν ἀοράτοις καὶ ἀκατονομάστοις ὑψώμασι.
Before introducing the Aeon, he adds Ennius’s celestial upper rooms, stacked heavens, stairs, rented accommodation, and the tall Roman apartment building called the Insula Felicula.
Expands a few source words into an architectural satire. This extended scene requires no additional Valentinian document.
He counts the same eight, then describes sibling marriages, primordial cells, and a crowd whose divinity is inseparable from alleged criminality.
Preserves the numerical conclusion and attaches a hostile interpretation to it.
3. Chapters 8–13: lists, Sophia’s crisis, and the collective Saviour
The borrowing continues across different kinds of material: catalogues, alternative reports, theological instruction, and narrative summary. Names of Horos are pulled forward from a later source paragraph; most of the source’s following biblical demonstrations are cut.
Two catalogues give the ten and the twelve names, in pairs.
He follows the catalogues closely. Several printed readings, however, have been restored or corrected by editors using Irenaeus.
Preservation with a textual qualification: an editorial restoration cannot count as independent evidence of Tertullian’s wording. See the textual note below.
Another account: Sophia’s impossible undertaking produces a formless substance, οὐσίαν ἄμορφον, female in character.
Sophia herself is first described as disfigured, pale, thin, and neglected; she then conceives and produces a female without a consort.
Recasts an abstract account as a bodily and emotional scene. The descriptor shifts in presentation; the table does not presume a different source text.
Monogenes produces Christ and the Holy Spirit to stabilize and instruct the Aeons.
He keeps their production and shared task, then ridicules a pairing of “two males” unless the Spirit is made female.
Creates a sexual objection using Latin Spiritus, grammatically masculine; Greek πνεῦμα is neuter. This is rhetoric, not fresh evidence about the school.
The Father’s incomprehensibility secures permanence; what is comprehensible explains generation and formation. The Spirit teaches gratitude and repose.
He repeats the distinctions, then argues that the account makes not comprehending God advantageous. Gratitude and repose follow.
Turns the source’s theological distinction into an apparent paradox, then returns to the source sequence.
Biblical proofs intervene before the account moves outside the Pleroma.
A theatrical intermission—applaud, or rather learn and reject—leads directly to the next scene outside the curtain.
Replaces the intervening proof-text section with a new transition and jumps to I.4.1.
4. Chapters 14–19: Achamoth, matter, and the Demiurge
Here the source is repeatedly turned into a contemporary scene. The distinction between inherited satire and new satire matters especially in the chapter about waters.
The source’s waters are discussed by type rather than by these three named examples.
He adds Nonacris, Lyncestis, and Salmacis, with traditional stories of lethal, intoxicating, and feminizing waters, then imagines stored rain as her tears in our cisterns.
Classical and domestic illustrations. Their presence proves elaboration, not independent inquiry into Valentinian scriptures.
Painting a mule from a donkey becomes a joke about deriving Ptolemy from Valentinus.
New comparison doubles as a school-history comment; it adds no new cosmological datum.
5. Chapters 20–25: creation, ignorance, elements, and Adam
These chapters show both close retention and tendentious completion. The fire passage is particularly useful because Tertullian identifies his own proposed addition as an argument he is supplying.
Fire inheres in all three elements, as ignorance inheres in the three passions. The source thus gives an account of fire’s relation to them.
Tertullian says they have not identified its particular originating passion; ego argumentabor, “I shall supply an argument”: perhaps her agitations included a fever.
Reframes the source’s explanation and adds a marked comic conjecture. Do not claim that Irenaeus said nothing about fire.
The components are listed as soul from the creator, earthy body from mud, flesh from matter, spirit from Achamoth.
He starts with Achamoth’s contribution, then gives the psychic, earthy, and fleshly components, ending with a fourfold Geryon.
Changes list order and adds a monster comparison. The difficult printed phrase attached to the earthy component should not carry an argument about a lost source.
6. Chapters 26–32: Christ, human kinds, conduct, and the end
Here the rearrangement becomes structural. Tertullian moves through I.6.1, I.7.2–5, then back to I.6.2–4 and finally I.7.1. His own announcement in chapter 29 makes the collection of dispersed material explicit.
The Saviour takes spiritual and psychic elements and a specially constructed visible body; nothing material, because matter cannot be saved.
The components and exemption remain. He argues that those needing salvation should be precisely its beneficiaries and links the teaching to denial of our flesh’s salvation.
Adds his own soteriological objection to the reported exclusion.
These spirits become brides of the Saviour’s attendant angels.
A staged series of questions culminates in “brides.” He adds Sabine abduction and imagines a Roman Marcus or Gaius, a bearded father and grandfather, assigned this role.
Dramatic enlargement. The named Romans are illustrative characters, not extra Valentinian witnesses.
The source’s end-time account does not conclude with Tertullian’s personal destination and the following scriptural contrast.
He imagines the divine figures taking offence and contrasts being stripped with being “clothed over,” echoing 2 Corinthians 5.
Ends the adapted narrative with his own theological identity and biblical application.
7. Chapters 33–39: a rearranged appendix of alternatives
The variant catalogue is not evidence of seven new investigations. It is largely a new ordering of material in Irenaeus I.11–12. References here follow the familiar modern/ANF subdivisions: Secundus is I.11.2 and the final Christ alternatives are I.12.4.
The source has displayed variant systems and criticized their inventions.
His conclusion makes the luxuriant maternal seed grow into a forest of Gnostic teachings.
Supplies a new closing image for the rearranged catalogue.
8. What disappears: the omissions are part of the evidence
A comparison restricted to verbal parallels would miss one of the most consequential editorial decisions. Tertullian transmits much of the cosmology while removing large portions of the scriptural argument that Irenaeus reports for it. The resulting treatise gives the reader a substantially different ratio of myth, exegesis, and refutation.
An extended reading of the Johannine prologue to establish the primordial Ogdoad.
No equivalent exposition appears in AV. A reader with only Tertullian’s adaptation would lose this substantial evidence of the reported group’s use of John.
A substantial account attributed to Valentinus himself, including two boundaries and distinctive Christological options.
Not reproduced as the opening of the variant catalogue. This matters because the surviving adaptation does not evenly represent every strand in its source.
Irenaeus’s Monotes parody using a gourd, emptiness, cucumber, and melon.
The parody is not reproduced as a block, although Tertullian supplies numerous other jokes. Silence here cannot make Irenaeus an originally humourless expositor.
Irenaeus’s theological answer about God’s unity of thinking and willing, with a Homeric comparison.
The adjacent two-consort theory is retained in AV 33, but this source refutation is not reproduced.
The Johannine example is the strongest control for arguments from silence. The presence of the prologue exposition in Irenaeus I.8.5 is positive evidence of the use that Irenaeus reports. Its absence from Tertullian cannot overturn that evidence. If the source had been lost and only the adaptation survived, the exegetical dossier would have been difficult to recover. That is a concrete problem to bear in mind when reconstructing Marcion’s Gospel or Apostolikon from selected polemical discussions.
9. The demonstrable change of order
Tertullian’s new sequence
Irenaeus’s source sequence
Consequence
AV 26
I.6.1
Three kinds and the Saviour’s composite nature.
AV 27–28
I.7.2–4
Incarnation, passion, prophetic inspiration, and the creator’s education come next.
AV 29
I.7.5, then back to I.7.3
A thematic collection of human kinds and privileged souls.
AV 30
Back to I.6.2–4
Conduct and the claim to salvation by nature are relocated.
AV 31–32
Back to I.7.1, supplemented by I.7.5
The consummation is held for the end of the main narrative.
The appendix alternates between source chapters while preserving many smaller sequences within the moved units.
Two statements expose the editorial procedure from inside the finished work. In AV 29.1 Tertullian says colligam nunc ex disperso, “I shall now gather from scattered places.” In AV 33.1 he explains that he deferred material so that its insertion would not disrupt the order or the reader’s attention. The primary comparison shows what that means: he preserves many short sequences but changes the architecture of the whole account. An argument that assumes the surviving order is automatically the source’s order would fail this control case.
10. Translation, inherited polemic, and textual restorations
The Greek source and the old Latin Irenaeus must be distinguished
The dependence on Irenaeus does not settle the language of the copy Tertullian used. The old Latin translation of Irenaeus is usually much more literal than Tertullian. Shared expressions such as the medical term appendix, or the reintroduction of Sophia’s name where Latin gender makes a referent awkward, do not by themselves prove dependence on that Latin translation. Riley treats direct translation from Greek as sufficient to explain the evidence, while acknowledging that use of the Latin version cannot be disproved. The safe result is literary dependence on Irenaeus; the intermediary language is a separate question.
This distinction matters for a hypothetical lost anti-Marcionite source. A Latin expression may be Tertullian’s choice of rendering rather than the wording of Irenaeus or Marcion. AV 27’s per virginem non ex virgine is a good control: the forceful Latin contrast explicates an already reported passage through Mary. It need not be a separately documented Valentinian slogan.
Irenaeus is already an adversary
One must qualify a simplifying contrast in older accounts of the dependence. Irenaeus is not merely a neutral reporter whom Tertullian transforms into a polemicist. Irenaeus I.4.3–4 already mocks the derivation of waters, and I.11.4 constructs a vegetable parody of the Aeons. He also uses objections, direct address, and hostile classifications. What the comparison demonstrates is Tertullian’s specific amplification, substitution, and relocation of such devices. It does not license assigning every joke or objection to him.
The apparatus prevents circular proofs
In Riley’s Latin apparatus at printed p. 35, line 1, the words from Ageratos through Autophyes in AV 8.2 are identified as a supplement by Pamelius from Irenaeus, absent from the principal witnesses designated P, M, and X. The same apparatus records another name corrected from Irenaeus. At printed p. 70, the apparatus records source-based additions in AV 37, including monada and singularitas. These may be persuasive restorations, but they are not independently transmitted matches. The dependence is secure from the surrounding structure and numerous uncontroversial correspondences; it should not be “proved” by words editors inserted from the proposed source.
References have also been normalized against the actual passages. In the usual chapter-and-paragraph numbering, AV 38’s Secundus notice corresponds to I.11.2, and AV 39’s Saviour catalogue to I.12.4. Riley’s online chapter labels give I.11.1 and I.12.3 at those two points; his introduction also places the opening hunc substantialiter passage in chapter 6, although it is in AV 7.3. Such slips are reasons to check a useful guide against the texts, not reasons to dismiss the documented dependence.
11. What this permits us to say about Irenaeus’s unavailable anti-Marcionite work
The ancient evidence supplies a real starting point. In Irenaeus I.27.4, he announces a special refutation of Marcion using Marcion’s own writings and the dominical and apostolic texts Marcion accepts. In III.12.12, he again promises a separate refutation from the portions retained by Marcion’s followers. Eusebius, Church History V.8.9, reports this promise. Eusebius IV.25 also names Irenaeus among writers who wrote against Marcion. That is positive evidence for his anti-Marcionite literary activity, but the notice supplies neither a quotation nor enough description to distinguish securely between the separate projected work and the extant anti-Marcionite polemic. The projected method is explicit; the separate book’s contents are unavailable for a matching comparison.
If that work was completed and reached Tertullian, AV shows exactly the kinds of reuse available to him. He could retain an older author’s doctrinal dossier and biblical arguments, shorten or expand them, reframe them in Latin, address Marcion as a present opponent, and integrate them with his own theological interests. The resulting work could sound immediate and original without all its underlying research being independent. That is a historically grounded possibility, not an inference from stylistic resemblance alone.
Observed operation in AV
A legitimate test in Against Marcion
What it would not establish by itself
Long source sequences survive beneath new rhetoric.
Separate the stable doctrinal or exegetical sequence from jokes, apostrophes, objections, and contemporary comparisons; compare the sequence with extant Irenaeus and other early witnesses.
That every stable sequence must come from one particular lost Irenaean book.
Whole source sections and biblical quotations are omitted.
Treat non-citation of a passage differently from an explicit claim that Marcion’s text lacked it. Ask whether the passage was relevant to the excerpted argument.
That a positively attested deletion is necessarily fictitious; positive textual claims need their own assessment.
Inherited identifications remain after their supporting quotations are dropped.
Distinguish a report of an interpretation from the exact biblical wording that allegedly supported it. AV 28’s centurion is the control example.
The exact reading of a Marcionite Gospel merely from the interpretation attached to it.
An abstract source claim becomes a sharper Latin formula.
Test whether prepositional, terminological, or argumentative oppositions are Tertullian’s exposition of a doctrine or actual quotations from an opponent.
That a polished Latin antithesis reproduces the words of a Greek Marcionite source.
New first-person comments occur inside borrowed passages.
Ask what “I find,” “we know,” “you say,” and similar expressions actually claim: reading, inference, hypothetical confrontation, or a real encounter.
That all first-person testimony is artificial. Explicit claims of observation remain evidence to evaluate, as AV 30’s eunuchs show.
The compiler explicitly gathers scattered material.
Check whether repeated discussions or shifts of order reflect thematic compilation rather than several different Marcionite copies.
The order of Irenaeus’s lost work, or Marcion’s own order, from Tertullian’s finished order alone.
The source’s qualification becomes a charge of contradiction.
Before assigning an inconsistency to Marcion, ask whether it arose from the summarizer’s treatment of agency, categories, or an acknowledged alternative.
That Marcion was consistent on every point; it supplies a diagnostic, not a universal acquittal.
The writer labels some additions as conjecture.
Keep deductions and sarcasm separate from reports and quotations when reconstructing an opponent’s text.
Permission to reconstruct a lost document by collecting every hostile inference as if it were a lemma.
Some information lies outside the demonstrably borrowed passage.
Allow a combined model: earlier anti-Marcionite research plus Tertullian’s own reading, information, or inference.
That literary dependence excludes access to Marcionite texts, or that such access makes dependence disappear.
Three propositions with different evidential status
Demonstrated: Tertullian used and extensively rewrote Irenaeus in Against the Valentinians. We can show his procedures because the corresponding source survives.
Plausible and testable: a completed anti-Marcionite work by Irenaeus could have been reused in a similar manner. The acknowledged use of Irenaeus elsewhere makes the proposal concrete, and Irenaeus’s announced project is relevant evidence.
Not demonstrated by this comparison: that the unavailable work was completed, that it was the principal or sole source of Against Marcion, that particular uncheckable sentences came from it, or that Tertullian never inspected Marcionite texts. Those claims require further evidence.
A useful next comparison would collect exact parallels between Tertullian’s Against Marcion and the surviving Irenaean anti-Marcionite passages, distinguish them from commonplace objections, and then compare other witnesses for shared unusual order, wording, and errors. Independent wording or explicit claims about physical copies would have to be tested on their own merits. Simply assigning all apparent primary-text evidence to the lost book would make the model impossible to disprove.
12. Conclusions
The securely documented model is not mere copying. Tertullian retains a research core and produces a new polemical work around it. The most diagnostic evidence is the repeated preservation of unusual sequences: generations, consorts, titles, near-dissolution and rescue, staged formations, invisible deposits of seed, the four-component Christ, and the catalogue of alternative Saviours. Around that core he supplies Roman buildings, magistrates, marriage law, theatre, named waters, direct questions, confessional comments, and new deductions.
The absences matter equally. He can transmit a doctrine while deleting its biblical demonstrations; preserve an interpretation while dropping the quotation; introduce a logical difficulty while compressing its qualifications; and combine material that originally stood apart. Consequently, Tertullian’s finished polemic cannot automatically be treated as an independent transcript of his opponent’s book. Its source-bearing clauses have to be distinguished from its editorial and argumentative clauses.
The strongest conclusion for the Marcion question is methodological: Against the Valentinians supplies a demonstrable precedent for extensive Irenaean mediation, including a precedent for the loss of biblical evidence through abbreviation. It makes the proposed reuse of Irenaeus’s anti-Marcionite research credible enough to investigate passage by passage. It does not supply the missing source, and it does not settle whether Tertullian also consulted Marcionite texts.
Texts and references
Mark T. Riley, Q. S. Fl. Tertulliani Adversus Valentinianos: Text, Translation, and Commentary (Stanford University PhD dissertation, February 1971), hosted by Roger Pearse: index; introduction; Latin and apparatus; English translation; commentary and bibliography. Particularly useful: introduction pp. 7–16; commentary under each chapter; apparatus pp. 35 and 70. This study’s comparative descriptions and translations are new.
Irenaeus, Against Heresies, Book I, chapters 1–12: complete Book I in the Roberts–Rambaut translation. Tables link to the corresponding New Advent chapter where possible. Chapter and paragraph references follow the familiar divisions used in these translations.
W. W. Harvey, Sancti Irenaei episcopi Lugdunensis libros quinque adversus haereses, vol. I (Cambridge, 1857), Greek fragments and ancient Latin translation. Greek text catalogue; digital Greek transcription. Harvey’s own internal divisions differ; the citations in this study are normalized rather than copied from its XML numbering.
Irenaeus, I.27.4 and III.12.12; Eusebius, Church History V.8.9, for the announced separate anti-Marcionite refutation; also IV.25 for Irenaeus among anti-Marcionite writers.
Prepared 7 November 2026. 129 numbered clause/unit comparisons. This HTML may be read directly, printed, or used as the basis of a Blogger post; the article body uses ordinary headings, paragraphs, tables, and links.
