The agreements do not all have the same origin or evidential value. Several belong to an older argumentative tradition already visible in Irenaeus. Others concern particular reported readings and remain relevant to reconstructing Marcion’s texts. The strongest result is that Tertullian–Epiphanius agreement cannot be treated automatically as two independent inspections of a Marcionite manuscript. The identity of any intermediary, and the contribution of Irenaeus’s projected separate refutation, remain questions to be established.
This study uses Williams’s annotations as a guide, checks the Irenaean antecedents, and separates reports of readings from the arguments constructed around them. It is a selected comparison of diagnostic cases, not a numerical census of every citation or a stylometric demonstration. AH means Irenaeus’s Adversus haereses; AM means Tertullian’s Adversus Marcionem. Gospel scholion numbers belong to Panarion 42.11; Pauline scholia and their refutations belong to 42.12.3. Dual Pauline numbers, such as 9/17, retain the two numberings displayed by the edition.
Williams’s annotations as a source guide
Williams’s opening note distinguishes a reconstructed Hippolytan framework, Irenaean additions, Eusebian information, and Epiphanius’s earlier canonical dossier. These are distinct source claims. His references to Pseudo-Tertullian concern the Adversus omnes haereses epitome, not Tertullian’s five-book Against Marcion. A comparative “cf.” is not itself a demonstration of direct borrowing.
| Topic | Williams 2009 locator | Earlier material indicated | Question for this investigation |
|---|---|---|---|
| Opening source assessment | p. 294 n. 1 | Hippolytan Syntagma; Irenaeus; Eusebius | A model of compilation; not a claim that every biblical lemma comes from Irenaeus. |
| Origin and theology | p. 294 n. 2; p. 296 nn. 8, 10–11, 13 | AH 1.27.1–2 | Distinguish inherited descriptions of a teacher from the wording of his books. |
| Soul-only salvation and rejection of the older scriptures | p. 297 nn. 17, 19 | AH 1.27.1, 3 | Doctrinal parallels are not automatically textual variants. |
| Inverted salvation in Hades | p. 298 n. 21 | AH 1.27.3 | Compare the named persons and the explanation, including alterations. |
| Arguments concerning divine territories | p. 298 n. 26; p. 300 nn. 27–29 | AH 2.1.2–5 | A particularly useful control for adaptation of an earlier argument. |
| Mutilated Luke | p. 302 n. 32 | AH 1.27.2 | The general identification and mutilation charge already belong to earlier heresiology. |
| Refutation from the opponent’s accepted writings | p. 302 n. 33 | AH 1.27.4 | Compare the program before assuming that its repetition announces a new investigation. |
| Chronology within the Pauline refutation | p. 352 n. 206 | AH 3.4.2–3 | Earlier heresiology also appears inside the later scriptural argument. |
The first column identifies the topics; the locators make the annotations checkable. The final column states this study’s source-critical use of them. The following comparisons are an independent assessment of the ancient passages, not an attribution of all the proposed relationships to Williams.
A demonstrable adaptation of an argument
The clearest control is the discussion of spatially limited divine powers. In AH 2.1.2–3, a deity bounded by another reality fails to contain everything and is inferior to the reality containing it. A Pleroma and what lies outside it require a separating third reality, producing a regress. In 2.1.5, the allocation of separate realms creates a further problem for the alleged antagonism between their possessors.
Epiphanius 42.7.3–7 takes over this reasoning while fitting it to his account of three principles. Consequently the further superior authority becomes a fourth. This is an observable adjustment of an inherited argument. It is more informative than two authors quoting the same biblical sentence, because the argument does not arise from the wording of Marcion’s Gospel or Paul. The number changes to fit the new target while the logical machinery survives.
Tertullian AM 1.15.2–4 belongs in the comparison: a place containing the higher god must exceed him, and the place itself becomes a claimant to divinity. His short formula maius est id quod capit eo quod capitur, “what contains is greater than what is contained,” expresses the common logic. Tertullian does not reproduce every step of Epiphanius’s revised sequence. The evidence supports an inherited argumentative complex more securely than one particular stemma.
Comparison of the passages and their evidential weight
| Case | Ancient references | Agreement and material difference | Assessment |
|---|---|---|---|
| 1. A containing realm greater than its god | AH 2.1.2–3; AM 1.15.2–4; Pan. 42.7.3, 6–7 | All three arguments make spatial enclosure compromise divine supremacy. Tertullian also makes the enclosing place a competitor for divinity. Epiphanius adds a superior distributor of the three gods’ territories. | Strong inherited argumentative complex. No Marcionite biblical reading is needed to generate this agreement. The precise route to each later author remains a separate question. |
| 2. A third boundary becomes a fourth authority | AH 2.1.3, 5; Pan. 42.7.4, 6–7 | Irenaeus’s argument against a Pleroma and what lies outside it introduces a third separating reality. Epiphanius, starting from three principles, asks for a fourth surveyor or boundary-setter. He also uses the argument from peaceful allocation of territories. | Especially clear Irenaeus–Epiphanius adaptation. Tertullian 1.15 shares the containment argument, not the whole rewritten third-to-fourth sequence. Do not turn this into a three-way verbal identity. |
| 3. The salvation of Cain and exclusion of the righteous | AH 1.27.3; Pan. 42.4.3–4 | Both reports reverse the expected salvation of scriptural figures and begin the excluded group with Abel, Enoch, and Noah. Epiphanius expands the lists; his rationale is not a verbatim reproduction of Irenaeus’s explanation that the righteous suspected another trial by their god. | Inherited heresiological report with editorial change. It is not a quotation from the Gospel or Paul. No corresponding identical catalogue in Tertullian is being claimed here. |
| 4. Saving another creator’s property | AH 5.2.1; AM 1.23.1–9; Pan. 42.6.1–2; 42.8.1–5 | The objection treats the salvation of another maker’s humanity as a problem of ownership and justice. Epiphanius develops it through the distinction between buying and redeeming; Tertullian disputes the rationality of goodness directed to another’s human creation. | Shared anti-dualist reasoning. Irenaeus supplies a surviving antecedent. This does not establish that the whole later discussion was copied from one particular lost treatise. |
| 5. Goodness and justice belong together | AH 3.25.2–4; AM 5.13.14–15; Pan. 42.6.4–7 | Irenaeus joins saving goodness to judgment. Epiphanius argues that a just judge must also be good and reward goodness. Tertullian applies the combination to Romans 7:12 and the law’s just and good commandment. | Older polemical framework, differently applied. The argument can shape the selection and interpretation of a Pauline lemma; it is not itself a special Marcionite reading. |
| 6. Refuting Marcion from what he retains | AH 1.27.4; AH 3.11.7; AH 3.12.12; AM 4.6.2; AM 5.4.2; Pan. 42.9.5–6; 42.10.2–7 | The reduced Gospel and Pauline collection are made to defeat their possessor. In each later author, the surviving material is represented as sufficient to establish continuity with the Creator. | A documented prior program. The method alone cannot distinguish inherited research from fresh collation, and does not cancel Epiphanius’s express claim of consultation. |
| 7. Lord of heaven | AH 4.2.2; AM 4.25.1; Pan. 42.11, Gospel scholion and elenchus 22 | All use the lordship of heaven to identify Christ’s Father with the Creator. Irenaeus gives the fuller saying with Father and earth. Epiphanius explicitly alleges their omission; Tertullian’s displayed quotation is shorter. | Shared argument; uneven textual testimony. A short citation in Tertullian is not equivalent to an explicit statement that both words were absent from his exemplar. |
| 8. Abraham’s daughter and Sabbath healing | AH 4.8.2; AM 4.30.1; Pan. 42.11, Gospel scholion and elenchus 39 | Irenaeus and Epiphanius infer that care for Abraham’s daughter shows that Christ is not alien to Abraham. Tertullian’s immediately comparable discussion defends the lawfulness of the Sabbath cure through the ox-and-ass saying. | Close Irenaeus–Epiphanius exegesis; partial Tertullianic parallel. Tertullian must not be credited here with repeating the same daughter-of-Abraham inference. |
| 9. The righteous and the kingdom | AH 4.8.1; AM 4.30.4–6; Pan. 42.11, Gospel scholion and elenchus 40 | Irenaeus argues from the patriarchs’ presence in the kingdom. Tertullian and Epiphanius both use a generalized reference to righteous people, but Tertullian describes entry; Epiphanius insists on their being inside rather than entering. | Shared controversy and a possible shared textual feature, with a material difference. Do not reconstruct one identical Greek sentence or one identical argument from the agreement. |
| 10. A testimony to you, followed by the fulfillment gloss | Luke 5:14; AM 4.9.9–10; Pan. 42.11, Gospel scholion 1 and elenchus 1(d) | Both have the second-person testimony formulation and connect it with Christ’s fulfillment of the Law, invoking the thought or wording of Matthew 5:17. | A valuable quotation-and-explanation cluster. It could travel in an earlier exegetical dossier. No exact surviving Irenaean source for this complete cluster has been identified in this study. |
| 11. Calling instead of judgment | Luke 11:42; AM 4.27.4; Pan. 42.11, Gospel scholion and elenchus 26 | Tertullian has vocationem; Epiphanius reports κλῆσιν in place of κρίσιν. This is a specific lexical agreement, not just a common doctrinal conclusion. | A genuine candidate for an inherited reading. It could come from Marcionite copies or a shared report of their text. A matching reading in surviving Irenaeus has not been established. |
| 12. The rock was Christ | AH 4.27.3; AM 5.7.12–14; Pan. 42.12.3, 1 Corinthians scholion and elenchus 9/17 | The wilderness narrative in 1 Corinthians 10 establishes continuity between Christ and the God active in Israel. The common material extends beyond the isolated identification of the rock. | An earlier exegetical dossier is plausible. The sequence within Paul is not an independently distinctive sequence, and the inference is readily suggested by Paul’s text itself. |
| 13. The body dies and therefore rises | AH 5.7.1–2; AM 5.9.3–4; Pan. 42.5.1–7 | Irenaeus distinguishes mortal bodies from souls and uses the sown grain. Tertullian and Epiphanius sharpen the argument through what falls or dies and can consequently rise. Their linguistic elaborations differ. | A shared resurrection argument, not a demonstrated variant. Irenaeus supplies the conceptual cluster; the exact etymological form should not be attributed to him without additional evidence. |
| 14. Bones in the risen Christ | AH 5.2.3; AM 4.43.6–7; Pan. 42.11, Gospel scholion and elenchus 78 | Tertullian and Epiphanius display a shortened bones formulation; Irenaeus’s corresponding argument mentions both bones and flesh. Epiphanius’s explanation itself also speaks of both. | Shared bodily-reality polemic; uncertain origin of the shorter citation. This comparison does not show that Marcion deleted flesh, or that surviving Irenaeus furnished the shorter wording. |
| 15. Ephesians and Laodiceans | AM 5.17.1; Pan. 42.9.4; 42.11.7–10; 42.12.3, Laodiceans scholion 40 | Tertullian treats Laodiceans as the changed title of the letter the church calls Ephesians. Epiphanius lists Ephesians among ten and then adds Laodicean material, including words identifiable as Ephesians 4:5–6. | Agreement on a name is not agreement on the reported collection. An extra separate book, a duplicated dossier entry, and a retitled letter cannot be silently equated. No surviving Irenaean explanation for this discrepancy has been identified. |
Two quotation clusters that deserve particular weight
Luke 5:14 is a useful test of transmission together with interpretation. Tertullian’s vobis and Epiphanius’s ὑμῖν make the testimony second person. Both then connect the healing and the Mosaic offering to the fulfillment of the Law, invoking Matthew 5:17. A Marcionite exemplar could explain a shared reading. By itself, that exemplar does not explain why the same added scriptural gloss accompanies it. A transmitted lemma-and-commentary unit is therefore a serious alternative. The comparison does not identify its author. Epiphanius uses Matthew 5:17 in his explanation; this must not be redescribed as his explicit testimony that Marcion removed that sentence from Luke.
Luke 13:28 combines agreement with an important disagreement. Irenaeus uses the named patriarchs to defend Abraham’s salvation. Tertullian speaks of righteous people entering the kingdom, while Epiphanius reports righteous people in the kingdom and expressly bases his argument on their being there already. Both later reports can preserve a generalized wording behind their polemic; they do not attest exactly the same sentence. Their arguments also differ: Tertullian links exclusion, punishment, and kingdom to one God, while Epiphanius recovers the patriarchs behind the general designation and adds the implication of bodily teeth. This case cannot be reduced either to pure copying or to two identical independent transcriptions.
Controls against overextending the Irenaean hypothesis
| Passage | Earlier comparison | What the Epiphanian control shows |
|---|---|---|
| Galatians 2:5 | AH 3.13.3; AM 5.3.3–5 | The positive yielding argument is not reproduced in Epiphanius’s Galatians entries. It cannot be counted as a three-author agreement here. |
| 1 Corinthians 15:49 | AH 5.9.3; AH 5.11.2; AM 5.10.10–11 | The specifically hortatory image-bearing argument is not reproduced in Epiphanius’s resurrection entry 16/24. The common subject of resurrection does not supply the missing argument. |
| 2 Corinthians 4:4 | AH 3.7.1–2; AM 5.11.9–11 | The shared repunctuation in Irenaeus and Tertullian is not reproduced in Panarion 42’s 2 Corinthians entries, which select 1:20, 4:5–6, and 4:13. It is not an established three-way parallel. |
| Luke 24:25 | AM 4.43.4; Pan. 42.11, Gospel scholion 77 | The reports agree in moving away from an explicit reference to the prophets, but Tertullian has third-person speech and Epiphanius first-person speech. Roth’s 2022 article, nn. 54–56, is an important check against harmonizing these reports into one exact Greek reading. |
The three Pauline controls matter because the preceding comparison of Tertullian with Irenaeus identified them as particularly diagnostic. Their absence from the corresponding Epiphanian arguments prevents their use as proof that all three authors inherited one continuous dossier. Silence does not disprove use of an earlier source; selection and abbreviation remain possible. It simply supplies no positive third witness for those particular features.
The compiled character of the canonical study
The catalogue and the rebuttals should be analyzed separately. Epiphanius presents an earlier selection of lemmata and then supplies explanations of why those lemmata defeat Marcion. Agreement between the later explanations can arise from older polemic even when a lemma was independently checked. Conversely, a writer may transmit a useful earlier lemma while adding an explanation of his own.
The chronological material in the Pauline resurrection refutation is a useful illustration. The reference to Hyginus and Marcion’s lateness belongs to the heresiological tradition compared with AH 3.4.2–3. It is not information obtainable by reading 1 Corinthians 15. Its presence inside the Pauline exposition demonstrates that source-based argument and discussion of a scriptural passage can occupy the same unit. It does not establish that the Pauline lemma itself was copied from Irenaeus.
The change of Pauline order is also explicitly acknowledged. At 42.12.2–3 Epiphanius says that the earlier selection put Romans first according to the church’s order, and that he now presents the material with Galatians first according to Marcion’s order. That statement explains a feature of the composition which should not be treated as an unacknowledged contradiction. It neither proves that his selections came from another author nor authenticates every assertion about the collection.
Epiphanius’s claims at 42.10.2 and 42.11.16 remain evidence: he represents himself as having handled the books and compiled material from Marcionite copies. A shared source can coexist with such consultation. The test is whether an individual report has features best explained by textual inspection, an earlier excerpt, conflation, memory, or editorial inference. Williams’s framework does not authorize deleting the consultation claim, and no dishonesty is demonstrated merely by showing use of earlier literature.
Which source models survive the comparison
- Reuse of surviving Irenaeus. For the containment argument, the salvation catalogue, the prior refutation program, and some exegetical themes, an extant antecedent already exists. A lost work is not required to explain the existence of these themes.
- An Irenaean tradition through intermediaries. A handbook, excerpt collection, or later anti-Marcionite dossier could transmit selections and readings together. The reconstructed Syntagma is relevant to Epiphanius’s framework, but its role in any proposed route to Tertullian must respect chronology and cannot simply be assumed.
- The projected separate refutation of Marcion. Irenaeus’s announcements at 1.27.4 and 3.12.12 make this a concrete hypothesis to investigate. They do not supply the lost text, prove its completion, or identify a particular later sentence as its descendant.
- Shared polemic plus consultation of Marcionite copies. This combined model remains possible. Reading a manuscript does not make every argument original; using an older argument does not show that no manuscript was read.
The present evidence establishes inherited polemical material most securely. It supports testing a common textual dossier where an unusual reading and an unusual explanatory gloss occur together. It does not establish one exclusive intermediary or justify attributing every Tertullian–Epiphanius agreement to Irenaeus.
Passage for the paper
The agreement of Tertullian and Epiphanius cannot be assumed at the outset to constitute the agreement of two independent readers of Marcionite manuscripts. Epiphanius’s treatment incorporates earlier heresiology, and the surviving work of Irenaeus provides controls for recognizing its adaptation. The argument against spatially limited gods in Panarion 42.7.3–7 reworks the containment and boundary argument of Adversus haereses 2.1.2–5, even converting the further separating authority from a third into a fourth to accommodate the new presentation. Tertullian’s corresponding containment objection shows that this argumentative complex was already available to the anti-Marcionite tradition. Such agreement concerns inherited polemic rather than a distinctive feature of Marcion’s canon.
The scriptural agreements require a further distinction. At Luke 5:14, the shared second-person testimony formulation travels with the same fulfillment-of-the-Law explanation; at Luke 13:28, agreement in the reference to righteous people coexists with a difference between entering the kingdom and being within it. The former raises the possibility of a transmitted quotation-and-commentary unit, while the latter warns against constructing an identical text from related reports. The evidential weight of a particular reading therefore depends on identifying what each witness actually attributes to Marcion and discounting what may derive from their inherited explanatory tradition. This procedure preserves genuine textual evidence while preventing a shared polemical source from being counted twice as independent corroboration.
Sources and verification
Primary references in Table 2 identify the passages required for checking the comparisons. The readings are not reconstructed by translating Williams’s English back into Greek. Short original-language expressions are used only where needed; English descriptions of the comparisons are new.
- Frank Williams, trans., The Panarion of Epiphanius of Salamis, Book I (Sects 1–46), 2nd ed. (Leiden: Brill, 2009), Panarion 42, especially the pages and annotations identified in Table 1. Consultable combined PDF. Printed page numbers are used; PDF links include the preliminary-page offset.
- Karl Holl, ed., Epiphanius, vol. II, Panarion 34–64, GCS 31 (Leipzig: Hinrichs, 1922), heresy 42. Open Greek and Latin digital transcription. Some digital Greek has OCR defects; no inference is based on an isolated apparent OCR form.
- Irenaeus, Adversus haereses, passages linked in the tables. The New Advent reference translations are Roberts–Rambaut, Ante-Nicene Fathers 1 (1885). The argument concerns the surviving content; verbal conclusions across Greek and the ancient Latin version require separate textual control.
- Tertullian, Adversus Marcionem, Latin text linked by book in the tables. Ernest Evans edition and translation (1972), online index. The linked Latin Library text provides the chapter and subsection divisions used here.
- Dieter T. Roth, “The Testimony for Marcion’s Gospel in NA28: Revisiting the Apparatus to Luke in the Light of Recent Research,” New Testament Studies 68.1 (2022): 52–60, especially nn. 22 and 54–56. DOI and article record. These notes supply a check on citation abbreviation and the differing testimony to Luke 24:25, not an endorsement of the source genealogy proposed here.
The source genealogy is argued from identifiable parallels and differences. No percentage of dependence, exhaustive stemma, or certainty about the contents of an unavailable Irenaean work is claimed.
