HistoricalChristian.Faith

Were the composite citations drawn from a pre-formed testimonia collection?

Type: supporting argument (off-timeline) · Assessment: plausible — that the composite citations, and their single (sometimes wrong) attributions, were not freshly fused by the Evangelist but inherited from a pre-formed collection of Christian proof-texts — a “testimony book” — is a real, genuinely-advocated mechanism (J. Rendel Harris) that is now underwritten by hard evidence: the testimonia genre is no longer conjecture but documented fact, since Qumran yielded an actual pre-Christian sheet of strung-together proof-texts (4QTestimonia), one of whose blocks is itself a pre-conflated composite. So the reading is coherent and not ruled out. What it cannot reach is demonstration for any single conflated citation — no surviving testimonia source contains Mark’s Malachi–Isaiah blend or Matthew’s Jeremiah-headed Zechariah oracle — and the strong form Harris built (one fixed written “Testimony Book” standing behind the whole New Testament) is the part the field has set aside, on C. H. Dodd’s showing that such collections were “the result, not the presupposition” of early Christian exegesis. Rightly, this mechanism does not rival the deliberate-composite reading but sits upstream of it: it relocates the deliberate fusion to a compiler rather than dissolving the design — and, since the Qumran collection carries no misattribution at all, it explains the conflation better than it explains the wrong prophet, which still needs the “name the senior prophet” convention or a genuine slip.

This argument weighs a mechanism that cuts across the whole doctrine rather than one crux: it asks not which prophet Mark or Matthew names but where the conflation came from — from the Evangelist’s own pen or from a source already carrying it. It is the classic Testimony Book / Testimonia hypothesis, and it surfaces as a live option on the doctrine’s own pages without being adjudicated there: the Mark 1:2 argument lists “a pre-formed collection of testimonies” among the open answers to why Isaiah (“in which case the grouping, not the naming, is what he inherited”), and the Matthew formula-quotations argument lists “a pre-existing collection of Christian testimonia” as one of four mechanisms behind Matthew’s non-Septuagintal proof-texts. The reception of the whole composite-citation phenomenon is traced on the Naming the Wrong Prophet timeline.

See also: the closest working parallel elsewhere on this wiki — the “excised” descent-prophecy (a logion Justin and Irenaeus quote as Scripture, its attribution wandering across Isaiah, Jeremiah, and no prophet at all), which that page reads as “a free-floating Christian testimonium drawn from a proof-text collection” — the same Isaiah/Jeremiah wandering that shadows Matthew 27:9.

The claim

J. Rendel Harris built his two-volume Testimonies on exactly the phenomenon this doctrine catalogues. Surveying the recurring “errors” of attribution in the Gospels and the fathers, he took a chapter to the paradigm case — heading it “Spoken by Jeremy the Prophet” — and argued that the misattribution is not the Evangelist’s slip but a fingerprint of his source:

The existence of a primitive book of Testimonies against the Jews, quoted by successive generations of controversialists… will explain a whole group of difficulties… When we find in the New Testament, or in the early Fathers, a group of passages quoted together, or a wrong ascription of a quotation, or a peculiar text-form… the natural explanation is that the writer is not going to the Scriptures direct, but is copying from a Testimony Book in which the passages were already grouped and (sometimes wrongly) labelled.

— J. Rendel Harris (with Vacher Burch), Testimonies (Cambridge, 1916–20).¹ On this reading, when Mark heads a Malachi–Isaiah blend “in Isaiah” and Matthew heads a Zechariah oracle “Jeremy,” neither is fusing texts at his desk: both are transcribing a proof-text that came to them already fused and already labelled, so that the conflation and the “wrong” name were in the collection before either Gospel was written.

Who advances it

The hypothesis is Harris’s (1916–20), anticipated in the nineteenth century by Edwin Hatch and others who noticed that the fathers’ Old Testament citations cluster and corrupt in shared ways. Harris pressed it hard: a single, unified, written Testimony Book — the “first known treatise on Christian theology” — lay behind the New Testament and the early apologists alike, and its tell-tale traces were precisely the grouped quotations, wandering ascriptions, and odd text-forms that this doctrine gathers. The idea was refined for Justin Martyr’s sources by Oskar Skarsaune, whose The Proof from Prophecy reconstructs the written testimony-traditions Justin demonstrably drew on — the same scholarship the descent-prophecy argument leans on. Standing against Harris’s strong form is C. H. Dodd, whose According to the Scriptures (his 1950 Stone Lectures) is the classic correction: the New Testament writers worked from whole fields of Scripture held in common, not from a fixed anthology of isolated proof-texts, so that “the composition of ‘testimony-books’ was the result, not the presupposition, of the work of early Christian biblical scholars.” Krister Stendahl reached the same verdict for Matthew in particular, rejecting a Testimony Book behind the fulfilment citations; L. W. Barnard and R. A. Kraft later argued an intermediate, evolutionary picture (oral clusters hardening into small written collections over time). The whole debate was then reopened, on the archaeological side, by the Dead Sea Scrolls.²

The evidence for — and the Qumran vindication of the genre

The single hardest fact in the hypothesis’s favour is that the genre is no longer conjecture. When Harris wrote, a pre-Christian “book of testimonies” was a paper hypothesis; Qumran turned it into an artefact. 4QTestimonia (4Q175), copied in Hasmonean script about 90 BC, is a single sheet that strings together five scriptural passages — Deuteronomy 5:28–29 and 18:18–19 (a prophet like Moses), Numbers 24:15–17 (Balaam’s star and sceptre), Deuteronomy 33:8–11 (the priestly blessing), and Joshua 6:26 (with an interpretation from an apocryphal “Psalms of Joshua”) — grouped around expected eschatological figures. It is, materially, exactly what Harris posited: a compact anthology of proof-texts excerpted from their scrolls and arranged by theme, a century before the Gospels. And its very first block is itself a composite: the compiler quoted a form of the Sinai material in which Exodus 20:21, Deuteronomy 5:28–29, and 18:18–19 had already been edited together into one continuous citation. So a Second-Temple testimonia sheet demonstrably (a) existed, (b) lifted texts out of context and grouped them, and (c) carried a pre-formed conflation of its own — the three moves the hypothesis needs.³

Christian testimonia collections are attested too. The third-century Testimonia ad Quirinum of Cyprian is an intact Christian proof-text book of just this kind, and Skarsaune’s reconstruction shows Justin Martyr drawing on earlier written testimony-sources rather than always going to the scrolls direct. The wandering attribution of the descent-logion in Justin and Irenaeus (Isaiah once, Jeremiah twice, unnamed thrice, never quite word-for-word) is the standard example of what citation-from-anthology looks like: a line that travels without secure provenance because it travelled in a collection, not a scroll. The same profile — a fixed wording under an unstable name — is what a testimonia source would predict for Matthew’s “Jeremy.”⁴

And the mechanism explains a real oddity economically. If the Evangelist is copying a labelled proof-text, then the conflation is a datum he received, not a liberty he took — which fits the ease with which Mark opens his Gospel with a mixed citation and moves on, and the fact that Matthew and Mark independently show the same convention (a composite headed with the senior prophet). It also fits modern composite-citation scholarship, which lists a testimonia source among the recognized ways ancient authors came by their combined quotations rather than fusing them fresh.⁵

Adversarial assessment

What holds. The testimonia genre is documented fact, not apologetic speculation: 4Q175 is a physical, pre-Christian sheet of grouped proof-texts, one block of which is a pre-formed composite, and Cyprian’s Testimonia is a Christian instance of the same thing. Skarsaune has shown at least one New Testament-adjacent writer (Justin) drawing on written testimony-sources, and the descent-logion is a working model of a conflated, wrongly-labelled line travelling in such a collection. So the claim that an Evangelist could have received a composite already fused and already (mis)labelled is coherent, precedented, and not ruled out. As an account of transmission, it is a live and respectable option.

What is overstated or unreachable.

1. No surviving source contains these composites. The hypothesis is an inference from the phenomenon, not a reading off a document. There is no extant testimonia collection — at Qumran, in Cyprian, or anywhere — that contains Mark’s Malachi–Isaiah blend headed “Isaiah,” or Matthew’s Zechariah oracle headed “Jeremiah.” The genre is proven; its application to these particular verses is postulated. That keeps the mechanism from rising above possible for any single crux.

2. Harris’s strong form is the part the field set aside. A single, fixed, written “Testimony Book” behind the whole New Testament outran the evidence, and Dodd’s reply is now the standard one: the shared groupings are better explained by a common exegetical tradition working over whole fields of Scripture — so that testimony-books are the deposit of that work, “the result, not the presupposition.” Stendahl rejected a Testimony Book for Matthew specifically. The moderate form — small, fluid collections and a shared proof-text culture, now backed by Qumran — survives; the maximal one does not. (This is the overreach to flag, not a rival thesis: the moderate reading is the interpretation being rated here.)

3. It relocates the design; it does not dissolve it — so it complements the deliberate-composite reading rather than defeating it. If Mark inherited a Malachi–Isaiah proof-text, someone still fused those texts on purpose and headed the blend with one name; the deliberate composite is real either way, only its author moves upstream from the Evangelist to a compiler. The two readings are not rivals with a winner to be declared but two locations for the same intentional act, and the evidence rarely lets us say which. On the mainstream “name the senior prophet” account the Evangelist did it; on the testimonia account his source did; on both, it was done deliberately.

4. And it does not, by itself, explain the wrong prophet. This is the sharpest limit. The one testimonia sheet we can actually inspect — 4Q175 — carries no attributions at all: it strings texts anonymously. A source that hands on unlabelled clusters explains the conflation handsomely but says nothing about how a single, wrong prophet’s name got attached. To reach the misattribution the hypothesis must borrow the very thing the mainstream reading already supplies — the convention of naming the senior or governing prophet for a blend — or fall back on a copyist’s slip. So even granting a testimonia source, the “wrong prophet” is a second phenomenon needing a second explanation.

Assessment

Plausible. That the conflated citations were drawn from a pre-formed collection of Christian proof-texts — so that the conflation, and sometimes the single wrong attribution, was already in the source — is a coherent, genuinely-advocated mechanism whose central presupposition is now demonstrated rather than merely imagined: Qumran’s 4QTestimonia proves that pre-Christian sheets of grouped, excerpted proof-texts existed and could carry a pre-formed composite, and Skarsaune shows written testimony-sources in use close to the New Testament. What holds it short of probable for any given crux is that no surviving collection contains Mark’s or Matthew’s composite, so the mechanism is inferred, not read off a document; that Harris’s maximal “one fixed Testimony Book” form is the part scholarship has rightly set aside, on Dodd’s showing that such books were the result and not the presupposition of early Christian exegesis; and that a testimonia source, since the Qumran exemplar labels nothing, explains the conflation far better than the wrong prophet — which still needs the “name the senior prophet” convention or a slip. Rightly, then, the testimonia mechanism does not compete with the doctrine’s deliberate-composite verdict but stands upstream of it: it can push the deliberate fusion back to a compiler, but the fusion stays deliberate, and the sound readings of Mark 1:2 and Matthew 27:9 are untouched by whether the Evangelist or his source did the fusing.

Sources