Abstract
This paper argues that the passage in To Theodore where Clement specifies the placement of the “mystic Mark” material operates in a recognizably “apparatus-like” register: it identifies Gospel units by incipit, bounds them by a terminus formula (“and the following until…”), marks verbatim quotation, links adjacent units by boundary language (“after these comes…”), and adjudicates exclusion (“is not found”). These are not merely rhetorical flourishes but features of a technical discourse of textual navigation and control, consonant with Alexandrian scholarly habits and closely paralleled in Origen’s synoptic handling of Gospel lemmata. When set beside the Ammonian/Eusebian tradition—known chiefly through Eusebius of Caesarea’s “Letter to Carpianus”—Clement’s language looks less like idiosyncratic private referencing and more like participation in a broader culture of Gospel segmentation and lookup. On this basis, the paper contends that Morton Smith’s publications, while accurately translating the relevant cues, did not adequately thematize their “apparatus” character; the result is an interpretive overemphasis on a discrete “secret Gospel” as a narrative artifact and an underemphasis on the letter’s procedural aim: refuting falsification by controlled citation and unit-checking.
Problem and historiography
Morton Smith’s 1958 discovery at Mar Saba Monastery and his 1973 publications—especially Clement of Alexandria and a Secret Gospel of Mark and The Secret Gospel: The Discovery and Interpretation of the Secret Gospel According to Mark—made To Theodore a central (and polarizing) artifact in debates about early Gospel traditions, Alexandrian Christianity, and modern forgery.
Within those debates, the methodological texture of Clement’s “placement instructions” has often been treated as transparent: a straightforward report that certain material occurs after Mark 10:32–34 and that another brief addition occurs after Mark 10:46, plus a denial that other allegedly libertine phrases occur. In Smith’s own translation, the passage is framed as an exemplar (“For example…”) in which Clement “refut[es] the falsifications by the very words of the Gospel,” then provides the placement brackets and verbatim excerpts.
This paper reframes the same lines as evidence for a genre of reference: pericope navigation that treats Gospel text as segmented into stable, retrievable units. The main historiographical claim is not that Smith mistranslated Clement but that he did not sufficiently recognize the scholarly technology implied by Clement’s phrasing, and so did not explore how an “apparatus” reading alters what the letter most directly evidences: not only additional narrative content, but also procedures for locating, bounding, quoting, sequencing, and excluding textual units.
Two broader scholarly contexts make such a reframing historically plausible. First is the Alexandrian environment in which “grammar” and philological technique are not peripheral but structurally integrated into Christian exegesis and polemic; recent work on Clement’s relationship to Greek grammatical tradition emphasizes precisely the role of grammatical operations and disciplines of reading in Alexandrian Christian interpretation. Second is the late-antique emergence of explicit Gospel-navigation systems, most famously those explained in Eusebius’ “Letter to Carpianus,” which itself credits Ammonius with an earlier synopsis-like arrangement of Gospel “sections.”
Clement’s apparatus-like procedure in To Theodore
The critical passage appears where Clement announces that he will answer Theodore’s questions “disproving the slanders through the same discourses of the gospel” and then gives a set of placement and control instructions. In the Greek text, five tightly coordinated moves appear.
First, Clement anchors a unit by incipit—not by numeric reference (which did not exist in antiquity as chapter-and-verse) but by the opening words of a Markan segment: “Ἀμέλει μετὰ τὸ ἦσαν δὲ ἐν τῇ ὁδῷ ἀναβαίνοντες εἰς Ἱεροσόλυμα…”. This makes the unit’s identity dependent on recognizability of opening words: the reader is expected to be able to find the place by incipit recognition. In Smith’s translation this appears as “after ‘And they were in the road going up to Jerusalem,’” and he supplies the modern bracket [Mark 10:32–34], which is helpful for contemporary readers but also illustrates a key point: the ancient coordinate is the incipit label, not a numbered verse.
Second, Clement bounds the referenced stretch by a terminus formula—“καὶ τὰ ἑξῆς ἕως, μετὰ τρεῖς ἡμέρας ἀναστήσεται”—constructing a compressed “range” bracket (incipit → and the following until → endpoint clause). The formula is not a vague “etc.” but a boundary operator: it treats the segment as a delimited object whose end can be indicated by citing an internal clause at the terminus. Smith captures the same bounding logic (“and what follows, until ‘After three days he shall arise’”).
Third, Clement explicitly signals verbatim transfer: “Ὧδε ἐπιφέρει κατὰ λέξιν” (“here it brings in word for word”). This is a meta-textual instruction about the status of what follows: not paraphrase, not summary, but controlled citation of the lettered text. Importantly, the idiom “κατὰ λέξιν … ἐπιφέρει” is not alien to Clement’s undisputed practice: in the Stromata Clement introduces quotation of an opponent’s words with comparable language (“κατὰ λέξιν … ἐπιφέρει”), showing that “word-for-word” and “he brings in” belong to an established Clementine citation repertoire rather than being unique to To Theodore.
Fourth, Clement links units into a sequence by boundary formulas. After the quoted “mystic Mark” block, he states: “Ἐπὶ μὲν τούτοις ἕπεται τὸ καὶ προσεπορεύοντο αὐτῷ Ἰάκωβος καὶ Ἰωάννης καὶ πᾶσα ἡ περικοπή.” The syntax treats the next Markan incipit (“and James and John came to him”) as a boundary marker for the adjacent unit, with “the whole pericope/section” following. In other words, Clement is not merely narrating that “then James and John came”; he is naming the unit that can be located by its opening words and then gesturing to its full extent (“and the whole pericope”).
Fifth, Clement polices inclusion/exclusion with an explicit “is not found” judgment: “Τὸ δὲ γυμνὸς γυμνῷ … οὐκ εὑρίσκεται.” By narrating what is (and is not) “found,” Clement treats the Gospel text as a checkable repository. This “repository posture” is reinforced by a second placement instruction that repeats the incipit-and-addition pattern: “Μετὰ δὲ τὸ καὶ ἔρχεται εἰς Ἰεριχὼ ἐπάγει μόνον…”—again, an incipit anchor (“and he comes into Jericho”) plus a constraint (“adds only…”) that implies controlled textual verification rather than free paraphrase.
Taken together, these moves instantiate a workflow: identify a Gospel unit by incipit, delimit its span, mark verbatim quotation, position an added unit between two adjacent canonical units by naming the boundary incipits, and deny spurious phrases by explicit non-occurrence. This is precisely the kind of prose one writes when treating Gospel narrative as a set of indexed segments rather than as an undifferentiated flow.
The Ammonian and Alexandrian apparatus horizon
The notion of a “Gospel apparatus” in late antiquity is most concretely documented by Eusebius’ epistle “To Carpianus,” which describes both an earlier work by Ammonius and Eusebius’ own cross-referencing system. Eusebius credits Ammonius with producing a “harmonized account of the four gospels,” constructed by placing the corresponding sections of Mark, Luke, and John alongside Matthew—at the cost of disrupting their sequential order. Eusebius presents his own solution: preserve the order of each Gospel while providing numbered “sections” and ten “tables/canons” that allow the reader to look up parallels by reference-number rows.
Three features of this Eusebian description are especially relevant for reading Clement’s placement language.
The first is vocabulary and conceptuality. Eusebius speaks of Gospel “sections” (περικοπαί) and “passages/places” (τόποι), and he explains the practical use-case: a reader “open[s] one of the four gospels at some point,” wishes to know what other gospels contain “similar things,” and uses the apparatus to “immediately discover” and “find” the related passages. Clement’s “οὐκ εὑρίσκεται” (“is not found”) and his repeated incipit anchoring, in turn, make best sense in a world where “finding” a passage is an explicit reading operation supported by stable segmentation.
The second is the principle of preserving narrative integrity while enabling comparison. Eusebius describes Ammonius’ method as one that damaged the sequential order of the three non-Matthean Gospels, whereas his own method keeps each Gospel’s “body and sequence” intact while still enabling comparison. Clement’s method in To Theodore is strikingly aligned with the “preserve sequence + insert controlled references” ethos: he does not retell Mark; he names the relevant Markan boundary points (incipits) and inserts (or denies) small, controlled textual blocks at those seams.
The third feature is that apparatus-use is not merely theoretical; it expects a reader capable of operating by “place.” In Origen’s Commentary on Matthew, the procedural presence of such “place-based comparison” becomes explicit. In Book 16, Origen frames pericopes by incipit and terminus (“καὶ τὰ ἑξῆς ἕως τοῦ…”) and aligns them across Gospels using equivalence language (“τὰ δὲ ἰσοδυναμοῦντα…”) and “similar” markers (“τὸ δ’ ὅμοιον…”). He then issues an almost programmatic instruction: “σὺ δὲ παραθεὶς τὰ εὐαγγέλια ἀλλήλοις κατὰ τοὺς τόπους τούτους καὶ συγκρίνων…”—“set the Gospels beside one another at these places and compare them.”
This matters for Clement because it shows that the apparently “technical” grammar of incipit labels and bounded spans is not isolated. In Origen it is part of an Alexandrian exegetical culture that understands Gospel text as divisible into comparable units and expects readers to navigate by “places.”
Whether we should call Clement’s practice “Ammonian” in a strict historical sense depends on chronology and on the debated identity/dating of Ammonius’ work. Matthew R. Crawford argues that Ammonius likely composed a four-column synopsis (a “Diatessaron-Gospel”) and that Eusebius later reworked the parallels as “raw data” into his numbered canon tables; Crawford also notes scholarly uncertainty about priority and the identity of Ammonius, while leaning toward Eusebius’ report that Ammonius was Origen’s teacher. This uncertainty counsels caution: one need not posit Clement’s direct use of Ammonius to see that Clement’s phrasing coheres with the same Alexandrian “segmented text + place-based lookup” mentality visible in Origen and later institutionalized by Eusebius.
Morton Smith’s translation and the apparatus blind spot
Smith’s translation of the key passage is, in many respects, an excellent witness to the apparatus-like character of Clement’s prose. He renders the incipit anchor and range bracket (“after ‘And they were in the road going up to Jerusalem,’ and what follows, until ‘After three days he shall arise’”), includes the explicit “word for word” marker, and then preserves the boundary-link language (“After these words follows…,” “And after the words… adds only…,” “are not found”). In other words: Smith did not miss the signals at the level of translation.
The issue is interpretive emphasis. As summarized by Borger A. Pearson, Smith’s larger thesis treated the “secret Gospel” as a key witness to an early stage in Mark’s development and posited that canonical Mark had “expunged” material from a longer form; Pearson further notes that while many scholars accepted the letter’s authenticity, Smith’s broader reconstruction of earliest Christianity and his interpretation were not accepted by “a single reputable scholar” (in Pearson’s judgment). The apparatus-like placement language thus became, in Smith’s hands and in subsequent debate, a principal mechanism for reconstructing a longer narrative Mark with specific insertions, rather than being foregrounded as evidence for a text-critical and navigational procedure.
Two interpretive consequences follow.
First, the addition of modern chapter/verse labels in Smith’s translation—though pedagogically useful—encourages modern readers to think the coordinate system is essentially ours, i.e., that Clement is “really” pointing to Mark 10:32–34 as numbered verses, rather than to a unit defined by opening words and bounded span. Yet Clement’s own coordinate method is precisely the type of method Eusebius describes as necessary before chapters/verses: identify a delimited “place” by intrinsic textual cues so that it can be found and compared.
Second, Smith’s programmatic exploitation of the letter for a history-of-tradition thesis tends to obscure the letter’s declared immediate aim. Clement introduces the excerpt not as “here is a secret rite text,” but as evidence for refutation: he will answer Theodore’s questions by “refuting the falsifications by the very words of the Gospel,” and he denies that certain phrases “are not found.” That is, Clement frames his procedure as textual policing—what is genuinely present where, what is spurious, and what is absent. The apparatus reading simply takes Clement at his word about what he is doing: not merely narrating, but controlling citation boundaries and verifiability.
This is where the “Ammonius apparatus” lens produces a constructive reinterpretation. If Clement is writing in a milieu where Gospel texts are commonly navigated through segmented units (by incipit and span) and where “place-based collation” is a recognizable scholarly operation (as Origen explicitly says), then Clement’s otherwise highly compressed, quasi-hyperlink prose is not anomalous. The letter reads less like an improvisation to help Theodore find an obscure text and more like exploitation of a technical convention: “here is the unit; here are its boundaries; here is what the contested text adds; here is what is not present.”
Implications and conclusion
Reading Clement’s placement instructions as “apparatus discourse” has three implications.
It clarifies why Clement’s signal cluster is so dense. In a pericope-apparatus environment, one expects compressed unit labels, bounded spans, and adjacency formulas because these are precisely what enable fast lookup and controlled comparison—functions Eusebius makes explicit and Origen presupposes in practice.
It shifts the evidential center of gravity. The letter certainly provides (if authentic) remarkable testimony to additional Markan material; but it also testifies, more securely and directly, to a method: Mark is treated as segmented, checkable, and policed by what can and cannot be “found.” That methodological testimony does not require us to decide immediately between competing macro-theories (early Mark vs later cento, genuine Clement vs pseudepigraphon); it is a feature visible at the level of syntax and discourse function.
Finally, it reframes the evaluation of Smith’s contribution. Smith’s translation faithfully conveys the apparatus-like signals (incipit anchoring, span bounding, verbatim flagging, adjacency linking, and exclusion policing). The “misunderstanding,” insofar as the term applies, is best located not in philological incompetence but in interpretive weighting: Smith elevated the placement cues primarily as coordinates for reconstructing a longer Mark and for supporting a developmental hypothesis about the canonical Gospel, while leaving comparatively underdeveloped the possibility that these cues belong to a conventional Alexandrian technology of textual navigation—one that sits plausibly in the same broad trajectory that includes Ammonius’ synopsis-like work and culminates in Eusebius’ canon tables.
