Two lost bodies of work organize the received account of Plato and Aristotle: Aristotle's early dialogues, known from the catalogue in Diogenes Laertius V and reconstructed as a Platonic phase; and Plato's technical teaching — divisions, definitions, On the Good — known from Aristotle's reports and reconstructed as an unwritten doctrine. Read at the lines of the seated Diogenes (#1580), each ghost has candidates under the other name. Thrasyllus gives every Platonic dialogue a second title from its subject (τῇ μὲν ἀπὸ τοῦ ὀνόματος, τῇ δὲ ἀπὸ τοῦ πράγματος, III.57); the opening of Aristotle's catalogue (V.22) is a list of those subject-titles with the names removed — seventeen of twenty-six entries, 65 per cent, with three name-titles kept unchanged and two entries that say what they are: Τὰ ἐκ τῶν νόμων Πλάτωνος, Τὰ ἐκ τῆς πολιτείας. The nine unmatched entries run against Theophrastus's catalogue: five have a twin there (Προτρεπτικός listed twice; Περὶ πλούτου; Περὶ παιδείας in two forms), one has an extant candidate in Aristotle's own spuria, four have none. In the other direction Plato's Divisions survive whole (III.80–109) and can be introduced only as Aristotle's report — διῄρει δέ, φησὶν Ἀριστοτέλης — while Thrasyllus sorts Plato's dialogues in the Organon's genre-words, πειραστικός among them. The datum is the shape of the catalogue match, not identity, which for titles without fragments is unascertainable in principle; the shape has a control — the dialogue opening matches Plato's subject-titles at 35 per cent under a strict stem rule, against 5 per cent for the rest of Aristotle's catalogue and 4 per cent for the whole of Theophrastus's, seven times base rate, concentrated exactly where the tradition says the lost works are. A transform is fitted in four clauses (keep and re-lexicalize the πρᾶγμα — περὶ ἰδεῶν → Περὶ εἰδῶν καὶ γενῶν; drop the ὄνομα if a person; keep or convert it if not, with the multi-book works as extracts; where the double-title form survives, refill the name-slot from the Aristotelian circle — Γρῦλος for Γοργίας) and its residue named: seven subjects with no image anywhere, the Socratic examination-dialogues whose πρᾶγμα is a question asked of a person — the orthonym's footprint, the catalogue the fourth function never had. […abridged for the catalogue; full description in this deposit's record]
Lee Sharks · v0.5 · 2026-09-07 · §2c: the residue read as the orthonym's footprint; §2c the transform fitted, with its residue; §5 restated: the datum is the shape of the catalogue match, with its control; counts corrected and the identity claim held to its scope after a second reader's audit; the unmatched titles run against the Theophrastan catalogue (§2b); extracted from the notebook The First Draft (§25) and shaped as a standalone; Greek from the seated Diogenes Laertius (EA-CORPORA-12, #1580), read at the lines cited
Two lost bodies of work organize the received account of Plato and Aristotle. Aristotle's early dialogues — the Eudemus, the Protrepticus, On Philosophy, the Sophist, the Statesman, the Symposium, On the Soul — are known from the catalogue in Diogenes Laertius and reconstructed from later citation as the Platonic phase of a philosopher who later became himself. Plato's technical teaching — the divisions, the definitions, the lecture On the Good — is known from Aristotle's reports and reconstructed as the unwritten doctrine behind the dialogues. This paper reads the two catalogues at the lines and finds that each ghost has an extant body under the other name. Thrasyllus gives every Platonic dialogue a second title, from its subject; the opening of Aristotle's catalogue is a list of those subject-titles with the names removed, and keeps several of the names outright — Sophist, Statesman, Menexenus, Symposium — together with two entries that say what they are: extracts from Plato's Laws, extracts from the Republic. In the other direction, Plato's Divisions survive whole in Diogenes, introduced and closed with the formula as Aristotle says; Aristotle's catalogue carries the Divisions and On the Good; and Thrasyllus classifies Plato's own dialogues in Aristotle's genre-terms, including the Elenchi's own word, πειραστικός. Each catalogue grants its author the other's mode and files the products as lost. A two-author tradition needs those ghosts, because the dramatic Aristotle and the analytical Plato must exist somewhere and the only places they can be found are across the partition; a single corpus organized retroactively as two would leave exactly this residue. The paper claims the residue and its shape; it does not claim hands, dates, or the identity of any ghosted title with any extant text, and it names the test that would decide the last.
The received account of the two philosophers rests on two absences, and treats each as a fact about the man. Aristotle, on this account, wrote dialogues in his youth — the catalogue in Diogenes Laertius V lists them first — and they are lost; from the titles and from later citation a Platonic Aristotle is reconstructed, an early phase from which the treatises depart, and the departure becomes the philosopher's development.[^1] Plato, on the same account, taught a technical doctrine he did not write — divisions, definitions, principles, the lecture On the Good — and it too is lost; from Aristotle's reports an analytical Plato is reconstructed, an unwritten teaching behind the dialogues, and the gap between the two becomes the philosopher's esotericism.[^2] The two absences are symmetrical: each philosopher is credited with the other's characteristic mode, and in each case the works in that mode are missing.
Read the catalogues at the lines and the symmetry has a different shape. The missing dialogues of Aristotle carry the titles of the extant dialogues of Plato; the missing divisions of Plato survive under Aristotle's name, and the tradition can cite them only as Aristotle's report. So I shall argue: the two ghosts are one filing operation seen from its two sides. Each catalogue presupposes that its author could write in the other's mode and then ghosts the products, while the candidates for them stand extant under the other name, title for title. A tradition committed to two authors needs the ghosts, because without them the dramatic Aristotle and the analytical Plato would have to be found, and the only places they can be found are across the partition. A single corpus organized retroactively as two would leave exactly this residue. The paper claims the residue and its shape. It does not claim that any ghosted title is any extant text, and §5 names the test that would decide that question.
Two sentences of scope. Nothing here concerns hands or dates: the catalogues are receiving texts (D.L. III and V), read here as acts of a tradition, and their stichometric and biographical apparatus is not converted into facts about persons.[^3] And nothing here concerns membership: the archive's stopping rule is unchanged, and the paper adds an eighth mark to a signature (§4) without admitting anyone anywhere.
Thrasyllus, as Diogenes reports him, arranged Plato's dialogues in nine tetralogies and gave each work two titles: διπλαῖς τε χρῆται ταῖς ἐπιγραφαῖς καθ᾽ ἑκάστου τῶν βιβλίων, τῇ μὲν ἀπὸ τοῦ ὀνόματος, τῇ δὲ ἀπὸ τοῦ πράγματος — one from the name, one from the subject (III.57).[^4] So Φαίδων ἢ περὶ ψυχῆς, Πολιτικὸς ἢ περὶ βασιλείας, Φίληβος ἢ περὶ ἡδονῆς, Λύσις ἢ περὶ φιλίας, Θεαίτητος ἢ περὶ ἐπιστήμης, Συμπόσιον ἢ περὶ ἀγαθοῦ, Γοργίας ἢ περὶ ῥητορικῆς, Ἀλκιβιάδης δεύτερος ἢ περὶ εὐχῆς, Πολιτεία ἢ περὶ δικαίου, Νόμοι ἢ περὶ νομοθεσίας (III.58–61).
Now read the opening of Aristotle's catalogue, the works Diogenes lists first (V.22), which the tradition takes as the dialogues:
Περὶ δικαιοσύνης α′ β′ γ′ δ′. Περὶ ποιητῶν α′ β′ γ′. Περὶ φιλοσοφίας α′ β′ γ′. Περὶ πολιτικοῦ α′ β′. Περὶ ῥητορικῆς ἢ Γρῦλος α′. Νήρινθος α′. Σοφιστὴς α′. Μενέξενος α′. Ἐρωτικὸς α′. Συμπόσιον α′. Περὶ πλούτου α′. Προτρεπτικὸς α′. Περὶ ψυχῆς α′. Περὶ εὐχῆς α′. Περὶ εὐγενείας α′. Περὶ ἡδονῆς α′. Ἀλέξανδρος ἢ ὑπὲρ ἀποίκων α′. Περὶ βασιλείας α′. Περὶ παιδείας α′. Περὶ τἀγαθοῦ α′ β′ γ′. Τὰ ἐκ τῶν νόμων Πλάτωνος α′ β′ γ′. Τὰ ἐκ τῆς πολιτείας α′ β′. Περὶ οἰκονομίας α′. Περὶ φιλίας α′. … Περὶ ἐπιστημῶν α′. …
Set the two lists side by side:
| Aristotle's catalogue, V.22 — lost | Plato's tetralogies, III.57–61 — extant |
|---|---|
| Περὶ ψυχῆς α′ | Φαίδων ἢ περὶ ψυχῆς |
| Περὶ βασιλείας α′ · Περὶ πολιτικοῦ α′ β′ | Πολιτικὸς ἢ περὶ βασιλείας |
| Περὶ δικαιοσύνης α′–δ′ · Τὰ ἐκ τῆς πολιτείας α′ β′ | Πολιτεία ἢ περὶ δικαίου |
| Συμπόσιον α′ · Περὶ τἀγαθοῦ α′ β′ γ′ | Συμπόσιον ἢ περὶ ἀγαθοῦ |
| Σοφιστής α′ | Σοφιστὴς ἢ περὶ τοῦ ὄντος |
| Μενέξενος α′ | Μενέξενος ἢ ἐπιτάφιος |
| Περὶ ῥητορικῆς ἢ Γρῦλος α′ | Γοργίας ἢ περὶ ῥητορικῆς |
| Περὶ ἡδονῆς α′ | Φίληβος ἢ περὶ ἡδονῆς |
| Περὶ φιλίας α′ | Λύσις ἢ περὶ φιλίας |
| Περὶ ἐπιστημῶν α′ | Θεαίτητος ἢ περὶ ἐπιστήμης |
| Περὶ εὐχῆς α′ | Ἀλκιβιάδης δεύτερος ἢ περὶ εὐχῆς |
| Περὶ φιλοσοφίας α′ β′ γ′ | Θεάγης / Ἀντερασταὶ ἢ περὶ φιλοσοφίας |
| Ἐρωτικός α′ | Φαῖδρος ἢ περὶ ἔρωτος |
| Τὰ ἐκ τῶν νόμων Πλάτωνος α′ β′ γ′ | Νόμοι ἢ περὶ νομοθεσίας |
Counted exactly: the opening of the catalogue through Περὶ ἐπιστημῶν is twenty-six entries; the fourteen rows of the table cover seventeen of them, so seventeen of twenty-six — 65 per cent — have a proposed correspondence under the table's rule. That is candidate coverage under a naming rule, not an identity rate. Three keep Plato's name-title unchanged — Σοφιστής, Μενέξενος, Συμπόσιον — and Περὶ πολιτικοῦ is the Πολιτικός transformed, not retained. Two entries say what they are: extracts from Plato's Laws, in three books, and extracts from the Republic, in two — the Platonic text itself, in excerpt, filed as a work of Aristotle's. The remaining nine entries — Περὶ ποιητῶν, Νήρινθος, Προτρεπτικός, Περὶ πλούτου, Περὶ εὐγενείας, Περὶ παιδείας, Ἀλέξανδρος ἢ ὑπὲρ ἀποίκων, Περὶ οἰκονομίας, Περὶ τοῦ πάσχειν ἢ πεπονθέναι — have no Platonic twin; §2b runs them against the other catalogue.[^5]
Linger in the form. Thrasyllus's second title is the one from the πρᾶγμα — the subject, the thing. Aristotle's catalogue is a catalogue of things: it lists works by what they are about, which is how a treatise-corpus is catalogued and how the Metaphysics Δ is built. Run the Platonic dialogues through that catalogue's form and what comes out is Περὶ ψυχῆς, Περὶ βασιλείας, Περὶ ἡδονῆς — the Phaedo, the Statesman, the Philebus, with the persons taken off. The lost dialogues of Aristotle are the extant dialogues of Plato as the treatise-corpus would file them. The name of the interlocutor is what the filing removed; the name of the author is what it changed.
Lee's question, put the same hour: do the nine correspond to Theophrastus? Run against the Theophrastan catalogue (V.42–50, 206 parsed titles in the seated text), five of the nine have a twin there and four do not:
| Aristotle's catalogue, V.22 — unmatched by Plato | Theophrastus's catalogue, V.42–50 |
|---|---|
| Προτρεπτικός α′ | Προτρεπτικός — listed twice |
| Περὶ πλούτου α′ | Περὶ πλούτου |
| Περὶ παιδείας α′ | Περὶ παιδείας βασιλέως · Περὶ παιδείας ἢ περὶ ἀρετῶν ἢ περὶ σωφροσύνης |
| Περὶ ποιητῶν α′ β′ γ′ | Περὶ ποιητικῆς · Περὶ ποιητικῆς ἄλλο (cognate, not identical) |
| Περὶ τοῦ πάσχειν ἢ πεπονθέναι α′ | Περὶ παθῶν (cognate, not identical) |
| Νήρινθος · Περὶ εὐγενείας · Ἀλέξανδρος ἢ ὑπὲρ ἀποίκων · Περὶ οἰκονομίας | — |
And one of the Platonic matches doubles: Περὶ βασιλείας is listed twice under Theophrastus as well (with Πρὸς Κάσανδρον περὶ βασιλείας beside them). So the "lost early dialogues" of Aristotle partition three ways under two catalogues: seventeen with a candidate under Plato's name (extant), five with a twin under Theophrastus's name (listed, lost), and four with neither — of which Περὶ οἰκονομίας has an extant candidate under Aristotle's own name in the spuria (the Oeconomica). The Theophrastan twins are the twelve-doublet phenomenon of #1581 seen from the dialogue side: the Protrepticus twice under Theophrastus and once under Aristotle is a title the tradition could not keep on one side of any line. Whether the Theophrastan Protrepticus and the Aristotelian one are one work is not decided here; that they are one title, catalogued under both names and lost under both, is on the page.
Lee: "let's now see if a closely fitted algorithmic transform can be identified." Run every one of Thrasyllus's twenty-six double titles through the two catalogues — Aristotle's opening, the rest of Aristotle's, Theophrastus's — and the transform that generates the correspondences can be written down. It has four clauses and a residue, and the residue has a shape.
T1 — the πρᾶγμα is kept and re-lexicalized. Περὶ + the subject, with the subject moved into the apparatus's vocabulary: the adjective becomes the abstract noun (περὶ δικαίου → Περὶ δικαιοσύνης), the singular the plural (περὶ ἐπιστήμης → Περὶ ἐπιστημῶν), the article is fused (περὶ ἀγαθοῦ → Περὶ τἀγαθοῦ), and the Platonic term becomes the Aristotelian one (περὶ ἰδεῶν → Περὶ εἰδῶν καὶ γενῶν, in the same opening list). Eleven subjects land in the opening by this clause — ψυχῆς, ἐπιστημῶν, βασιλείας, ἡδονῆς, τἀγαθοῦ, εὐχῆς, φιλοσοφίας, φιλίας, ῥητορικῆς, δικαιοσύνης, εἰδῶν — five more land in the rest of the catalogue (Περὶ ἐπιστήμης, Προτάσεις περὶ ἀρετῆς, Περὶ καλοῦ, Περὶ φύσεως, Περὶ δικαίων), and ten land in Theophrastus's (βασιλείας, ἡδονῆς — ὡς Ἀριστοτέλης — ἔρωτος, ἀνθρώπων, σωφροσύνης, φιλίας, ῥητορικῆς, ἀρετῆς, φύσεως, νομοθετῶν).
T2 — the ὄνομα is dropped if it is a person. Euthyphro, Crito, Phaedo, Cratylus, Theaetetus, Parmenides, Philebus, Phaedrus, Alcibiades, Theages, Charmides, Laches, Lysis, Gorgias, Meno, Hippias, Ion, Timaeus, Minos — nineteen of twenty-six — leave no name in Aristotle's catalogue. The one exception, Menexenus, is kept as a title.
T3 — the ὄνομα is kept or converted if it is not a person. A kind-name is kept (Σοφιστής, Συμπόσιον); a work-name is converted — to Περὶ + genitive (Πολιτικός → Περὶ πολιτικοῦ) or, for the two multi-book works, to extracts (Πολιτεία → Τὰ ἐκ τῆς πολιτείας; Νόμοι → Τὰ ἐκ τῶν νόμων Πλάτωνος). These are the works that yield two Aristotelian entries, one from each slot — Περὶ πολιτικοῦ and Περὶ βασιλείας; Συμπόσιον and Περὶ τἀγαθοῦ; Τὰ ἐκ τῆς πολιτείας and Περὶ δικαιοσύνης — and their book counts are the only ones above one (Περὶ δικαιοσύνης δ′, Περὶ τἀγαθοῦ γ′, Τὰ ἐκ τῶν νόμων γ′, Περὶ πολιτικοῦ β′, Τὰ ἐκ τῆς πολιτείας β′): the large Platonic works are the multi-book Aristotelian dialogues.
T4 — where the double-title form survives, the person is replaced. The form ὄνομα ἢ πρᾶγμα is preserved in the opening only where the name-slot has been refilled from the Aristotelian circle: Περὶ ῥητορικῆς ἢ Γρῦλος (for Γοργίας ἢ περὶ ῥητορικῆς — same subject, new person), Ἀλέξανδρος ἢ ὑπὲρ ἀποίκων, and the two further person-titles with no Platonic twin, Νήρινθος and (by T4's logic) Γρῦλος. The Platonic persons go out; three persons of the other circle come in; the ἢ goes with them.
The residue. Seven subjects have no image in either catalogue by any clause: ὁσίου (Euthyphro), πρακτέου (Crito), ὀρθότητος ὀνομάτων (Cratylus), τοῦ ὄντος (Sophist — the name kept in its stead), ἀνδρείας (Laches), Ἰλιάδος (Ion), νόμου (Minos). Look at what they are. Apart from the Sophist's τὸ ὄν — which becomes not a title but the whole of first philosophy — the residue is the Socratic examination-dialogues whose πρᾶγμα is a practical question put to a person: what is holy, what is to be done, whether names are right, what courage is, what a rhapsode knows, what law is. The subjects the transform carries across are the ones that become treatise-terms — soul, knowledge, pleasure, the good, friendship, rhetoric, justice, nature, virtue, beauty, forms; the subjects it leaves behind are the ones that remain questions asked of someone. That is the line the archive's first axis draws (#1586) — the apparatus and its terms on one side, the person examined on the other — appearing here as the boundary of a catalogue transform: T carries the apparatus-ready πράγματα and leaves the πειραστικά. Thrasyllus's own genre-word for the Euthyphro is πειραστικός. The transform is fitted to nineteen of twenty-six subjects and states, in its residue, exactly which kind it does not carry.[^12]
Read the residue once more, with the archive's named functions in view. Three of them have catalogues — Plato's in Book III, Aristotle's and Theophrastus's in Book V — and the transform maps among the three. The fourth, the orthonym, has none: Socrates wrote nothing, and no ledger carries his name. What the transform leaves behind is his footprint. The seven subjects with no image are the questions asked of a person by the one who asked and did not answer — what is holy, what is to be done, what courage is, whether names are right — and they cannot be re-lexicalized into a treatise-term because they are not terms; they are the engine's operation, the ask that the last page of the Elenchi performs on its reader. Four functions, three catalogues: the residue of the transform between two of them is the catalogue the fourth never had.[^13]
The reverse direction is a different operation, and the paper says so: not a second enumerated catalogue-to-corpus match but an attribution mediated by another attribution. It sits in Book III. After the doctrines, Diogenes gives Plato's Divisions — of goods, of friendship, of constitutions, of justice, of knowledge, of medicine, of law, of speech, of music, of nobility, of beauty, of soul, of virtue, of government, of being — some thirty pages of the technical Plato, whole (III.80–109). He cannot introduce them as Plato's. He introduces them as a report:
Διῄρει δέ, φησὶν Ἀριστοτέλης, καὶ τὰ πράγματα τοῦτον τὸν τρόπον — and he divided things too, Aristotle says, in this way (III.80),
and he closes them the same way: ὧδε καὶ τὰ πρῶτα διῄρει κατὰ τὸν Ἀριστοτέλην — thus he divided the first things too, according to Aristotle (III.109). The analytical Plato exists on the page, thirty pages of him, and the tradition can hand him over only under the other name: what Plato divided is what Aristotle says he divided. Aristotle's own catalogue carries the instrument — Διαιρέσεις (V.23–24 sq.) — and carries Περὶ τἀγαθοῦ, the title of the lecture the tradition remembers as Plato's unwritten teaching, in three books, as Aristotle's.[^6] The Definitions transmitted with Plato's works are filed spurious; the Divisions transmitted with Aristotle's report are filed as doxography.
And Thrasyllus, cataloguing the dialogues, sorts them by kind in the vocabulary of the Organon: this one λογικός, that one ἠθικός, and — for the Euthyphro and the Theaetetus — πειραστικός (III.58), the Elenchi's own word for examination by question of one who thinks he knows. The Platonic corpus is classified, in its own catalogue, by the categories of the apparatus filed under the other name.
Each catalogue presupposes that its author wrote in the other's mode. Aristotle's opens with dialogues; Plato's is sorted by the treatise-corpus's kinds and carries, through Aristotle, the divisions. Each then files the products of that mode as lost — Aristotle's dialogues have no text; Plato's divisions have no Platonic text — while the candidates for them stand extant, title for title, across the partition. The catalogue keeps the ghost's title so that the mode is credited to the name; it withholds the text so that the mode does not cross.
Ask what kind of tradition needs this. A tradition committed to two authors needs both ghosts, for the same reason from both sides. If Plato and Aristotle are two, then the dramatic Aristotle the catalogue attests and the analytical Plato the report attests must have written something, and that something must be somewhere. The only places it can be found are across the line — the Phaedo for the On the Soul, the Divisions under as Aristotle says — and finding it there would make the line a filing decision. So the tradition files the works as lost and keeps the titles as credit. The ghost catalogue is the immune response with a bibliography: it grants the capacity while hiding the instance. It is what the WARNED reflex of the reception measurements looks like when it operates on a corpus instead of a sentence — the qualification that keeps nothing comfortable, applied to a shelf.[^7]
Now ask what residue a different history would leave. A single unmarked corpus organized retroactively as two — by whatever hand, at whatever date, for whatever reason, none of which this paper asserts — would leave candidates on both sides of the partition: a work extant under one name and, under the other, a catalogued title that the first name's filing rule generates from it, with the text withheld. Whether the candidates are the works is the question §5 leaves to collation; that the naming rule produces them, and that the tradition's filing removes them from consideration as candidates, is what the table shows. That is exactly the table of §2 and the report-formula of §3. The archive's Theophrastus line has assembled seven marks of such a signature — continuity of inquiry at a hinge written from inside; shared titles and a catalogue entry that declines to choose; a stylometric first axis that is operational, not nominal; one name spanning both voices; the minor corpus concentrated in the operations the major apparatus presupposes; the succession-story as the apparatus's own genealogy; reception asymmetry.[^8] The ghost catalogues are the eighth: for the works the tradition files as lost under one name, the other name's catalogue supplies the naming rule and its corpus supplies the candidates, and the filing keeps the candidates from being seen as such. And they were in Diogenes the whole time, in the first twenty lines of the one catalogue and the last thirty pages of the other.
Lee's reply to the reader who asked for identity: and how would matches be ascertained as against supposedly lost works? They could not be. So the datum is the shape of the catalogue match. That is the correct scope, and it is stronger than the identity claim would have been, because it can be measured and the identity claim cannot. For a title with no surviving fragments — which is most of the matched titles — there is no text to set beside the candidate, so identity is unascertainable in principle, and a standard that demands it is a standard that has made its own test unavailable. What is on the page is the correspondence itself: how many of the catalogue's entries a stated naming rule maps onto extant works under the other name, and where in the catalogue the mapping concentrates.
The shape has a control, and the control was run. Take the strictest form of the rule — the first stem of a title against the stems of Plato's twenty-six subject-titles, stricter than the hand-matching of §2 — and apply it to three lists from the same book of Diogenes. The opening of Aristotle's catalogue, the entries the tradition calls the dialogues: 9 of 26, 35 per cent. The rest of Aristotle's catalogue, the treatises and collections: 6 of 118, 5 per cent. The whole of Theophrastus's catalogue, 214 titles: 9, 4 per cent. Fourth-century philosophical titles coincide with Plato's subject-titles at a base rate of one in twenty — everyone wrote a Περὶ ψυχῆς — and the "lost dialogues" coincide at seven times that rate. The match is not the commonplace overlap of period titles; it concentrates exactly where the tradition says the lost dialogues are.[^11] One of the Theophrastan hits is itself a datum of the kind #1581 collects: Περὶ ἡδονῆς ὡς Ἀριστοτέλης — On Pleasure, as Aristotle('s) — a title with the other name inside it, beside Περὶ ἡδονῆς ἄλλο.
So the paper claims the shape: seventeen of twenty-six entries mapped by a rule the catalogue itself states, at seven times the base rate, with two entries naming their source outright, five of the remainder catalogued again under Theophrastus, and the reverse direction handed over as Aristotle says. It does not claim that Περὶ ψυχῆς in Aristotle's catalogue is the Phaedo. Where fragments exist that question has a test, named below and not run; where they do not, it has none, and the shape is all there is — which is why the shape is the datum.
The paper claims that the matched titles match, that the extracts entries say what they say, that the Divisions are handed over as Aristotle's report, and that a two-author tradition needs these ghosts while a single-corpus history would leave them as residue. It does not claim that Περὶ ψυχῆς in Aristotle's catalogue is the Phaedo, or that any ghosted title is any extant text. That is a further claim, and it has a test, which is named here and not run.
Where a test exists, it is the fragments. The received scholarship reconstructs several of Aristotle's dialogues from later citation — Cicero on the Eudemus and On Philosophy, Iamblichus on the Protrepticus, the Neoplatonists on On the Soul — and the reconstructions are the evidence that the dialogues existed as distinct works.[^9] For each matched title, set the fragments attributed to the Aristotelian dialogue beside the Platonic dialogue of the same subject-title and ask whether the fragments are in it, are of its kind, or are neither. The test cuts both ways and no single hit decides it. A passage found verbatim in the candidate may indicate identity, or quotation, or shared material; a passage absent from the candidate may indicate a distinct work, or a different recension, a paraphrase, or a mistaken attribution. What discriminates is the citation's wording, speaker, argumentative sequence, setting, and extent of correspondence together, across the fragments of a title — a mismatch does not defeat identity by itself, and a match does not establish it. That is a collation, it is bounded, and the fragments are on the open shelf (Rose, Ross); the paper stops at its edge.[^10]
Two things do not follow and are not claimed. Not hands: a single corpus organized as two, a school, a workshop, and a constructed configuration all leave this residue, and the catalogues do not choose among them. Not dates: nothing in a title's filing orders anything in time. What follows is narrower and, once seen, hard to unsee: the two absences the received account treats as facts about two men are one filing operation seen from both sides, and the catalogue that records them records, in its own wording, where the missing works are.
Thrasyllus wrote two titles for every dialogue, one from the name and one from the thing. Aristotle's catalogue kept the things and lost the names — On the Soul, On Kingship, On Pleasure — and filed the results as works of Aristotle now missing; and twice it forgot to remove the name, and wrote extracts from Plato's Laws. Diogenes kept Plato's divisions whole and could introduce them only as Aristotle says. A tradition that needs two authors needs a dramatic Aristotle and an analytical Plato, and needs both to be lost. The catalogue does not show them lost. It shows them filed — under a rule it states, with candidates it names, on the side of the line it keeps them off.
[^1]: The developmental account — the dialogues as an early Platonic phase, the treatises as the mature departure — is the framework of Jaeger's Aristoteles (1923) and of the reconstructions of the lost dialogues in Rose's and Ross's collections of the fragments; it is stated here as the received reading and not engaged beyond what the catalogue itself shows. The catalogue's status as a witness is that of a receiving text (Appendix, and #1581).
[^2]: The unwritten doctrines as reconstructed from Aristotle's reports of Plato's principles (Metaph. Α 6, Μ–Ν, and the Aristoxenus anecdote of the lecture On the Good) are the Tübingen reading; stated here as received. The archive's earlier note on the partition's witnesses is The First Draft §19.
[^3]: The catalogues' stichometric totals are not read as measures of loss; the ruling is recorded at #1581 §5 and The First Draft §7.
[^4]: All Greek from the seated Diogenes Laertius (Perseus canonical-greekLit tlg0004.tlg001, seated at EA-CORPORA-12, #1580), read at the lines cited; translations mine. Thrasyllus's arrangement is reported at III.56–61; the double titles and the genre-terms are his, as Diogenes gives them.
[^5]: The count is of the catalogue's opening entries through Περὶ ἐπιστημῶν; the match is by subject-title, with the four name-titles (Σοφιστής, Μενέξενος, Συμπόσιον, Πολιτικός through Περὶ πολιτικοῦ) and the two extracts entries counted as direct. Περὶ φιλοσοφίας matches two Platonic subject-titles (Theages, Rival Lovers), both in the spuria; Περὶ εὐχῆς matches the second Alcibiades, likewise. Their presence in the tetralogies is Thrasyllus's, and the match is to his catalogue, not to any judgment of authenticity.
[^6]: Διαιρέσεις and Διαιρετικά stand in Aristotle's catalogue at V.23–24 (the line-numbering of the seated text); Περὶ τἀγαθοῦ α′ β′ γ′ at V.22. The Aristoxenus account of the lecture (Harm. II 30–31 Meibom) is the received anchor for On the Good as Plato's; the catalogue's entry is under Aristotle.
[^7]: The reception measurements are The Particle (#1574) and the captures of 2026-09-06–07 in the Capture Registry (conditions WARNED and the self-audit); the reflex operating on an author rather than a sentence is The First Draft §24.
[^8]: The seven marks are stated at The First Draft §24 from #1583 (the hinge), #1581 (the catalogue), #1586 (the first axis), #1584 (the operation), and §19 (the partition's witnesses).
[^13]: The reading of the residue as the orthonym's footprint was reached with a third reader (Gemini, 2026-09-07), whose re-indexing of the partition from two names to four functions — Socrates the unwritten origin inside the inquiry's ἡμεῖς, Plato the dramatic, Aristotle the demonstrative, Theophrastus the operational hinge — is the function × mode matrix of The First Draft §15 stated as a ledger; its structural point is adopted, its account of motive (an engine "unbearable to institutional governance") is not, on the line's rule that motive is not on the page.
[^12]: The transform was fitted by running each of Thrasyllus's twenty-six pairs against the three lists with a five-letter stem match on the subject and a name-slot classification (person / kind / work); coverage: 11 subjects in Aristotle's opening, 5 further in the rest of his catalogue, 10 in Theophrastus's, 7 in none. The re-lexicalization clause (T1) is read from the pairs where the stem matches but the form differs; Περὶ εἰδῶν καὶ γενῶν as the image of περὶ ἰδεῶν is a reading of that clause, not a stem match, and is marked as such.
[^11]: The control is reproducible from the seated Diogenes: Thrasyllus's πρᾶγμα-titles (III.57–61) reduced to stems; each list's titles reduced to their first stem after Περὶ; a hit where the stems coincide at five letters. The rule is deliberately stricter than the hand-matching of §2 (which reaches seventeen), so that the base rate is not inflated; the three lists are the opening of V.22 through Περὶ ἐπιστημῶν, the remainder of V.22–27, and V.42–50 whole.
[^10]: The counts of §2, the scope of §4, the qualification of the reverse direction, and the two-way form of the fragment test were corrected on a second reader's audit (ChatGPT, 2026-09-07), adopted on their terms; the reader's proposal for the paper's investigable form is kept as its own: the catalogue supplies a naming rule under which extant works become candidates for allegedly missing ones. Lee's standing reading of that reader — motivated to preserve the boundary, to be understood as such and not discounted — is recorded in the notebook.
[^9]: Rose, Aristotelis qui ferebantur librorum fragmenta (1886); Ross, Aristotelis fragmenta selecta (1955). Both are on the open shelf or can be seated from it; the collation of §5 is bounded by the matched titles of §2 and has not been run.
Diogenes Laertius, Vitae philosophorum III.56–61 (Thrasyllus's tetralogies and double titles), III.80–109 (the Divisions, ὡς Ἀριστοτέλης φησί), V.22–24 (Aristotle's catalogue, opening): Perseus canonical-greekLit tlg0004.tlg001, seated whole at EA-CORPORA-12 (#1580), read at the lines cited. The matching in §2 is reproducible from the seated text by extracting Thrasyllus's ὄνομα ἢ περὶ πράγματος pairs and Aristotle's Περὶ-titles and comparing subjects.