Two Greek texts seated as one corpus so that the four openings of a planned facing-page edition resolve to bytes on both sides, together with the plan of that edition. Theophrastus, On First Principles (First1KGreek tlg0093.tlg006, 3,370 tokens) and Aristotle, Metaphysics Λ (Perseus tlg0086.tlg025 book XII, 5,071 tokens), each whole, each with its own numbering — Usener as carried by van Raalte for the fragment, Bekker for Λ. The four openings, every locus verified against the seated bytes: (A) the axiom, fragment 7 b 19-22 δοκεῖ δὲ … πάσχειν against Λ 7's τὸ πρῶτον κινοῦν ἀκίνητον εἶναι καθ᾽ αὑτό, where the fragment adduces αἴσθησις against the received axiom and where the Greek manuscripts split four ways (μὴ ἀεὶ Wimmer, μὴ δεῖ PJ, μὴ δὴ CA, καὶ ὡς δεῖ L) while Laks-Most read μὴ δεῖν on the Arabic's authority — et ut vid. Isḥāq — so that the Arabic breaks the tie at the axiom; (B) the first demand, μέχρι μὲν δὴ τούτων οἷον ἄρτιος ὁ λόγος … τὸ δὲ μετὰ ταῦτ᾽ ἤδη λόγου δεῖται πλείονος, a prior account marked complete to a point and the next question named; (C) the third demand and the method refused, τὸ δὲ κατὰ τὸ πλῆθος τῶν σφαιρῶν τῆς αἰτίας μείζονα ζητεῖ λόγον· οὐ γὰρ ὅ γε τῶν ἀστρολόγων, against Λ 8's derivation of the number of the spheres from Eudoxus and Callippus; (D) the landing, εὐλογώτερον οὖν εἶναί τινα συναφὴν καὶ μὴ ἐπεισοδιῶδες τὸ πᾶν, against Λ 10's ἐπεισοδιώδη τὴν τοῦ παντὸς οὐσίαν ποιοῦσιν at 1076 a 1 — the word van Raalte cites as a lexical parallel and which the edition prints as the place the aporiai arrive. The seat also records an editorial hazard found in the course of the work: the Greek Unicode range contains the koronis U+1FBD, so a tokenizer built on that range keeps elided forms as two-character tokens and any elision-normalization table silently fails to match them; a difference-of-centroids direction fitted on these two texts loaded entirely on δ᾽/ἀλλ᾽/δι᾽ against δέ/ἀλλά — on edition, not on voice — until it was caught. The plan of the edition is carried in the body: what it is, why an edition and not a paper, the four openings, text and apparatus policy, commentary policy, the two outstanding acquisitions (Alon 1985 for Isḥāq at the hinge lemmas; Gutas 2010 for the Arabic entire), the order of work, and what the edition would and would not settle.
Lee Sharks · plan v0.1 · 2026-09-08 · not the edition; the specification for it
A facing-page critical presentation of two Greek texts in the order their own argument exhibits, with a commentary that states the relation at each locus rather than adducing it as a parallel.
Recto: Theophrastus, On First Principles (the work known as his Metaphysics), whole — 4 a 2 to 12 a 2 in Usener's numbering as carried by van Raalte.
Verso: the passages of Aristotle, Metaphysics Λ, that stand opposite each demand — principally Λ 6–8 (1071 b 3 – 1074 a 31) and Λ 10's close (1075 b 37 – 1076 a 4).
The claim the layout makes, and the only one it makes: the fragment holds the position of the question and Λ the position of the attempt. Where the fragment says an account is needed, Λ supplies one; where it says the astronomers' account will not do, Λ 8 is the astronomers' account. This is an order of inquiry, exhibited by the texts. It is not a claim about hands, dates, or composition, and the introduction says so before it says anything else.
The relation cannot be stated in the notation the discipline uses for relations between texts. A commentary that cites by cp. can produce a hundred parallels and has no form in which to write and here the second answers the first. Van Raalte cites Λ five times inside the hinge commentary — 1072 b 19-30, 1073 a 5-7, 1074 b 25-7, 1075 b 20-4, and 1076 a 1, which is ἐπεισοδιώδη, the word the fragment's aporiai land on. Every citation is correct and every one is a comparandum. The seam is unrepresentable in that apparatus, and a paper arguing so would be answered with more parallels.
A facing-page edition does not argue the point. It puts the demand and the supply on the same opening and lets a reader with Greek decide. That is a form the objection cannot be raised against, because the objection is about notation and the edition changes the notation.
The commentary is dense at four places and quiet everywhere else. These are the openings the edition exists for.
Opening A — the axiom. Fragment 7 b 19-22 ‖ Λ 7, 1072 a 23-6.
δοκεῖ δὲ … πάσχειν. The received axiom is that the first mover is unmoved. The fragment adduces αἴσθησις against it: in the imparting of movement, subject and object need not be distinct. Apparatus: μὴ ἀεὶ Wimmer · μὴ δεῖ PJ · μὴ δὴ CA · καὶ ὡς δεῖ L · μὴ δεῖν Laks–Most, et ut vid. Isḥāq · κινεῖ MSS., κινεῖται Usener. Five readings of two words; the Greek splits four ways; the Arabic breaks the tie, and it breaks it at the axiom.
Opening B — the first demand. Fragment 7 b 23 – 8 a 7 ‖ Λ 7–8 transition.
μέχρι μὲν δὴ τούτων οἷον ἄρτιος ὁ λόγος — up to this point the account is, as it were, complete. Then: τὸ δὲ μετὰ ταῦτ' ἤδη λόγου δεῖται πλείονος περὶ τῆς ἐφέσεως, ποία καὶ τίνων — what comes after this now needs a fuller account. A prior account marked complete to a point, and the next question named. Van Raalte on this stretch: "the text of this passage is distressingly corrupt, a fact that probably is not unconnected with the complexity of the reasoning." Her summary of the chapter is quoted whole in the introduction, because it is the counter-reading at its strongest.
Opening C — the third demand and the method refused. Fragment 8 a 21-3 ‖ Λ 8, 1073 b 17 – 1074 a 14.
τὸ δὲ κατὰ τὸ πλῆθος τῶν σφαιρῶν τῆς αἰτίας μείζονα ζητεῖ λόγον· οὐ γὰρ ὅ γε τῶν ἀστρολόγων — the account of the number of the spheres seeks a greater λόγος; the astronomers' will not do. Facing it: Λ 8's derivation of the number of the spheres from Eudoxus and Callippus, fifty-five or forty-seven. The fragment names the method Λ 8 runs on and rules it insufficient. The commentary notes, without leaning on it, that Λ 8 has been the most-suspected chapter of the book since Jaeger, on internal grounds, and that the suspicion falls where the demand points.
Opening D — the landing. Fragment 11 b 24 – 12 a 2 ‖ Λ 10, 1075 b 37 – 1076 a 4.
εὐλογώτερον οὖν εἶναί τινα συναφὴν καὶ μὴ ἐπεισοδιῶδες τὸ πᾶν — more reasonable that there be some connection and that the whole not be episodic. Facing it: Aristotle's ἐπεισοδιώδη τὴν τοῦ παντὸς οὐσίαν ποιοῦσιν. Van Raalte cites this at p. 376 as a lexical parallel for a phrase at 11 b 25-6. The edition prints it as the place the aporiai arrive.
Base text. The fragment: van Raalte's text (Brill, Mnemosyne Suppl. 125, 1993), with Usener's paragraph and page numbering as she carries it, since it is the numbering the commentary literature uses. Λ: the Oxford text, with Bekker numbers.
Apparatus. Printed at every locus where a reading is contested, not only at the four openings — the credential is completeness, not selectivity. Sigla and manuscript matters follow Laks–Most, as van Raalte's do. Each entry records: the manuscripts, Ross–Fobes, Usener, Wimmer, Sylburg where relevant, Laks–Most, van Raalte, and Isḥāq where he speaks.
The Arabic. Isḥāq ibn Ḥunayn's version descends from a Greek uncial outside the surviving stemma and is therefore a witness, not a translation to be consulted afterwards. It is cited as a siglum among the others. Source: Alon, "The Arabic Version of Theophrastus' Metaphysica", Jerusalem Studies in Arabic and Islam VI (1985) 163–205, which is where van Raalte reads it, and Gutas (Brill, Philosophia Antiqua 119, 2010) where Alon is insufficient. Neither is on the archive's shelf yet; see §6.
Nothing conflated. Two texts, two apparatuses, two numbering systems. The edition never prints a composite.
Quiet, and at loci. The commentary's whole business is to say, at each of the four openings, what stands opposite and in what relation — demand and supply, method named and method used, word and word. It does not characterize either author, does not adjudicate hands, and does not use the vocabulary of influence, borrowing, dependence, or reply-to.
Three rules, from the archive's own practice:
1. Relations established, readings withheld — the register of #1581 and #1592. The commentary states what stands where and stops.
2. What is not claimed, stated at full strength. Three counters get their own section in the introduction, each in its strongest form before any answer: (i) aporetic demands are ordinarily raised against a standing position, which is what dialectic does and what οὐ γὰρ ὅ γε usually signals; (ii) both texts may draw on a shared school discussion, in which case neither is prior and the demand answers something we do not have; (iii) priority in the order of inquiry is not priority in composition.
3. The best counter-reading quoted whole. Van Raalte, p. 361: "Chapter V in two respects modifies the provisional conclusion formulated at the end of Chapter I." She sees the reopening motion exactly, and reads it as self-correction across chapters of one work. That is what this edition says instead of, and it is printed in her words, in the introduction, before the case is made.
Missing — two acquisitions, both small.
Not missing.
1. Seat the two texts as one corpus, aligned at the four openings, with van Raalte's numbering keyed to the fragment and Bekker to Λ. This is the next action.
2. Acquire Alon 1985; collate Isḥāq at 7 b 19-22, 8 a 2-7, 8 a 8-14, 8 a 21-3, 11 b 24 – 12 a 2.
3. Draft the four commentary openings.
4. Draft the introduction: the order, the warrant, the three counters, van Raalte quoted whole.
5. Complete the apparatus across the fragment.
6. Deposit; then offer.
Would settle: that the two texts can be presented in the order of the inquiry without violence to either, that the demands and their supplies fall in that order at four loci, and that the relation is invisible in the notation the tradition uses to record relations between texts.
Would not settle: how many hands wrote the corpus, when either text was composed, or whether the demand precedes the supply in time rather than in argument. The edition makes an order visible. It does not date it.