This paper aims to show that Burnet's text at Pl. Ti. 55c7–d6 is at least questionable, and that opting for a different reading at 55d5 would shed light on an intriguing argumentative aspect of Plato's cosmological account.

The Timaeus encompasses a direct argument for the uniqueness of the world: at 30c–31b, Timaeus points out that the uniqueness and completeness of the sensible world depends on the fact that it is an image of the complete intelligible living being.Footnote ¹ This, in turn, is explicitly related to the intellect which is the cause of the world, that is, the Demiurge: the reason why one can be certain that the world is shaped according to a perfect intelligible paradigm is that God is good and wants his creation to be as beautiful as possible. Hence the reader is struck when, at 55c–d, Timaeus says:

The puzzle is not only why Timaeus might want to provide a new argument in favour of the uniqueness of the world but, above all, that, while at 30c–31b he concluded without any doubt that there is only one world, here he seems to be sceptical about the firmness of this conclusion; after all, he seems to declare that according to the reasonable account (κατὰ τὸν εἰκότα λόγον) one might consider it true that five worlds exist. This difficulty notwithstanding, the text established by Burnet, which I have quoted in Cornford's translation, has been accepted by virtually all translators apart from Rivaud.Footnote ³ To the best of my knowledge, Burnet's choice has never been submitted to criticism, whereas I would argue that, from a philological point of view, it is at least questionable.

Let us first briefly consider the textual tradition. Burnet in his edition considered as primary witnesses for the Timaeus not only A (Par. gr. 1807, late ninth century), F (Vind. suppl. gr. 39, before 1355) and Y (Vind. phil. gr. 21, late thirteenth/early fourteenth century) but also P (Vat. Pal. 173) and Vat. (Vat. Gr. 228). However, Jonkers has recently demonstrated that the manuscript tradition of the Timaeus is based on seven primary witnesses, but that P and Vat. are not among them.Footnote ⁴ More specifically, besides A and F, there are: V (Vind. phil. gr. 337, fifteenth century, from 34b4), a gemellus of A but contaminated with ancestors of C and F; C (Tubingensis Mb 14, eleventh century), which is closer to F but has been contaminated with an ancestor of A and V; and, along with Y, Θ (Vat. gr. 226, early fourteenth century) and Ψ (Par. gr. 2998, 1273–83), which depend on the same antigraph as Y and are closer to the line of F, still characterized by contamination—this group is indicated by the letter g.Footnote ⁵ Hence anyone dealing with the text of the Timaeus ought to be wary of mechanically applying stemmatic principles: a reading found in only one manuscript may be correct against all other primary witnesses, and editors ought to take indirect testimonies into account too. As a consequence, in our passage we should consider the evidence not just of Philoponus but also of the Armenian translation, which was produced on the basis of a manuscript related to A.Footnote ⁶

In our passage both θεός and θεόν are well attested. C's support of A cannot be decisive, since it may be due to contamination. V's agreement with A testifies to the fact that text A was in their common, late antique antigraph, and the support of the Armenian translation confirms (at least) that this is how the line was read.Footnote ⁷ On the other hand, the antiquity of text F is confirmed by Philoponus. The omission in g can be reasonably explained by referring to the grammatical difficulty presented by text F (to which I shall now turn), and in any case cannot be regarded as a plausible option.

No reading is clearly difficilior, but from a linguistic point of view text F presents an evident problem, since it lacks a subject which is usually supplied by referring to δόγμα. This word, however, occurs some lines earlier and in a completely different syntactical context. One could also consider τὸ μὲν οὖν δὴ παρ’ ἡμῶν, coordinated with ἄλλος δέ, as a circumlocution for ἡμεῖς; but the phrase is unnatural in Plato,Footnote ⁸ and the presence of οὖν δή makes this option even less plausible.Footnote ⁹ The syntax is much clearer if, on the contrary, we adopt text A, where θεός is the subject of μηνύει.Footnote ¹⁰ However, this is not enough to make us opt for θεός, and the philosophical analysis proves crucial.

In favour of text F, it is tempting to invoke the principle according to which Timaeus’ εἰκὼς λόγος is intrinsically unstable and negotiable (presumably this is the reason why Burnet opted for this reading and translators maintained it): Timaeus would be conscious that he is providing just a reasonable (or, in this perspective, probable) account, so it is fair for him to leave room for other options (namely, the five-worlds hypothesis).Footnote ¹¹ This is a dangerous line of reasoning, though. If it were correct, we would be forced to project the same sceptical perspective onto the preceding statement about the uniqueness of the world (30c–31b), which is directly based on the Demiurge's goodness and the nature of the intelligible living being—that is, on intelligible objects, which are neither grasped through perception nor judged by belief. On the contrary, the argument at 30c–31b can be regarded as a strong foundation, since—as noted above—it involves assumptions related to the Demiurge's action and the causal role of the paradigm which could hardly be regarded as weak from a Platonic point of view. Moreover, recent discussions of the notion of εἰκὼς λόγος have modified the standard view of the epistemic value of Timaeus’ εἰκὼς λόγος, emphasizing that it might also be regarded as an account aiming to approximate to the Demiurge's practical reasoning.Footnote ¹²

All this makes a different solution desirable, and the solid attestation of text A makes it more than appropriate to consider what Plato would say if this were the correct reading. Text A states that the thesis of the uniqueness of the world is based on what the Demiurge reveals and on the nature of the world itself. The only way that the Demiurge can reveal the world's natural uniqueness is by spelling out the metaphysical reason why the world is such. As we have seen, this is exactly what Timaeus has done at 30c–31b: the uniqueness of the world depends on its being by nature an image of a paradigm, and this in turn is ensured by the Demiurge's goodness. Hence text A would repeat the argument presented at 30c–31b; this is an advantage from one point of view, in that it solves an apparent inconsistency in the composition of Timaeus’ account. On the other hand, we might wonder why Timaeus would simply rephrase the argument of 30c–31b, and in particular why he should hint that someone may have different views.

My answer to both questions is that in this way Timaeus is (paradoxically) stressing the non-negotiability of the correct conclusion, which is stronger than any other reasonable account. Let us first focus on the difference between the unreasonable hypothesis of uncountable worlds and the (allegedly) reasonable hypothesis of five worlds. The former is ascribed to people who have no experience of such discourses and is ruled out without any discussion, while the latter is said to be reasonable and plausible as a belief deriving from a different point of view from Timaeus’. The standard interpretation of the origin and logic of the five-world hypothesis is still persuasive: the reason why here—and here alone—five worlds are possible (or, at least, are said to be conceivable) is that now the reader's attention is drawn to the cosmological role of five regular solids.Footnote ¹³ Hence the difference between these two hypotheses is that the former is simply devoid of any mathematical grounds—from Plato's point of view, at leastFootnote ¹⁴—while the latter can rely on a specific logic, that of the importance of mathematical structures in the physical world. Within certain limits, this is indeed a logic which Timaeus endorses. However, while it may lead to the development of some consistent beliefs, it cannot be used as the only parameter for a cosmogony, which can be consistently construed only if the stable mathematical reasoning underlying it in turn depends on even more stable and valuable premises, namely, the Demiurge's goodness and reasoning and the intelligible model. It is in this sense that we can say that the five-world hypothesis is reasonable, yet epistemologically inferior to that of a single world: an abstract consideration of physics might lead to the five-world hypothesis, but since the number of the elements and their appropriate geometrical structures are established by the Demiurge, the Demiurge's very reasoning is ontologically and axiologically prior to any account based solely on geometrical physics.

One wonders whether this—that is, requiring confirmation from God—may not be too much for Plato. However, at 72d4–9 we are presented with a similar scenario. After his description of the tripartition of the soul and the related connection of the ‘parts’ of the soul to different organs, Timaeus explicitly states that the only criterion allowing us to be certain of the truth of his conclusion would be God's approval (τὸ μὲν ἀληθὲς ὡς εἴρηται, θεοῦ συμφήσαντος τότ’ ἂν οὕτως μόνως διισχυριζοίμεθα). In this case, however, such approval cannot be received, and Timaeus limits himself to relying on his reasonable account. However, the reason why one cannot have God's approval is related to the specific issue which Timaeus is dealing with here. Indeed, the generation of mortal living beings is not produced directly by the Demiurge but by the lower Gods, who are called to imitate the Demiurge's reasoning and action, though dealing with much more unstable objects (41a7–43a6). In other words, the Demiurge cannot give his confirmation here since he is not the divinity who produced human beingsFootnote ¹⁵—and, more generally, since the objects at stake (that is, individual living beings) have a lower ontological status than the world itself. This, though, does not apply to the generation of the world, which is directly accomplished by the Demiurge, whose reasoning has been clearly and fully unravelled at 30c–31b: briefly, God has already confirmed the metaphysical reasons why there is just one perfect world.Footnote ¹⁶

Hence text A is not only (at least) as well attested as text F philologically and more suitable syntactically but also much more effective philosophically. Keeping text A generates a significant philosophical pay-off: not only does Timaeus’ account of the uniqueness of the world become much more consistent, but it also confirms that his reasonable account (and, hence, Plato's cosmology) in any case relies on strong metaphysical premisses, which limit the instability of our grasp on the sensible world as far as possible and make Timaeus’ account stronger than (or at least not inferior to) any other reasonable account (29c7–d3).