Why “What is F?” does not by itself give you a Form: the five steps that get you there, the two claims Plato leaves unargued in the middle dialogues, and the one place he states them as a conclusion
Failing to say what F is does not lead to a Form by itself. It leads there only through further premises — and even then the chain delivers a definiendum that is not sensible, not an object that exists, let alone one that exists separately.
Steps 1 and 2 (the priority of definition; a definition that never fails) are Socratic. Step 3 (compresence) and step 4 (so the definition is not of anything sensible) are argued in the Phaedo and Republic. Step 5 hides two further claims — that there is such an object, and that it is separate — which the middle dialogues agree to, hypothesise or posit, but do not argue.
The one place Plato states that Forms exist as the conclusion of an argument is Timaeus 51b–52a: if understanding (nous) differs from true belief, “these things do by all means exist, themselves by themselves” (51d; trans. Moss 2021: 224).
Charith’s own questions and arguments are labelled as his. Quotations were checked against the books in the library, and page numbers are the printed ones. Anything reported from general knowledge rather than a held source is marked reported.
Does the Socratic failure to say what piety or justice is lead, by itself, to Plato’s Forms? This part writes the route out as an argument, one premise per step, and asks of each step what it establishes and who argues for it.
The first arrow is Socrates’ own: to live rightly one must know what the right is. The second arrow does not follow. Knowing what justice is requires a definition; nothing yet requires a Form. A correct definition could in principle be an account of a concept, of a property that sensible things share, or of a separately existing Form. Which of these it has to be is exactly what the argument must settle.
So the step needs further premises. Written out in full, the chain runs:
Each step is marked by its status in the dialogues: whether it is a Socratic requirement, a premise Plato argues for, or a claim he asserts without argument.
Steps 1–4 are valid as far as they go, but their conclusion is a condition on the definiendum: whatever F is, it is not a sensible thing or property. That is not yet an object. Step 5 adds existence (“there is such a thing as the F”) and, on the usual reading, separation (“it can exist whether or not the sensibles do”). Neither follows from steps 1–4. Dancy has to state existence as a separate premise (ARE) of his own reconstruction (2004: 18), and Fine insists that “Difference does not imply separation” (1993: 60).
The first arrow in the chain is Socrates’ own. It creates the need for a definition. It does not yet say what the definition is a definition of.
(IA) To know that . . . F —, one must be able to say what the F, or F-ness, is. Dancy 2006: 72. He calls this the “Intellectualist Assumption”, “often referred to elsewhere as the ‘Socratic Fallacy’ or the ‘Principle of the Priority of Definition’”, citing Benson 1990 and 2000: 112–63. Texts: Euthyphro 4d9–e8, 5c8–d5, 6d6–e7, 15c11–e1.
Socrates, Dancy adds, “gives us a reason for insisting on it when we are trying to determine how to live” (2006: 72). That reason is the first arrow of Charith’s chain.
“(A) that if you know you are correctly predicating a given term ‘T’ you must ‘know what it is to be T,’ in the sense of being able to give a general criterion for a thing’s being T; (B) that it is no use to try and arrive at the meaning of ‘T’ by giving examples of things that are T.”
It gives a need for a definition. It does not give an object. Dancy puts the existential side of it exactly: “It is clearly a presupposition of the whole enterprise that there are things to be defined” (2004: 14). A presupposition of the enterprise is not a conclusion of any step in it. That is why, in Dancy’s reconstruction, existence re-enters later as a premise, not as a result (§05).
(IA) has a cost that the Meno makes visible: if one cannot know anything about F without its definition, how can one search for the definition at all? Dancy reads the Meno as the place where Plato gives (IA) up: “He is, it appears, now prepared to allow that we can use terms in the absence of an explicit definition, and that, when definitions are required, our way of getting at them is due to our prenatal grasp of what is to be defined” (2006: 81).
The Phaedo then attaches recollection to Forms: “our ability to answer ‘what is it?’ questions … depends on our prenatal acquaintance with a special realm of objects not encountered by the senses. These objects are the Forms” (Dancy 2006: 81–82). On this route, the Forms enter as the objects of a prenatal acquaintance, not as the conclusion of the five-step chain.
What Socrates asks of an answer. These requirements are already in the early dialogues. Together they rule out any answer that is sometimes F and sometimes not.
“1. There must be just one form of piety for all pious things (actions, people, and so on) (5d1-5). 2. It must be usable as a paradigm that we can consult to say whether something is pious or not (6e3-6). 3. It must be describable without explicit reference to properties—just and unjust, fine and shameful, good and bad—that introduce disputes (7c10-d7). 4. The single ‘form’ (eidos) or ‘character’ (idea) of piety is that ‘by which all pious things are pious’ (6d9).”
The fourth feature does the work. If the form is what makes pious things pious, it must never also make things impious. Irwin states the requirement Plato later makes explicit:
“A property G cannot be the explanation of x’s being F if either (1) G is present in y, but y is not-F or (2) G is not present in z, but z is F (97a5-b3, 100e8-101b2).”
Dancy locates the same requirement inside his Argument from Relativity (§05): “Premise (ARF) figures as a criterion of adequacy for a correct definition, and instances of (ARO) are used in showing that various proposed definitions fail to meet that criterion” (2004: 19).
Are these requirements already a theory of Forms? Allen says yes. Socrates “assumes, in pursuing his inquiry, that there is an ἰδέα, or εἶδος, a Form, of holiness”, a universal, a standard and an essence: “These assumptions constitute a theory of Forms” (1970: 67–68). “Logically, Forms play a regulative role in dialectic: as antecedents of ‘it’ in the question ‘What is it?’ they determine the kinds of answer which are acceptable, and more importantly, unacceptable” (68). Dancy says no: conceding that there is such a thing as the F “is not a piece of metaphysics, but simply a matter of fixing the topic for discussion” (2004: 19). The argument of this chapter does not need to settle that dispute. On either reading, step 2 is a condition on answers; it does not yet say that the answer names a separately existing object.
The first premise about sensible things, and the only step Plato argues from what we observe. It says nothing yet about Forms.
The texts are Phaedo 74b7–c5 (equal sticks and stones), Republic V 479a5–b10 (the many beautifuls are also ugly), Hippias Major 289a2–c6 and 293b5–e5, Symposium 211a2–5, and the “finger” passage at Republic VII 523a–524d.
“Plato attributes ‘compresence of opposites’ (as some modern writers call it), not merely compresence of opposite appearances, to the many Fs. … A beautiful (kalon) girl is also ugly (aischron) in comparison with gods (Hp.Ma. 289a2-c6). Burying our parents and being buried by our children are fine (kalon) in some circumstances and shameful (aischron) in others (293b5-e5). The many beautifuls (justs, equals, and so on) are both beautiful and ugly (Rep. 479a5-b10).”
Aristotle reports that Plato posited Forms because sensibles are in flux (Met. A.6 987a32–b10; M.4 1078b12–32; M.9 1086a37–b11). The three commentators agree that the flux in question is mostly compresence, not change over time:
| Scholar | Page | Claim |
|---|---|---|
| Dancy 2004 | 19 | “Flux as it appears in Plato is not straightforward change as you and I conceive it: baldly put, flux is relativity.” |
| Irwin 1999 | 169 | “If his arguments appeal only to compresence, not to change over time, we are justified in concluding that the type of flux he attributes to sensibles in arguing for nonsensible Forms is compresence.” |
| Fine 1993 | 57 | “in the middle dialogues Plato takes compresence but not succession to require the existence of forms.” |
If there were no compresence, step 3 would fail and this route to the Forms would close. Plato grants as much for some predicates. In Republic VII, sight never reports that a finger is also not a finger, so “finger” does not provoke the question “What is it?”. Fine: “only some predicates have sensible instances that suffer narrow compresence—‘thick’ and ‘thin’ are such predicates, but ‘finger’ is not” (1993: 59). Irwin draws the consequence: “no Forms for (A) are mentioned; and if the argument about equals relies on the problems raised in Republic VII, Plato should not require them” (1977: 150).
If the contradiction itself showed what the thing is, you would have the Heracleitean view: justice just is returning and not returning what you borrowed. Plato’s reply, on Irwin’s reading, is that “the Heracleitean confuses different embodiments of justice in different circumstances with the property that is embodied in these different ways” (1999: 170). The reply still needs step 2: a single property that is never also its opposite.
The conclusion of the compresence argument. It is a claim of difference: what F is, is not the same as any sensible F.
Dancy’s reconstruction makes the logical form explicit. It is the Phaedo 74b–c argument generalised; at 74b4–c6 the F is the Equal (2006: 82):
“Although the Argument from Relativity first appears in the Phaedo, the materials for its construction are to be found in the Socratic dialogues” (Dancy 2004: 19). The existential premise sits above the argument. The argument does not produce it.
“1. Equal sticks and stones sometimes being the same appear equal to one man, and not to another (74b7-10). 2. But the equals themselves have never appeared to you unequal, or equality inequality (74c1-3). 3. Then these equals (sc. the sticks and stones) and the equal itself are not the same (74c4-5).”
Fine reaches the same result by way of explanation: “Since some sensible properties of F suffer compresence, reference to them does not explain why anything is F, and so they cannot be what F-ness is. Since explanation is possible, in these cases things are F in virtue of a non-sensible property, the form of F” (1993: 58). And she marks the limit: “The argument from compresence shows that forms are different from both sensible particulars and sensible properties” (60). It shows that they are different, and no more than that.
A category condition: whatever answers “What is F?” is not a sensible thing or a sensible property. The conclusion is a non-identity — (ARC), “not the same as”. It is not an existence claim and not a separation claim. In the three-level terms of §11, steps 2–4 argue at level 2.
Step 5 says more than step 4. “So there is a non-sensible object that is never unjust” contains two claims that steps 1–4 do not supply.
Yes: step 3 does not mean the Forms exist separately. Step 3 is only about sensible things. Even step 5 gives you a non-sensible object that is different from them, not one that could exist without them. That extra step needs a further premise.
| Scholar | Pages | In their words | Existence is… |
|---|---|---|---|
| Allen | 1970: 67–68, 81, 106–7 | “existential import is taken for granted, not demonstrated” (81); “the existential assumption is precisely the assumption that real definitions may be found” (107). In general: “‘One over Many’ is the nub of the argument for their existence” (1960, §IV, in Allen ed. 1965). | Assumed |
| Vlastos | 1991: 63–64, 76; 1954: 320 | Once “regularly conceded … without contest”, existence becomes “a highly contestable thesis”: a “posit” in the Republic, a “hypothesis” in the Phaedo (63), that is, “an unasserted premise” (64). Plato’s reason for “a single Form”, reconstructed: “(A1) If a number of things, a, b, c, are all F, there must be a single Form F-ness, in virtue of which we apprehend a, b, c, as all F” (1954: 320). | Posited |
| Irwin | 1977: 145; 1999: 143 | “if we believe just men are just by justice, we must recognize something called ‘justice’ or ‘the just’” (1977: 145). “Throughout the Platonic dialogues the character Socrates assumes the existence of ‘forms’” (1999: 143). | Assumed |
| Fine | 1993: 50, 58 | “the possibility of knowledge requires explanation, and this, in turn, requires the existence of forms—real properties and kinds” (50); “Since explanation is possible, in these cases things are F in virtue of a non-sensible property, the form of F” (58). | Argued |
| Dancy | 2004: 18–19; 2006: 81 | (ARE) is “the Argument from Relativity’s Existential presupposition” (2004: 18); conceding it “is not a piece of metaphysics, but simply a matter of fixing the topic for discussion” (19). In the Phaedo the Forms first appear “without argument” (2006: 81). | A premise |
Only Fine reconstructs an argument whose conclusion is existence, and her argument runs from the possibility of knowledge and explanation (a transcendental premise), not from compresence alone.
| Scholar | Page | On separation |
|---|---|---|
| Fine 1993 | 60–61 | “The argument from compresence shows that forms are different from both sensible particulars and sensible properties. But it does not show that forms are separate … Difference does not imply separation” (60); “Plato never argues, or even says, that forms are separate” (61). |
| Vlastos 1991 | 75–76 | Separation is the Form’s “existential independence of any actual or possible instantiation of it in the world of time” (75). Plato asserts it “feeling no need to offer a separate argument for it, since the existence of his Forms independent of any other constituent of reality … is the immediate consequence of their essence” (76; cf. 74 n. 137, “the essence of Beauty entails its own existence”). |
| Irwin 1977 → 1995 | 149; 1995: 375–76 n. 32 | In 1977, “The Phaedo uses the argument from A-flux to justify the separation of Forms” (149). By 1995: “None of the passages we have discussed offers any explicit basis for Aristotle’s claim that Plato ‘separated’ Forms from sensibles”. Aristotle at Met. 1086b7–10 “suggests that Plato took it for granted” that proving Forms “apart from” (para) sensibles was enough to prove them separate. |
| Dancy 2004 | 13 & n. 47, 247 | Plato “makes these forms distinct from perceptible things … which is at least part of what Aristotle means by saying that Plato ‘separated’ the forms” (13). Aristotle means “not just a matter of the forms being distinct … but of their being, we might say, radically distinct” (13 n. 47). The Imperceptibility Thesis “makes this a theory: the things about which Socrates has been asking are to be distinguished from, detached from, or even ‘separated’ from perceptible things” (247). |
| Devereux, in Fine ed. 1999 | 201 | In the Phaedo (e.g. 78d5–6, 100b5–7), “themselves by themselves” is “not accompanied by an explicit denial of immanence”. |
That sentence (Dancy 2006: 70) reports Aristotle; it is not Dancy’s argument. Its “or” joins two different claims. Distinct means that the F is not the same as any sensible F, which is what the Argument from Relativity proves: “(ARC) The Beautiful is not the same as any ordinary beautiful [thing]” (2006: 71). Separate means that the F can exist whether or not any sensible F exists (Fine 1993: 51), which nothing in the argument addresses.
The 2004 book shows he means “distinct”: distinctness is “at least part” of separation (13); separation rests on imperceptibility (247); and the flux argument is called “an argument for the existence of separate Forms” (14), though its conclusion is that “the forms are distinct from perceptible things” (14). The slide goes back to Aristotle: Met. A.6 “says only that flux … shows that forms are different (hetera …) from sensibles; separation is not mentioned” (Fine 1993: 60). Fine’s “Difference does not imply separation” applies directly to Dancy’s “or”.
The logic is right. “Sb ≠ FB” does not show that FB exists. One can talk coherently about FB and its relation to Sb — as a standard, a concept, or an ideal — without FB being in the ontology. That is why Dancy must add existence as a separate premise (ARE), and why Fine says that difference arguments do not give separation.
Plato raises the same worry himself, at Timaeus 51c: “Is it in vain on each occasion that we say there is a certain intelligible Form of each thing, and is this nothing but a word?” (trans. Prior 1985: 113). The argument at 51d–e is aimed at exactly that possibility (§08).
One correction: Plato does say that Forms exist in other places — Phd. 65d, 100b; Rep. 596a, and some twenty more passages (§07) — but as agreements, hypotheses, posits and the like. Only Timaeus 51d–e makes existence the stated conclusion of an argument. It is therefore the strongest evidence of Plato’s commitment, though not a proof that Forms exist, and it cannot by itself tell us what the early dialogues were committed to.
An existence premise (there is such a thing as the F) and, for Plato’s Forms, a separation premise (it can exist without sensibles). Steps 1–4 supply neither. In the middle dialogues, existence is agreed to, hypothesised or posited, and separation is asserted, typically through “itself by itself”, without argument. The next part looks at where Plato says it and at the one place where it is the conclusion of an argument.
Plato says that Forms exist in many places. The literature argues over four of them: in three the existence claim is agreed to, hypothesised or posited, and only in the fourth is it the conclusion of an argument. About twenty more passages, from the early dialogues to the Seventh Letter, concede it, posit it, reveal it in a speech or a myth, raise it as a question, or argue for it only conditionally.
The table below gives the four passages the literature argues over. Tables 1–3 collect what scholars say about the first three, and §08 takes the fourth. Tables 4–6 add about twenty more passages found in a search of the library.
| Passage | What Plato says | How existence enters | Argued? |
|---|---|---|---|
| Phaedo 65d | “Do we say that there is something itself just, or nothing? Of course we say [there is], by Zeus.” (tr. Dancy 2004: 246) | Agreed to, at once | No |
| Phaedo 100b | “hypothesizing that there is a beautiful itself by itself and a good and a tall and all the others” (tr. Dancy 2004: 300) | A hypothesis | No |
| Republic X 596a | “we are accustomed to posit a single Form for each plurality to which we apply the same name” (tr. Vlastos 1981: 86 n.) | A posit (whether it even asserts existence is disputed) | No |
| Timaeus 51d–e | “these things do by all means exist, themselves by themselves” (παντάπασιν εἶναι καθ’ αὑτὰ ταῦτα; tr. Moss 2021: 224) | The conclusion of an argument | Yes |
| Book | Page | What they say Plato does here |
|---|---|---|
| Dancy 2004 | 247 | Finds two claims: “(TF1) The things about which Socrates has been asking ‘what is it?’ exist; and (TF2) … are not perceptible.” It is (TF2) “that makes this a theory.” |
| Dancy 2006, in Benson ed. | 81 | “we first encounter the Forms … at 65a9–66a10, without argument” |
| Vlastos 1991 | 64 | Existence was introduced “by a question which had anticipated and received instant, enthusiastic, agreement” |
| White 1976 | 65 | Socrates “induces … Simmias, to agree without argument that there are such things as the Just Itself” |
| Natorp 2004 | 154 | “the existence of conceptual determinacies themselves is taken as absolutely established” |
| Prior 1985 | 30 | “he first gets the assent of Simmias to the existence of ‘something Just itself’” |
| Allen 1970 | 148 | “eliciting Simmias’ agreement that there exists such a thing as justice, alone by itself” |
| Irwin, in Fine ed. Plato 1 | 144 | These are “precisely the ‘beings’ or ‘essences’ … that Socrates sought to define”. What is new is that they are “inaccessible to the senses”. |
| Nehamas, in Fine ed. Plato 1 | 174 | “Socrates directly introduces the Forms as intelligible entities by the example of the just itself” |
| Vlastos 1981 | 347 n. 22 | Lists Phd. 65D among the “formulaic assertions of the existence of Forms” |
| Kahn 1981, reported in Moss 2021 | 94 n. 18 | In ousia here, “Veridical idea, existence claim, and predicative syntax are all taken up” |
| Book | Page | What they say |
|---|---|---|
| Dancy 2004 | 300 | “it is pretty clearly the claims that are hypothesized”: “(HF) There is such a thing as the F”, which he calls “the Theory of Forms” |
| Dancy 2004 | 291 | At 100b7–9, “the existence of the Forms carries immortality with it” |
| Vlastos, “Reasons and Causes”, in Vlastos ed. 1971 | 139 | A “‘hypothesis,’ that of the Theory of Forms”, whose content is “each of the Forms exists” |
| Vlastos 1991 | 63–64 | “In the Phaedo he calls it a hypothesis”, i.e. “an unasserted premise” |
| Fine, Introduction to Plato 1 | 22 | The hypothesis that there are forms is “justified to some extent” by 74b–c, “But the justification is incomplete” |
| Crombie 1963 | 530 | “His hypothesis is … that there are forms and that nothing comes to acquire a property except through the presence of a form to it” |
| Natorp 2004 | 154 | Introduced “without any proof”, yet “continually taken as ‘adequately proven’ premises” |
| Benson, in Benson ed. | 93 | “He identifies the hypothesis that the Forms exist (100b5–9), and infers from it…” |
| Ferejohn, in Benson ed. | 155 | Socrates “asks Cebes to reaffirm his earlier commitment to the existence of the Forms” |
| Devereux, in Fine ed. Plato 1 | 201 | Here “themselves by themselves” is “not accompanied by an explicit denial of immanence” |
| Silverman 2002 | 58 | Participation is tagged “as a consequence of the hypothesis that there are Forms” |
| Prior, in Benson ed. | 28 | One of two passages where “Socrates claims to be the author of the theory”, which conflicts with Aristotle’s account |
| Teloh 1981 | 122 | “The second voyage introduces the Forms as causes (aitiai, 100b)” |
| Book | Page | What they say |
|---|---|---|
| Vlastos 1991 | 63 | Existence, once “regularly conceded”, becomes “a highly contestable thesis. In the Republic he calls it a ‘posit’” |
| Vlastos 1981 | 347 n. 22, 355 | Among the “formulaic assertions of the existence of Forms” |
| Robinson & Denniston, in Vlastos ed. 1971 | 9 | “This form is the very thing itself meant by the name” |
| Wedberg, in Vlastos ed. 1971 | 29 | “postulating one unique Idea for each plurality of objects to which we apply a common name” |
| Cross & Woozley, in Vlastos ed. 1971 | 84, 94 | Their translation reads “assuming the existence of a single … Form”. Classifying things and having general words “can only be accounted for by the existence of the Forms”. |
| Teloh 1981 | 88, 122 | Forms are “introduced on at least four occasions as ‘each being one’”. Plato “introduces one Form to explain the occurrence of each distinct type of nonsortal quality”. |
| Fine 1993 | 111, 117 | Reports the view that it is “a semantic one over many argument”. Her own reading: “the Socratic procedure of positing a single form of F to explain the common feature”, with nothing yet about separation. |
| Fine, Introduction to Plato 1 | 19 | It “need not commit Plato to the view that there is a form corresponding to every predicate” |
| Cohen, in Fine ed. Plato 1 | 296–97 | “his most famous formulation” of the one-over-many principle |
| Owen 1957, in Allen ed. 1965 | note | “The one passage in which Plato seems unequivocally to require a Form for every predicate”; but “taken literally it contradicts Politicus 262a–3e” |
| Lloyd 1954, in Allen ed. 1965 | note | “Plato never said there is an Idea corresponding to every general name … Rep., 596a says εἰώθαμεν τίθεσθαι” (“we are accustomed to posit”) |
| Crombie 1963 | 282 | “phrases which seem to suggest the argument from the one over many … But is this clinching?” |
| White 1976 | 27 n. 40 | “it only says that we posit a Form for each ὄνομα … not why we should do so” |
| Prior 1985 | 60 | A Form “for every multiplicity of things having a common name … He apparently later came to reject this criterion” |
| Hampton 1990 | 98 | Whether it is “any more than a procedural recommendation or hypothesis is unclear” |
| Silverman 2002 | 334 n. 6 | “The accustomed method is to hypothesize that each Form is some one thing” |
| Sayre 2006 | 42, 237 n. 9 | A “customary procedure”; “the language of positing” marks “the beginning phase of inquiry” |
| Shields, in Benson ed. | 409 | “at least … a semantic purport, if not also a metaphysical one”; it would be “injudicious … to rely upon just this one passage” |
| Meinwald 2016 | 256 | Plato “did not mean … to present a dogmatic and official view concerning either the extent of Forms or the criteria for positing them” |
| Marmodoro 2021 | 78 n. 8 | Two possible readings. Existential, “‘we posit that there is an εἶδος’”, is the majority view and hers. Predicative, “‘we posit that the εἶδος is one’”, is held by Smith 1917, Burnyeat and Sedley 2013. |
| Harte, in Fine ed. Oxford Handbook 2019 | 463 | On the usual reading, the range of Forms becomes “wide to the point of potential absurdity”; but “the passage need not be translated in this way” |
Two library books do not cite any of these passages: Dancy 2004 never cites Rep. 596a or Tim. 51, and Fine 2014 cites none of the four. Fine’s closest remark concerns the Meno: she thinks it “unlikely” that Socrates is “offering a semantic argument for the existence of forms” (2014: 44).
At 65d, existence is agreed to without argument. At 100b it is hypothesised; Dancy’s (HF) has the same form as his (ARE). At 596a it is “posited”, and scholars even dispute whether the sentence asserts existence at all. Of these four, only Timaeus 51d–e is read as an argument that concludes existence.
Yes, many. The search turned up about twenty more passages: in eleven other dialogues and the Seventh Letter, and further passages in the Phaedo, Republic and Timaeus. None of them does what Timaeus 51d–e does. In each one the existence of Forms is conceded, agreed, posited, revealed in a speech or a myth, raised as a question, or argued for only conditionally.
Plato’s words in tables 4–6 are from Cooper (ed.) 1997, with the page in brackets.
| Passage | What Plato says | How existence enters | Discussed by |
|---|---|---|---|
| Prt. 330c1–2 | “Is justice a thing or is it not a thing? I think it is.” (763) | Conceded at once | Vlastos 1991: 60; Dancy 2004: 65, 68 |
| H.Ma. 287c–d | “And is this justice something?”—“Very much so.” … “by that being something?”—“It is. Why not?” Then at once: “what is that, the fine?” (905) | Conceded, and the “What is it?” question follows immediately | Vlastos 1991: 59–60 (“Then this [thing], justice, exists?”); Dancy 2004: 78 (“the third or fourth existential admission in this passage”) |
| Euth. 5d, 6d–e; Meno 72c | “that form itself that makes all pious actions pious” (6); “one and the same form which makes them virtues” (873) | A form is assumed, but “exists” is never said | Allen 1970: 67–68; Moss 2021: 208 (“no evidence that he is yet working with the full-blown theory of Forms”) |
Dancy on this group: “Later, these existential admissions will be seen as admissions of the existence of forms” (2004: 65).
| Passage | What Plato says | How existence enters | Discussed by |
|---|---|---|---|
| Phd. 74a9–b1 | “we say that there is something that is equal … the Equal itself. Shall we say that this exists or not?—Indeed we shall, by Zeus” (64–65) | Agreed; it is the first premise of the recollection argument | Allen 1970: 148 (“initial premise”); Dancy 2004: 268; Prior 1985: 11; Fine 2021: 124 |
| Phd. 76d7–77a5 | “an equal necessity for those realities to exist, and for our souls to exist before we were born”; “Nothing is so evident to me … as that all such things must certainly exist” (67) | Taken as evident; what the argument concludes is that the soul existed before birth | Allen 1970: 156; Dancy 2004: 291; Ross 1951: 24 |
| Phd. 102a10–b2 | “it was agreed that each of the Forms existed” (88) | A recap of the 100b hypothesis | Ross 1951: 15; Dancy 2004: 306 |
| Rep. V 475e–476c, 479a | “each is one … and all the forms” (1102); the sight-lover “doesn’t believe in the beautiful itself” (1103, 1106) | Glaucon agrees; the sight-lovers deny it | Allen 1970: 149 (“begins by assuming the theory of Forms”); Prior 1985: 11; Ross 1951: 37 |
| Rep. VI 507b | “a single form of each, believing that there is but one” (1128) | Posited, as at 596a | Teloh 1981: 88; Moss 2021: 208 |
| Smp. 210e–211b | “First, it always is … itself by itself with itself” (493) | Revealed in Diotima’s speech | Schofield, in Fine ed. Oxford Handbook 2019: 59 |
| Phdr. 247c | “a being that really is what it is … visible only to intelligence” (525) | Told in a myth | Teloh 1981: 125; Harte, in Fine ed. Oxford Handbook 2019: 470 |
| Passage | What Plato says | How existence enters | Discussed by |
|---|---|---|---|
| Cra. 439c–440d | “Are we or aren’t we to say that there is a beautiful itself …?”—“I think we are”; “if there is always that which knows and that which is known … it doesn’t appear to me that these things can be at all like flowings”; “whether I’m right … isn’t an easy matter to investigate” (155) | Agreed first. Then a conditional argument (if there is knowledge, there are unchanging objects), which is left undecided | Irwin, in Fine ed. Plato 1: 149; Fine 1993: 136; Kahn 2013: 58; Schofield, in Fine ed. Oxford Handbook 2019: 59 (“For anything like that, we have to wait until the end of the Cratylus”) |
| Rep. V 476e–480a | Socrates’ argument with the sight-lover | An argument for existence only on Fine’s reading | Fine 2003: 21, 81 (“if the sightlovers will agree that there is knowledge, they will now accept the existence of forms”) |
| Prm. 130b | “certain forms themselves … likeness itself is something, separate from the likeness we have?”—“I do indeed” (364) | Young Socrates says yes, and Parmenides then attacks the view | Prior 1985: 58; Kahn 2013: 4; Fine 2003: 34 n. 71 |
| Prm. 132a1–4 | “whenever some number of things seem to you to be large … you conclude that the large is one” (366) | A one-over-many inference, credited to Socrates and then attacked | White 1976: 6 |
| Prm. 135b–c | “if someone … won’t allow that there are forms … he will destroy the power of dialectic entirely” (369) | Conditional: without forms there is no dialectic | White 1976: 6; Teloh 1981: 169; Silverman 2002: 3; Kahn 2013: 58 |
| Phlb. 15a–b | “Firstly, whether one ought to suppose that there are any such unities truly in existence” (403) | Whether they exist is itself posed as a problem | Fine, Introduction to Plato 1: 27; Kahn 2013: 185 |
| Sph. 246b–c, 248a | “true being is certain nonbodily forms” (268); the “friends of the forms” say being and coming-to-be “are separate” (269) | Other people’s view, which the Visitor criticises | Crombie 1963: 258; Kahn 2013: 106; Prior 1985: 128 |
| Plt. 285e–286a | “the things that are without body, which are finest and greatest, are shown clearly only by verbal means” (329) | Stated as what dialectic is for | Kahn 2013: 148 |
| Tim. 27d–28a | “What is that which always is and has no becoming …?” (1234) | Laid down as a starting premise; argued for only at 51b–e | Johansen 2004: 49; Johansen, in Fine ed. Oxford Handbook 2019: 290 |
| Ep. VII 342a–b | “the object itself, the knowable and truly real being” (1659) | Asserted, but the letter may not be by Plato | Ross 1951: 62; Prior 1985: 60 |
Timaeus 51d–e is still the only passage where existence is the stated conclusion of an argument. Cratylus 440 and Parmenides 135b–c are “if … then” arguments that never assert the conclusion. Republic V argues for existence only on Fine’s reading, and Parmenides 132a states an inference only to attack it. White makes the same point: “nowhere in Plato’s works … does one find what could be called a full-fledged argument for the existence of Forms fully spelled out. There are fragments of such arguments in, for example, Phaedo 73-74 and the much later Parmenides 132a and 135b-c” (1976: 6). He does not mention Timaeus 51.
The early concessions are the existence presupposition said out loud (level 1 in §11). In the Hippias Major the “What is it?” question comes straight after. Vlastos contrasts existence “regularly conceded … without contest” in these dialogues with “a highly contestable thesis” in the middle dialogues (1991: 63).
Separation. Only two of these passages call the Forms “separate”: Parmenides 130b and Sophist 248a. Fine says 130b does not use the word “in the relevant sense” (2003: 34 n. 71).
Left out. Republic X 597c–d argues that each Form is unique, not that it exists. Aristotle’s reports of Platonist arguments, as in the Peri Ideōn, are not in the dialogues; Fine 1993 examines them.
The one passage where the existence of Forms is the conclusion of an argument. Timaeus raises the rival view before answering it: that talk of Forms is “nothing but a word”.
The argument comes after a test. At 29b Timaeus lays down that accounts of what is stable are “stable and unshifting”. At 49b he asks which of fire, water, earth and air can be named “in a way that employs a reliable and stable account”. None can: they keep turning into one another, so each is only “what is such”, never “this” (49d–e). Asked “What is it?” of a shape moulded in gold, the safest answer is “gold” (50a–b). Only the receptacle, which “does not depart from its own character in any way” (50b), can be called “this”, and what enters it are “imitations of those things that always are” (50c). That is what raises the question at 51b. (Translations here are Zeyl’s, 2000.)
So the what-is-it test is run on bodies first, and the bodies fail it. Only then does Plato argue that the Forms exist. Kahn: “Perhaps his only direct argument for the existence of Forms is in Timaeus 51D–E … In general, however, Plato’s arguments proceed from the Forms, not to them” (1996: 330–31). Authorship, date, the disputed lines 49c7–50b5 and the state of the Timaeus folder are taken up in a separate working study, Reading the Timaeus.
Is there any such thing as Fire just in itself, or any of the things of which we are always speaking thus: that each of them is just in itself? Or do those things at which we look, and all other things which we perceive by the body, alone have this sort of reality, but there is nothing else besides these in any way at any time? Is it in vain on each occasion that we say there is a certain intelligible Form of each thing, and is this nothing but a word? Timaeus 51b7–c5, trans. Prior 1985: 113. Crombie reads the rival view as holding that universals are “only ‘expressions we use’ (logoi)” (1963: 35).
If reason and opinion are different things, then Forms imperceptible by us, objects of thought alone, must exist by themselves; but if true opinion differs in no wise from reason, the things we perceive by means of the body must be deemed to be the most secure realities. Now we must say that reason and opinion are two things, since they differ both in origin and in nature. The one is produced by teaching, the other by persuasion; we can justify the one by true reasoning, the other is unreasoned; the one cannot be moved by persuasion, the other can; only God and a few men possess the one, all men possess the other. We must therefore agree that there are Forms distinct from the copies of them. Timaeus 51d3–52a4, as given by Ross 1951: 124. Ross: “Plato here treats as the essential reason for believing in the existence of Forms the difference between knowledge and true opinion” (124–25).
Broadie sets out the logic as two questions: the “criterial question” — “are intelligence and true opinion two kinds (or is there only true opinion)?” — and the “original question”, whether there are Forms of fire and the like. “The answer, affirmative or negative, to CQ will yield the corresponding answer to OQ” (2012: 275). Pitteloud writes it as modus ponens, but with a weaker conclusion: “(P1) If intellect and true opinion are two different kinds (duo genē), then Forms are different from sensible objects. (P2) Intellect is different (chōris) from true opinion. (C) Forms are different from sensible objects” (2025: 189 n. 11).
The move from two faculties to two kinds of object. Kahn calls it “Plato’s characteristic realism: how could two cognitive faculties be distinct if they did not have distinct objects?” (2013: 195). Pitteloud: “The truth of (P1) is not evident as it is possible to conceive of two different faculties of the soul relating to the same object” (2025: 189 n. 11). Crombie: “it does not pretend to be conclusive. It rests on the admittedly controvertible fact that understanding is not the same thing as right opinion” (1963: 521). Annas: Plato “does not consider the option that our conception of knowledge might not answer to anything” (2003: 84). Against Crombie, Vlastos calls the argument “as pure an example of ‘deductive metaphysics’ as one could hope to find” (1981: 376).
Mostly yes. In the early and middle dialogues the Forms’ existence is assumed, agreed to or hypothesised (Phd. 65d, 100b; Rep. 596a), and arguments like Phd. 74 show only that the F differs from sensible things. The clear exception is the late Timaeus (51b–e). Rep. V 476e–480a comes close.
The scholars agree. Prior: “In the early and middle dialogues, he does not regard the existence of the Forms as subject to serious philosophical doubt” (1985: 113). Crombie: “the doctrine that there are forms is nowhere the object of what purports to be a deductive argument. Usually it is put forward as an article of the Socratic faith. There is indeed an argument for it in the Timaeus (51–2)” (1963: 521). Teloh: “Plato never questions that Forms exist” (1981: 180). On Republic V, Cross & Woozley call the argument “implicit in Book V” (in Vlastos ed. 1971: 90), and White thinks Rep. 477ff. “more substantial” than the Timaeus (1976: 25 n. 33).
About twenty scholars in the library use the passage. Quotations were checked against each book; the search covered every text-searchable PDF in the library.
| Scholar | Where | How they use it |
|---|---|---|
| Cherniss 1936 | in Vlastos ed. 1971: 20, 23 (also in Allen ed. 1965) | “an avowedly brief and casual proof of the separate existence of Ideas”; later, “the epistemological proof of the existence of Ideas” |
| Cherniss 1944 | 211, 213 n. | “If knowledge is other than sensation and opinion … they must have a real and independent existence (Timaeus 51b–52a …)” |
| Ross 1951 | 103, 124–25 | “Plato here treats as the essential reason for believing in the existence of Forms the difference between knowledge and true opinion” |
| Crombie 1963 | vol. 2: 35–36, 521 | Sets out the argument (“therefore there exist self-consistent universals”), then rejects it as not conclusive |
| Cross & Woozley 1964 | in Vlastos ed. 1971: 72, 90 | “This argument for the Forms is implicit in Book V of the Republic. It is put clearly and explicitly in the Timaeus (51D)” |
| Vlastos 1966 | Platonic Studies 1981: 376 | Against Crombie: “as pure an example of ‘deductive metaphysics’ as one could hope to find” |
| White 1976 | 25 n. 33 | “more like a statement of the conclusion of an argument than the argument itself. More substantial is Rep. 477ff.” |
| O’Brien 1984 | 144 | “the argument which Plato offers briefly, but explicitly … If knowledge cannot be reduced to true opinion, then there must exist forms” |
| Prior 1985 | 113–14 | “He affirms the existence of the Forms … The answer is perfunctory and disappointing in its brevity” |
| Kung 1989 | in Anton & Preus eds.: 320 | “no science and no knowledge at all without positing the existence of other abstract items … as indeed Plato himself says in the Timaeus (51d3–7)” |
| Vlastos 1991 | 264 | The existence of Forms becomes “a question of burning importance which must be raised and answered de novo (Ti. 51B–C)” |
| Fine 1993 | 137 | “Plato argues that knowledge and true belief differ in various ways that require the existence of forms… Admittedly, Plato does not spell the argument out in any detail” |
| Kahn 1996 | 330–31 | “Perhaps his only direct argument for the existence of Forms is in Timaeus 51D–E”; Phaedo 74B–C argues only “the non-identity of Forms and sensible things”; “In general, however, Plato’s arguments proceed from the Forms, not to them” |
| Annas 2003 | 84 | Forms are “implied by our recognition of the differences between knowledge and true belief”, but see the critics above |
| Johansen 2004 | 74, 167 | “the discreteness of true belief and nous… provides the evidence that there must be forms over and above the sense particulars” |
| Modrak 2006 | in Benson ed.: 138 | A narrower conclusion: invoked “to establish that Forms by themselves are not objects of perception” |
| Hartenburg 2011 | 101, 138 | “Plato bases an argument for the existence of the forms on the difference between knowledge (nous) and true belief”; here “he makes it explicitly and does not assume it” |
| Broadie 2012 | in Patterson ed.: 275 | The answer to the criterial question “will yield the corresponding answer” to whether there are Forms |
| Kahn 2013 | 185, 195 | “The proof of the existence of Forms at 51b–e”; the step from two faculties to two objects is “Plato’s characteristic realism” |
| Gerson 2013 | 15 n. 23 | “if νοῦς… is different from true belief, then Forms must exist. But if Forms exist, then materialism is false” |
| Rowe 2015 | xv | The claim that there is such a thing “in itself” is made “by Timaeus, presumably on Plato’s behalf” |
| Moss 2021 | 224, 235 | “epistemological and ontological distinctions stand and fall together” |
| Pitteloud 2025 | in Larsen & Politis eds.: 189 n. 11 | Modus ponens, but concluding only that “Forms are different from sensible objects”; “The truth of (P1) is not evident” |
If a definition is all that “What is F?” needs, why did Plato want an existing object as well? This part lists the work the existence claim does for him, asks whether the analogy with numbers in science helps, and maps the whole chain onto the three levels of the question.
On Plato’s own terms, the existing Form does work that a definition by itself cannot do. Remove it and each of the following goes missing:
| Without Forms… | Where Plato presses it | What the existing Form supplies |
|---|---|---|
| knowledge has no object | Rep. V 476e–480a | Knowledge is of what is; belief is of what is and is not. Fine gives the epistemological version: “Since knowledge in these cases is possible, there must be forms” (1993: 58). |
| truth becomes relative | Cra. 386d–e; Tht. 152a | A nature that things have in themselves, not “relative to us, and dragged by us up and down by our appearance” (Cra. 386e1–2, as quoted by Irwin 1999: 169). |
| nothing explains why F things are F | Phd. 100c–d | The cause: “it is because of the beautiful that all beautiful things are beautiful” (100d7–8, as quoted by Fine 1993: 58). |
| there is no standard to judge by | Euth. 6e3–6; Phd. 74d–75b | A paradigm “that we can consult to say whether something is pious or not” (Irwin 1999: 145). |
| learning is impossible | Meno 80d–e; Phd. 72e–78b | Objects of a prenatal acquaintance that recollection recovers (Dancy 2006: 81–82). |
| discourse collapses | Prm. 135b–c | Whoever denies the Forms “in every way destroys the power of discourse” (as quoted by Teloh 1981: 169). |
| knowledge and true belief collapse into one | Tim. 51d–e | If they are two kinds, Forms exist; if not, sensibles are “the most secure realities” (§08). |
Aristotle accepts the need and rejects the separation. He says “it is not possible to acquire knowledge without the universal” (Met. M.9, as quoted by Fine 1993: 60). But in his view “only substance particulars are separate”, and universals exist when and only when they are instantiated (Fine 1993: 61). So every item in the table can be met by a stable, non-sensible definiendum that is not a separately existing Form. The pressure runs from level 2 toward some level-3 theory of the definiendum, but it does not by itself select Plato’s theory.
Correct, at least in practice. Science works without anyone deciding whether numbers exist, and idealisations like point masses are openly treated as fictions. Field’s Science Without Numbers (1980) and van Fraassen’s The Scientific Image (1980) go further: they argue that mathematics and unobservable entities only have to be useful, or match what we observe, not be real reported.
The other side is the Quine–Putnam indispensability argument: you are committed to whatever your best theory has to say exists (Quine’s criterion, the one Prior uses) reported. That is the same kind of argument Plato makes at Parmenides 135b–c. Joan Kung ties it to the Timaeus as well: Plato “thinks we shall have no science and no knowledge at all without positing the existence of other abstract items in addition to perceptible ones, as indeed Plato himself says in the Timaeus (51d3–7)” (1989: 320). So in both cases the real question is whether the thing is truly indispensable.
Even if Quine is right, science would be committed to abstract objects, not to Forms. Modern abstract objects are not models that things copy, they are not instances of their own property (the number 2 is not two of anything), and they cause nothing. So here too the pressure goes from level 2 to some level-3 theory, not necessarily Plato’s.
The three levels of the ti esti question (see What Justifies the Question?) sort the five steps by what each one needs and what each one delivers.
The Socratic failure to say what F is creates a demand for a definition that no sensible thing can meet. That demand fixes a category, not an object: the existence of the Form is presupposed, its separation is asserted, and only the Timaeus states either as the conclusion of an argument.
Documents on Ti Esti that take up each step in more detail.
The chain question, questions 2–8, argument 1 (Sb and FB) and the three-level test are Charith’s. The answers and the organisation of the page were drafted with Claude from the conversation, and every quotation is the cited author’s own. Page numbers are the printed pages of the editions held in the library.