1. THE SOUL'S PRENATAL KNOWLEDGE

In response to Meno's Paradox (80d5–e5), Socrates introduces recollection with the following compact statement (81c5–9):

The prenatal seeing of ‘all things’ implies (ἅτε) that the soul has learned everything, which makes possible its post-natal recollection of ‘virtue and other things’. Socrates says that what we recollect post-natally is something we ‘knew earlier’ (πρότερον ἠπίστατο). So, however we understand this passage, we should grant that the soul knew prenatally and accordingly had prenatal knowledge. What, then, is the prenatal knowledge of?

Some have seen in the phrase ‘virtue and other things’ a covert reference to the Platonic forms. One might think, for example, that, since this is a disembodied (cf. 86a3–4) prenatal vision of a classic example of a Platonic form, these must be forms. On the other hand, Plato does not say that these are forms, nor does he use the language he typically uses to introduce forms, nor does he attribute traditional qualities of forms to these objects. On this basis, one might deny that these are forms.Footnote ⁶

I sympathize with each of these positions, but find it difficult to conceive of what ‘virtue and other things’ can mean if not abstract, immaterial objects, which might suggest that these are Socratic universals.Footnote ⁷ Aristotle tells us, however, that Socrates did not separate the universals (Metaph. 1078b30–1), and that whatever is ‘seen’ here appears to be ‘separate’. So, to account for this passage, we need abstract intelligible objects that correspond in some way to ‘virtue and the others’, which can be seen by the disembodied soul—learned and known—and yet are not the full-blown Platonic forms. While no solution will please everyone, I propose to treat these objects as ‘proto-forms’, halfway between Socrates’ and Plato's favourite (respective) entities. These entities share the Socratic requirement that they correspond to our definitions, but they are abstract and immaterial; also, to some extent, ‘separate’ from this world. They are important enough to require a different route or path to access, or originally acquire information about, them. Yet they are not quite the entities of mystical devotion that they are in the Phaedo.Footnote ⁸

What, then, happens when we ‘see’ a proto-form, and why would Plato call that ‘learning’ and ‘knowing’? A frequent suggestion, which I accept, is that the soul acquires mental content corresponding to these proto-forms through this vision.Footnote ⁹ The sense of ‘learn’, then, is accordingly something like ‘take in content’.Footnote ¹⁰ Why this counts as knowing is more complicated, especially because the Meno prescribes demanding criteria for knowledge, implying that a mere vision of something would be insufficient for this title (97c–98c). Some have accordingly thought that Socrates’ use of the word ‘know’ here is loose or imprecise,Footnote ¹¹ but his careful discussion of the distinction between knowledge and opinion at the end of the Meno makes this unlikely. Instead, we should think of the disembodied and embodied periods of the soul's existence as having different criteria for knowledge: the soul can be said to know in its disembodied life because it is not distracted by the body (cf. Phd. 65c), or because it is currently enjoying a ‘direct access’ to the proto-forms—or, perhaps, because there is no falsehood in the disembodied realm. Any of these options would explain why Socrates can treat the vision of proto-forms as knowing but then have more demanding requirements for knowledge in the post-natal, embodied period of the soul's existence.

These proto-forms, or the content derived from ‘seeing’ them, is what Socrates will soon say is ‘found again’ in recollection. We find again (and awaken and get back) what we ‘saw’ and ‘learned’ and ‘knew’ prenatally. The suggestion, then, is that there are special objects, proto-forms, that the soul sees in its prenatal, disembodied existence. This prenatal vision presumably endows the soul with some sort of mental content, content that refers to these proto-forms; and the acquisition of this proto-form-related mental content is, at this prenatal stage of the soul's existence, sufficient for ‘having learned’ and ‘known’.Footnote ¹²

2. ANEURISKEIN AS ‘FINDING AGAIN’

Continuing from the previous section's quotation, the initial statement of recollection suggests a close association between recollection and ‘finding again’ (ἀνευρεῖν) (81c9–d5):

According to Socrates, recollecting one thing implies the possibility of ‘finding again’ all the others. In this we can see that ἀνευρίσκω is used for a stage of recollection. For why should recollection of one thing help us find another thing, if the finding in question were not a type of recollection? Recollecting courage should not in principle help me find piety. If, for example, I were searching for my keys and spontaneously remembered (or ‘recollected’) what I ate for breakfast, that recollection of waffles would not normally help me find my keys. However, if we think of the ‘finding again’ (ἀνευρεῖν) as a specific stage of recollection, then the inference is straightforward: recollecting one thing, like courage, does help make more palatable the possibility of also coming to recollect piety—as incontrovertible proof in the occurrence of a miracle would make more palatable the belief in the possibility of further miracles. If one has recollected, one has evidence that recollection is possible.Footnote ¹³

The next and last use of the verb ἀνευρίσκω is in the dialogue between Meno and Socrates about the progress of the slave (84c10–12):

The verb ἀνευρίσκω can just mean ‘find the answer’ here, and that is how this passage is usually understood. But given the use of ἀνευρίσκω above, we should consider the possibility that this verb again refers to recollection or a particular stage of it. One reason to think this is that the present passage parallels three other exhortations for Meno to watch the slave ‘recollect’ with Socrates (81b6–7, 82e12–13, 84a2–b1). This therefore suggests that this passage should also be understood as an exhortation to watch the slave recollect. So, both uses of ἀνευρίσκω in the Meno seem to refer to a particular stage of recollection.

In the next two sections, I examine two distinct stages of recollection: a mental phenomenon of true opinions ‘popping into the head’ (similar to ‘aha!’ moments of memory recognition), and an epistemic procedure of converting a true opinion into knowledge.Footnote ¹⁴ Since Socrates will later imply that the slave lacks knowledge (85c2; cf. 85c10–d1), the slave has not recollected in the second sense; but since ἀνευρίσκω does refer to a particular stage of recollection, we should think of it as indicating the phenomenological or mental stage. As we will see further in the next section, the core of this stage is finding again or awakening mental content within oneself, the mental content derived from seeing the proto-forms.

3. ANAKINEIN AS ‘AWAKENING’ INNATE OPINIONS

This section argues that ἀνακινέω refers to the same phenomenological stage of recollection as ἀνευρίσκω. Unlike with ἀνευρίσκω, however, we here get a direct object, δόξα, and these δόξαι seem innate. If this is correct, then one important stage of recollection is finding again or awakening innate true opinions, which are presumably latent or unconscious before being found again or awoken. This is the crucial passage (85b8–d1):

In this passage, it seems like Socrates is introducing the view that δόξαι are innate.Footnote ¹⁵ For Socrates says twice that the δόξαι ‘were in’Footnote ¹⁶ the slave, which must refer to a moment prior to the slave's conversation with Socrates. For confirming with Meno that the slave had these opinions would be pointless if Socrates were not making the point that their presence ‘in’ the slave antedated the conversation. Nor could this prior moment of acquisition have occurred through teaching during an earlier, embodied period of the slave's existence, since Socrates later confirms with Meno that no one taught the slave geometry and ‘all the other disciplines’ (85e1–3). Since the slave demonstration is presumably supposed to be generalizable to all people, everyone should thus have the same δόξαι ‘within’ them innately. Coupling this with the previous two sections, then, we have good reason to think that these innate δόξαι are remnants of the soul's prenatal knowledge, which it received from seeing the proto-forms.

If these δόξαι are things the slave finds within himself, and if they are things he once knew (prenatally) and then forgot (but retained latently), we again have an instance of recollection in the phenomenological sense—finding again within oneself. Socrates here uses ἀνακινέω to refer to the act of ‘making explicit’ a latent opinion, an awakening from latency to explicitness. The dream analogy thus suggests that ἀνακινέω refers to the same phenomenological stage of recollection that we saw above in the discussion of ἀνευρίσκω: as dreams just come to us in sleep, so too do certain δόξαι in recollection.Footnote ¹⁷ Like the flash of insight that accompanies the ‘aha!’ experience of memory recognition (stereotypically explained through suddenly remembering the name of an actress that was on the ‘tip of one's tongue’), I take the point of the slave's discussion with Socrates to be that certain opinions just come to the slave, as if he were suddenly remembering them.Footnote ¹⁸ Socrates uses ἀνευρίσκω and ἀνακινέω to refer to this phenomenological stage of recollection.

The last part of the above passage discusses the ‘conversion’ of these δόξαι into ἐπιστήμη. It is here that we see why the slave lacks ἐπιστήμη and learn more about the later epistemic stage of recollection. While the slave has clearly recollected in some sense—since he has ‘found again’ and ‘awakened’ some δόξαι within him—he has not yet recollected in another sense, because he does not yet know (85c10–d11). He does not yet know because he has not been questioned ‘often and in many ways’, which is apparently required for the conversion of true opinion into knowledge. Most scholars think that Socrates later connects this passage to tying opinions down with a ‘reasoning about the explanation’ (98a3–8).Footnote ¹⁹ If so, the slave would presumably need to answer a variety of similar questions, but also understand exactly why the solution of finding the diagonal yields the correct answer. In the next section, we will see that Socrates refers to this second epistemic stage of recollection with the verb ἀναλαμβάνω.

4. ANALAMBANEIN AS ‘GETTING BACK’ PRENATAL KNOWLEDGE

Immediately after the passage cited in the previous section, Socrates says the following (85d3–11):

This passage is usually taken as conclusive proof of Plato's commitment to innate knowledge, not innate true opinion that can later become knowledge.Footnote ²¹ If the slave is ‘taking up/recovering knowledge out of himself’, or ‘taking up/recovering the knowledge in himself’, it is easy to see why commentators have supposed that knowledge is literally present in the soul waiting to be ‘taken up’: this translation suggests that the object is already there.Footnote ²² But treating ἀναλαμβάνω as ‘get back’ does not imply that the object ‘gotten back’ is present within the slave before it is gotten back, and it certainly does not imply that it is present in the same form.Footnote ²³ Taking the verb in this way also suggests a different connotation of the prepositional phrases ἐξ αὑτοῦ and ἐν αὑτῷ.Footnote ²⁴ So if we translate ἀναλαμβάνω as ‘get back’, nothing in this passage (or anything else in the Meno) implies innate knowledge.Footnote ²⁵

Thinking of ἀναλαμβάνω as ‘get back’ makes better sense of the flow of the text. We learned above that the slave saw the proto-forms in his prenatal existence, which Socrates called ‘learning’ and ‘knowing’ in the same breath. I interpreted this as the acquisition of mental content corresponding to the proto-forms: seeing the proto-forms leads to the acquisition of mental content corresponding to those proto-forms, the acquisition of which entitles the subject to be described as having learned and known the proto-form in question. We then saw that Socrates seemed to call this innate content δόξα during the slave demonstration, and he then outlined a procedure whereby these δόξαι can be converted into knowledge. After outlining this conversion process, Socrates referred to a future conversion of the slave's true opinions into knowledge. Since, if the above analysis is correct, the slave prenatally had knowledge but then lost it, the conversion of his true opinions into knowledge would be correctly described as an instance of ‘getting back’, which is why I have translated ἀναλαμβάνω in this way here. The slave is getting the original prenatal knowledge back, that is, getting back the title of ‘knowledge’ for the same content he possessed all along.

On the traditional translation of ‘take up’ or ‘recover’, when Socrates asks ‘Well then, with no one teaching him but only asking him questions, he will know, himself ἀναλαβὼν knowledge ἐξ αὑτοῦ?’, and Meno answers without hesitation ‘Yes’, nothing prepares Meno to assent so quickly. For Socrates has not indicated that we have innate knowledge, but only prenatal knowledge and innate δόξα. Thinking of ἀναλαμβάνω as ‘get back’, however, makes good sense of Meno's assent.Footnote ²⁶ For this comes immediately after the previous block quote, 85b8–d1, where Socrates indicated that the slave had innate true opinions and can (eventually) convert them into knowledge. Hence Socrates does not indelicately shift to speaking about knowledge from speaking about opinion, as if he were unconcerned with the terminology, but carefully transitions from the one to the other through the discussion about converting true opinion into knowledge.Footnote ²⁷ He discusses prenatal knowledge (81c5–e5), innate true opinions (85b8–c10), a conversion procedure (85c10–d1), and then discusses ἀναλαμβάνειν knowledge (85d3–11). This careful transition prepares Meno (and the reader) to answer Socrates’ questions. It accordingly makes sense to think of ἀναλαμβάνω as getting back prenatal knowledge. ἀναλαμβάνω refers to this later, epistemic stage of recollection that the slave has not yet achieved. There is accordingly no commitment to innate knowledge in the Meno.

This is one of the main advantages of my account over Bronstein and Schwab's recent position (n. 1). While I agree that Socrates is committed to innate content, I disagree about whether he ever calls it knowledge before the conversion procedure. They think he does at 85d3–11, whereas I deny it. While this might seem trivial, determining whether the Meno is committed to innate knowledge is important for a number of reasons. First, denying innate knowledge and maintaining that ‘knowledge’ only has one sense is important for the consistency of the Meno, for Socrates later claims that, if he knows anything at all, it is that knowledge and correct opinion are different (98b1–5).Footnote ²⁸ My view thus avoids implicitly charging Plato with being loose about his epistemic terminology in the very dialogue where he so carefully articulates the difference between knowledge and true opinion.

More importantly, determining what Plato's view of knowledge is in the Meno raises important questions about how it relates to the other dialogues. For example, if knowledge is not innate in the Meno, then that suggests a shift between the epistemologies of the Meno and the Phaedo, where knowledge does seem to be innate. Why might he suddenly have accepted a theory of innate knowledge when he previously did not? Could it perhaps have something to do with the (further) development of the theory of forms? Similarly, we might also wonder about the relationship between the Meno and the frequent Platonic view that knowledge requires an agent's awareness in some sense (for example Lach. 190c, Ion 532b–c, Symp. 202a). If the Meno denies innate knowledge, it maintains this common pattern, which makes the Phaedo's position even more unusual and raises questions about the ordering of the dialogues. This might imply, for example, that the Phaedo comes after the Symposium, since the latter still holds the ‘awareness’ requirement for knowledge. Such enquiries could have enormous implications for how we understand the relationship between knowledge and the Platonic forms, as well as how the dialogues relate to each other.

5. TYPES OF MENTAL CONTENT

I have attempted to solve a number of puzzles about the theory of recollection in the Meno by thinking of ἀναλαμβάνω as referring to a distinct (and later) stage of recollection than ἀνευρίσκω and ἀνακινέω, and by paying attention to Socrates’ labels for the soul's mental content at each stage. I have argued that (1) the soul acquires this content prenatally from ‘seeing’ the proto-forms and that Plato calls this content ἐπιστήμη during the soul's prenatal existence; (2) that it has this content innately and that Plato calls the content δόξα while it is latent or unconscious, and also (at least initially) when it is made explicit to conscious thought; and (3) that humans can convert this mental content back into knowledge (‘getting back’ their prenatal knowledge). In this section, I want to explore what this mental content might be.

We have little to go on. Socrates tells us that the soul saw ‘virtue and other things’ in its prenatal existence (§1), and we know that Socrates’ conversation with the slave had something to do with recollection (§§2, 3). We can conclude that the innate content, then, is at least ‘about’ (περί) virtue and other moral things, and probably also about mathematical and geometrical ‘things’.Footnote ²⁹ This raises the question ‘What can moral and mathematical mental content be?’

There are at least the following three types of mental content to consider. Plato might think that the content corresponding to any moral or mathematical/geometrical proposition is innate, which I call ‘full content’. He might consider innate certain principles explaining the relationships between numbers, geometrical objects, moral concepts and states of affairs instantiating those moral concepts, which I call ‘principle content’. Or he might think that the essences (that is, definitions) of certain concepts are innate, which I call ‘essential content’. I want to briefly explain these and then argue that Plato has something like principle and essential content in mind here.

The strongest version of innate content, ‘full content’, would state that any proposition of mathematics, geometry, or virtue is innate. On this view, there is an infinite number of innate δόξαι, as the following examples show:

• 4 is the sum of two prime numbers.

• 6 is the sum of two prime numbers.

• 400 is the sum of two prime numbers.

• … Helping one person is a virtuous action.

• Helping two people is a virtuous action.

• Helping three people is a virtuous action.

• …

A more moderate version of innate content might explain our mathematical and moral δόξαι according to general principles from which they might be derived, hence the term ‘principle content’. On this view, certain principles are innate. For example, Plato might take the following to be innate:

• Any even number is the sum of two prime numbers.

• Helping people is a virtuous action.

From these principles, one can consider the number four hundred and deduce that it is the sum of two prime numbers (for example 89 + 311), since it is an even number. Similarly, one can see that helping eight people is virtuous, because helping people (any number) is virtuous.

On the last formulation, ‘essential content’, what is innate, is information about the essences of things:

• Virtue is …Footnote ³⁰

• Triangles are three-sided figures.

The core idea here is simply that we have information about what certain things are, their essences, within us.

In what follows, I suggest that Plato thinks that our δόξαι about mathematicals and the virtues are innate both in the principle- and in the essential-content sense. What the slave has innately and ‘finds again’ and ‘awakens’, then, is essential and principle content corresponding to mathematical and geometrical proto-forms. This is as it should be, since the first option (full content) has a notorious problem in that it implies infinite innate content. While this is implausible on its face, it also makes Plato's account of innatism susceptible to a version of Malebranche's ‘retrieval problem’: if the soul contains an infinite amount of content, how can one ever traverse everything within oneself in order to find the specific thing one seeks?Footnote ³¹

6. WHAT THE SLAVE RECOLLECTS

I said above that I do not think that the slave recollects the answer in the geometry lesson.Footnote ³² Perhaps the biggest reason for this is that it would imply a ‘full content’ view and accordingly introduce an infinite number of innate items that the soul possesses at birth. For if that answer is innate, presumably every answer to any geometrical problem should also be innate. But there is no reason to attribute such an extreme view to Plato. For he uses the plural δόξαι in saying that the slave has opinions ‘in him’ that were ‘awoken’ in the conversation with Socrates. If the answer were awoken, we would need to choose something else that was not only in him but was also awoken in the conversation with Socrates to make sense of the plural. We could, perhaps, say that the wrong answers were recollected in addition to the correct answer and that this is why Socrates uses the plural δόξαι,Footnote ³³ but another answer suggests itself when we turn to the Greek.

We have seen that, in addition to using ἀνάμνησις/ἀναμιμνήσκω for both stages, Plato uses ἀνευρίσκω and ἀνακινέω for the phenomenological stage of recollection and ἀναλαμβάνω for the epistemic stage. When it comes to finding the answer, however, he uses ἐξευρίσκω (84b10). This suggests either that ἐξευρίσκω is a similarly technical term used to indicate one of the stages of recollection, or that the slave does not directly recollect the answer itself. But there is little reason to think that ἐξευρίσκω is used for a stage of recollection here. For, unlike the ἀνα- prefix, no interesting patterns involving the ἐξ- prefix emerge from analysing the recollection segment in the Meno, and the ἐξ- prefix on its own suggests a quite different sense. Smyth notes that the ἐξ- prefix often has the sense of ‘in consequence of’ or denotes the ‘cause or ground of judgement’;Footnote ³⁴ and the verb ἐξευρίσκω in Plato is often used to figure something out by means of something else (for example Prt. 353b1).Footnote ³⁵ I thus suggest that the answer itself is not found; rather, the slave ‘finds again’ and ‘awakens’ something within himself that he uses to ‘find’ or ‘figure out’ (ἐξευρίσκω) the answer.Footnote ³⁶ To do this, let us consider the slave's responses to Socrates, since these might show us what sorts of δόξαι were ‘found again’ and ‘awakened’ in his discussion with Socrates.

The slave gets the correct answer when Socrates draws a sixteen-foot (4 × 4) square composed of four four-foot (1 × 1) squares and shows him how a diagonal cut across a given square bisects its area. When Socrates then draws a diamond figure within the sixteen-foot square by placing diagonal cuts in each of the smaller four-foot squares, the slave understands that the diamond figure has an area of eight feet, and thus that the line they were looking for is the diagonal of a four-foot square.Footnote ³⁷ If we examine this discussion closely, Socrates does seem to have awoken a number of δόξαι that were used in the slave's progress towards the correct answer. Here is a sample of some of Socrates’ questions, which the slave answers affirmatively:Footnote ³⁸

1. 1. Do you know (γιγνώσκεις) that a square area is ‘like this’? (82b9–10)

2. 2. Does a square area have four equal sides, all four of them? (82b10–c2)

3. 3. Doesn't it have these two lines down the middle? (82c2–3)

4. 4. Can't such an area be larger or smaller? (82c3–4)

5. 5. Two feet and one foot would be once two feet? (82c8–d1)

6. 6. Two by two would be twice two feet? (82d4)

7. 7. How much is twice two feet? [The slave answers correctly.] (82d2–3)

(1)–(4) seem to be about the essence of the figure ‘square’. Socrates is ‘teaching’ the slave (or helping him recollect) that a square has four equal sides, essentially, and that the size is not important; instead, what is important is the relationship between the sides. (5)–(7) seem to be about the principle of multiplication, where Socrates is ‘teaching’ the slave how to multiply numbers. These are the types of δόξαι that seem to be elicited throughout the geometry lesson. If so, we should think of these δόξαι as ‘essential content’ and ‘principle content’. The slave has innate general δόξαι about mathematical principles and geometrical figures within him: Socrates awakens these in the discussion, and the slave uses these newly awoken δόξαι to get the correct answer.Footnote ³⁹

Such opinions are presumably included under Socrates’ exhortation to watch the slave recollect (82b6–7 and 82e12–13), suggesting that the slave has already begun recollecting and is continuing to do so.Footnote ⁴⁰ The key use of ἀνευρίσκω at 84c10, however, occurs after the slave has demonstrated his understanding of these mathematical concepts. This suggests that the slave (also) recollects in the phenomenological sense after he reaches aporia in the discussion with Socrates, that is, after demonstrating his facility with the relevant mathematical concepts. However, we still do not need to think that the answer itself is recollected. Given the tension between the innate δόξαι being plural but not wanting to attribute the full content view to Plato, it seems more reasonable to think that the slave is instead recollecting the general principle that diagonal cuts bisect squares’ areas, or that the way to find the (irrational number) ‘side’ of a square is to take the diagonal of its quarter. He may have also, as we saw above, recollected the very essence of ‘square’ or the principle of ‘multiplication’. These are reasonable candidates for the δόξαι found again, without implying that ‘full content’ view that the answer itself is innate.Footnote ⁴¹

This new interpretation of the slave using the content he recollected, or found again, to figure out the answer yields an interesting insight about Plato's model of mind. We should not see it as an ‘infinite storehouse’ of specific propositions, but rather as containing a large but finite set of principles (the number of which, I suspect, corresponds to the number of proto-forms).Footnote ⁴² The mind, in making judgements about particulars, consults the principle and essential content available to it, that is, the content that has already been awoken or found again. The essential content, for example, might say that ‘x is a triangle if and only if it has three sides’; perception says that ‘the thing in front of me has three sides’; and the thinker then judges that the thing in front of them is a triangle.

It is nevertheless difficult to see why Socrates would think that these δόξαι are innate. One possibility is that, newly infatuated with mathematics, Plato saw the multiple realizability of mathematical figures (that is, the fact that they can be instantiated in so many places and ways and shapes and sizes, all at once) and the infinite applicability of mathematical principles to be so amazing that they require an other-worldly origin.Footnote ⁴³ Or he may think that the slave demonstration was a mere analogy and that what is really innate is moral content. It is easier to see why one might think that the meaning of moral terms and the rightness of certain moral principles are innate: these are things that cannot really be taught, and yet many of us do understand them, both of which are prominent themes throughout the beginning, middle and end of the Meno.