BOOK I
In contrast with the rest of the Republic, book i strongly resembles a Socratic dialogue.  Socrates considers and rejects various conventional views about justice -- both Cephalus's law conception as telling the truth and repaying one's debts (331cd) and Polemarchus's more Homeric conception as benefiting one's friends and harming one's enemies (332a-336a).  These definitions suffer from the "compresence of opposites," the F and not-F problem.

1. Some G-things are F.
2. Some G-things are not-F.
3. Hence, F and G are not the same.

THRASYMACHUS
Like Callicles, Thrasymachus notes a tension between eudaimonism and the recognition of other-regarding virtues, such as justice.

1. Justice is a virtue.
2. Virtues benefit their possessor.
3. Conventional-justice often requires the agent to benefit others at her own expense.
4. Conventional-justice = justice.

(a) Justice is the advantage of the stronger (338c1-2);
(b) Justice is the advantage of the ruler (339a1-2); and
(c) Justice is another's good (343c1-3).

Notice that in attacking justice Thrasymachus makes two significantly different claims:

(i) Justice is not a virtue, because justice is not always more advantageous than injustice (348c).
(ii) Injustice is a virtue, because injustice is always more advantageous than justice (348b8-349a3).

BOOK II
What role do Glaucon and Adeimantus play?  How does their challenge compare with Thrasymachus's?  On the one hand, they make things easier insofar they require only the comparative claim that justice always be better than injustice, rather than Socrates's sufficiency claim that justice is sufficient for happiness (a complete good).

• Sufficiency: Justice is sufficient for happiness (a complete good).
• Dominance: Justice is always better than injustice.

On the other hand, Glaucon and Adeimantus make things harder for Socrates insofar as they retreat to Thrasymachus's weaker challenge to justice [(i) above].  Moreover, they require that justice be shown to be better in itself.  They distinguish three kinds of goods (357b-):

(a) things good for their consequences,
(b) things good in themselves, and
(c) things good both for their consequences and in themselves.

To make their demand clear, Glaucon and Adeimantus explain the instrumental value of justice.  I benefit from another's justice, not my own.  The reason I have to be just is that the justice of others is typically conditional on my own.  Others won't be just toward me if they see that I am not just toward them.  So in order to secure the benefits of other people being just toward me I must be just toward them.  In this way, justice is good for its consequences.  But this defense of justice is counterfactually unstable, as the ring of Gyges makes clear (359b-360d).  For then I could receive the benefits of others' justice without incurring the costs of my own.  But this shows that the conventional defense of justice is unstable; justice is really a second-best option behind the first-best option of practicing undetected injustice.  But this is to praise the appearance of justice, rather than justice itself (360e-362c, 362e-363e, 365b-367a).  Glaucon and Adeimantus demand a counterfactually stable defense of justice.

To understand counterfactual stability, imagine that we can rank possible worlds in terms of their value and that their value is a function of virtue and externals, including the appearance of justice or injustice.  Moreover, partition these worlds into just and unjust worlds, all those possible worlds in which the agent is just and all those possible worlds in which the agent is unjust.  Socrates must show that the worst just world (the world in which the agent is just but has the reputation for injustice and suffers other assorted bad fortune) is better than the best unjust world (the world in which the agent is unjust but has the reputation of justice and is the beneficiary of other assorted good fortune).  And this would be to show that justice in itself, independently of its attendant circumstances, is good.  This would show that justice is a b-good, as well as an a-good, which would show that it is in fact a c-good.

Is this demand coherent?  Can justice be shown to contribute to the agent's happiness without making its value depend upon some of its consequences?  Their demand is coherent if they distinguish between intrinsic and extrinsic consequences of justice.  For then Socrates can answer their challenge if he can show that justice is valuable for its intrinsic consequences.

JUSTICE
Socrates pursues the eudaimonist justification of justice indirectly, by looking at the justice, not of an individual, but of a city.  He appeals to a macro-micro analogy (368d, 441cd).

1. First, identify the tri-partite structure of the ideal city.
2. Next, interpret the political virtues as relations among the parts of the ideal city.
3. Next, provide an independent account of the structure of the soul that represents it as having a tri-partite structure isomorphic to the structure of the ideal city.
4. Therefore, conclude that the virtues for an individual agent stand to the structure of the soul as the political virtues stand to the structure of the ideal city.

JUSTICE AND THE DIVISION OF THE SOUL
Plato's tri-partite division of the soul appeals to a principle of opposites (436b).

1. The same thing cannot do or undergo contraries at the same time and in the same respect (436b).
2. The soul undergoes contraries at the same time: desire and aversion or rejection (e.g. for drink) (437b-c, 439b-c).
3. Hence, the soul must have different respects.
4. Hence, the soul must have parts.

At 437e-438a Plato distinguishes between desire for F and desire for F qua good; at 439c-d he distinguishes between rational and nonrational desires; and at 442bc he distinguishes desires that are good-independent and those that are good-dependent.  This will yield three, and not just two, parts of the soul if we distinguish

• optimizing desires (associated with the rational part)
• good-dependent but not optimizing desires (associated with the emotional part)
• good-independent desires (associated with the appetitive part)

Plato uses the division of the soul into rational (good-dependent) and nonrational (good-independent) parts to define (psychic) virtue as rule by the rational part of the soul.  What does he have in mind by control of the rational part?

• Instrumental control.  The rational part decides how to best satisfy existing desires; it determines only how to satisfy appetitive desires most efficiently.
• Directive control.  Reason forms desires as well as restrains existing ones; both formation and control of desires are guided by reasoning about the overall good of the agent.

A person is wise by virtue of his rational part exercising directive control over the other parts of his soul (442c).  A person is courageous by virtue of his emotional part honoring and enforcing the dictates of the rational part against his appetites (442bc). A person is temperate by virtue of his appetitive part obeying and agreeing in the rule by the rational parts (442cd).  Justice is this relation among the parts of the soul in which each part does its job (441de, 443d).  Might this account of psychic order be compatible with the inseparability and unity of the virtues?  Imagine the virtues as involving a single tri-partite pyramid structure just viewed from four angles: the bottom (temperance), the middle (courage), the top (wisdom), and the bird's eye point of view (justice).

A FALLACY IN THE REPUBLIC?
We can see why the proper ordering of an individual's soul -- psychic justice (p-justice) -- is good in itself.  But what does p-justice have to do with the demand to benefit others made by conventional justice (c-justice) and challenged by both Thrasymachus and Glaucon and Adeimantus?

Plato thinks c-injustice results from p-injustice (442e-443b; cf. G 507a-e).  But does all c-injustice arise from reason being a slave to the appetites?  Mightn't there be c-injustice that is not impulsive but well calculated?

DEVIANT HAPPINESS
The account of deviant men in Republic viii-ix is relevant here.  Whereas the desires of the appetitive and spirited parts of the soul reflect practical reasoning from a partial perspective, only the desires of the rational part reflect practical reasoning from an impartial and comprehensive perspective.  Because each part of the soul engages in reasoning, it is possible for Plato to think of each of them as if they were agents (cf. 588c-d).   In books viii-ix Plato examines the lives of deviant people -- the timocrat, the oligarch, the democrat, and the tyrant -- who correspond to the non-rational parts of the soul, precisely because, in different ways, they are dominated by the non-rational parts.

Only when the rational part rules, is a person's conduct regulated by a concern for each part of his soul as part of his overall good (586d-587a; cf. 442c).  Plato illustrates this claim by appealing to an image of the soul as consisting of a man, a lion, and a many-headed beast (588c-590d).  It is fitting for the man to look after the beast, with the help of the lion; only in this way does the beast get what's best for it (589ab).  If Plato literally identifies the rational part with a man, he introduces an apparently vicious regress, because the man himself must have three parts, including a rational part, which itself is a man, which in turn has three parts ....  But the point of thinking of the rational part as a man having a whole soul is that the rational part alone has a comprehensive or impartial concern for the whole soul, including each of its parts.  The rational part does not deny the appetitive and spirited parts but trains them and satisfies them as is appropriate for them as parts of an ordered whole (591c-592a).  Here the health analogy becomes important.  In caring for my whole body, I am concerned with the health and functioning of my bodily parts as parts of a larger, functioning whole.  I will not eat as much as I can, but as much as contributes to optimal functioning of someone with my frame and build.  Caring for my right arm will not consist in making it as strong as possible but in strengthening it in proportion to the rest of my body.  Similarly, in pursuing the good of the whole soul, one will gratify diverse appetites and passions insofar as their gratification contributes to a well functioning tri-partite soul in which reason rules.

The account of deviant men in books viii-ix explains why Plato thinks that unhappiness is not due simply to compulsive action on the appetites and passions.  He can argue that even calculated restraint in the service of honor, wealth, appetites, and lust does not secure happiness, because it does not represent a comprehensive concern for the various elements of the soul and their role in the agent's own good.  If so, Plato can claim that various sorts of c-injustice, not simply compulsive c-injustice, involve p-injustice.  But has he shown that all c-injustice involves p-injustice?  Moreover, if Plato is to avoid the fallacy of irrelevance, he must show not only that c-injustice results from p-injustice but also that p-justice produces c-justice.  Can he do this?  Will the p-just person not only refrain from harming her neighbors (i.e. avoid c-injustice) but also do those positive acts of beneficence that c-justice requires?  This depends upon the nature and results of directive rationality.

EROS AND ANOTHER'S GOOD
We get some picture of directive rationality if we look briefly at Plato's account of love (eros) in the Symposium (esp. 206e-212c) and Phaedrus (esp. 243c-257b).   The Republic invites this appeal in two ways.  (1) Sometimes eros is linked with sexual desire and the appetites, rather than the rational part (439d6-8).   But Plato also recognizes philosophical eros.  Genuine philosophers are lovers of truth (485a-d, 501d1-2; cf. Ph 68ab).  Indeed, the Symposium describes an ascent or progression from sexual desire to philosophical eros (204d-, 211c).  In the Republic Plato claims that the philosopher will not have consummated his eros until he has had intercourse with true reality and begotten intelligence and truth (490ab).  (2) But Plato also thinks that eros is other-regarding.  He remarks upon the love that a virtuous person will have for others (402d-403c, 412d).  Plato tries to connect (1) and (2).

Plato describes an ascent of desire through various stages (Symposium 210a-212a).

1. love of a particular beautiful body
2. love of bodily beauty, as such,
3. love of all beautiful bodies
4. love of spiritual beauty, that is, what is fine or beautiful in souls
5. love of fine laws and institutions
6. love of all kinds of knowledge
7. love of what is fine, as such

Such love benefits the beloved, because one benefits by becoming virtuous precisely insofar as one is better off being regulated by a correct conception of one's overall good.  But Plato also believes that the lover benefits from loving another (Phaedrus 245b), as he must if he is to reconcile interpersonal love with his eudaimonism.  The key to seeing how Plato can reconcile interpersonal love with self-love is to appreciate the way in which he thinks that reproducing one's virtuous traits in another is the next best thing to immortality (Symposium 206c1-209e5; Phaedrus 276e-277a).

Now although we speak of an individual as being the same so long as he continues to exist in the same form, and therefore assume that a man is the same person in his dotage as in his infancy, yet, for all we call him the same, every bit of him is different, and every day he is becoming a new man, while the old man is ceasing to exist, as you can see from his hair, his flesh, his bones, his blood, and all the rest of his body.  And not only his body, for the same thing happens to his soul.  And neither his manners, nor his disposition, nor his thoughts, nor his desires, nor his sufferings, nor his fears are the same throughout his life, for some of them grow, while others disappear.  ...  In this way every mortal creature is perpetuated, not by always being the same in every way, as a divine being is, but by what goes away and gets old leaving behind and in its place some other new thing that is of the same sort as it was [207d3-208b12].

1. I extend my interests by reproducing my valuable traits into a future self (intrapersonal reproduction).
2. The lover reproduces his valuable traits in his beloved (interpersonal reproduction).
3. Hence, the beloved extends the lover's interests.
4. Hence, benefits the lover confers on his beloved benefit the lover.

Does Platonic love express concern for the beloved's own sake?  Because Socrates treats intrinsic and instrumental concern as mutually exclusive and jointly exhaustive (Lysis 219c, 220ab), Socratic eudaimonism implies instrumental concern for the beloved.

1. Insofar as x is desired for the sake of some other thing y, then x is only instrumentally valuable (219c-220b).
2. If A and B are friends, A cares about B because of the contribution this makes to A's eudaimonia.
3. Hence, friends care about each other only instrumentally.

POLITICAL VIRTUE
Plato thinks society should be based upon a natural division of labor (NDL).  NDL is reached by consideration of a person's function (353a, 369e, 406-408, 420b, 421b, 433a, 435a, 590d).  F is the function of x iff:

1. It's best for x to F (406-8, 420b, 421b).
2. x can F better than anything else x can do (353, 435a).
3. x can F better than anyone/anything else can F (353, 369).
4. It's better for the state for x to F (406-8, 420b, 421b).

Notice that this account of political virtue does not ascribe political virtues to any particular class; the city as a whole is the bearer of these virtues.  This is required by the macro-micro analogy.  Though members of the ruling class may have psychic virtues, it is not any particular class (much less any members of a class) that have the political virtues.

THE JUSTIFICATION OF PHILOSOPHICAL RULE
Epistemological claims underlie Plato's defense of philosophical rule.

1. Knowledge of certain disputed (D-) matters (e.g. moral matters) is necessary for good ruling.
2. Knowledge of Forms is necessary for having knowledge of such matters.
3. Only philosophers have knowledge of Forms; the many are sightlovers and have no knowledge of forms.
4. Hence, only philosophers can have knowledge of D-properties; the many have only (at most) belief about D properties.
5. Hence, only philosophers can rule; the many are unfit to rule.

1. (For certain properties) attention to Forms yields knowledge; whereas, attention to sensibles yields only (at most) belief.
2. The views of the many are concerned with sensibles; the views of the philosophers are concerned with Forms.
3. Hence, (for certain properties) the many have only (at most) belief; whereas, philosophers have knowledge.

THE PHAEDO ON FORMS AND FLUX
According to Aristotle, Plato separates the forms, whereas Socrates does not (Meta 987a31-b10, 1078b12-79a4, 1086a30-b7).  Forms are ontologically separate from sensible particulars just in case the former can exist independently of the latter.  Aristotle claims that it was the search for definitions and knowledge that led Socrates to forms and that Plato's belief in sensible flux that led him to posit Forms.  The argument from flux goes something like this (74a9-c5).

1. Sensibles are in flux.
2. What is in flux is unknowable.
3. Hence, there can be no knowledge with sensibles.
4. There is (or can be) knowledge.
5. Hence, there must be nonsensible things -- Forms -- with which one can have knowledge.

1. The form of F is a property.
2. x is F iff it has this property.
3. If F = G, then all and only G-things must be F-things.
4. For any sensible property G, some G things will be F and some G things will be not-F.
5. Hence, F does not = G.
6. Hence, forms are not sensible properties.
7. Forms are necessary for knowledge.
8. Hence, no knowledge with sensibles.

Plato's main claim is that sensible properties contain both F and not-F instances.  In general, there is no reason to think that he believes that sensible particulars are both F and not-F.  Presumably, particular actions are not both just and unjust.  It's rather that in some cases returning what is owed is just, and in other cases it is not.  However, in cases involving relational properties Plato does think that sensible particulars can have contradictory properties.  Simmias is both tall (compared to Socrates) and short or not-tall (compared to Phaedo) (102b-d).  And Helen is both beautiful (compared with mortals) and ugly or not-beautiful (compared with gods) (Hippias Major 289a-e).

Later in the Phaedo Plato returns to the topic of Forms and the nature of formal explanation.

1. Plato contrasts sensible explanations of what makes all and only F-things F, which are subject to the compresence of opposites, with "safe" explanations of what makes F-things F in terms of their participation in F-ness, which are not subject to this problem.  Safe explanations are better than sensible explanations (96c5-e5, 100c3-101d3).
2. Plato mentions his desire (and inability) to find an account that explains the coming to be of all things in terms of a Final Good (97c1-98b6).
3. Plato mentions the need to appeal to formal, rather than material efficient causal, explanations (98b6-99d2).
4. In addition to safe explanations, there are informative explanations of one nonsensible property in terms of others (e.g. heat arises from fire) (96c5-105c8).

AGAINST THE SIGHTLOVERS, AGAIN
Now we are in a position to better understand Plato's second argument against the sightlovers.

1. Knowledge is set over what is, and ignorance is set over what is not  (477ab); belief is set over what is and is not (478a-e).
2. Hence knowledge entails true belief, ignorance entails false belief, and (mere) belief entails neither true nor false belief.
3. For any property F, the many base their account of F on sensible (or observable or behavioral) properties (479a, 479e).
4. For any D-property F, the sensible property (properties) that the many fix on in their account of F both is [F] and is not [F] (479ab).
5. Hence, for any D-property F, the many's account of F leads to some truths and some falsehoods (479cd).
6. Hence, for any D-property F, the many have only belief about F (479de).
7. Hence, for any D-property F, if one is to have knowledge of F, one must not base one's account of F on sensible properties.
8. Hence, for any D-property F, if one is to have knowledge of F, one must base one's account of F on nonsensible properties.
9. Forms just are nonsensible, explanatory properties.
10. Hence, for any D-property F, one has knowledge of F iff one bases one's account of F on the Form of F (479e).
11. For any D-property F, philosophers base their account of F on nonsensible (explanatory) properties.
12. Hence, for any D-property F, philosophers base their account of F on the Form of F (479e).
13. Hence, for any D-property F, philosophers have knowledge of F (479e).

'ESTI'
There is a question how to understand premise (1).  We translate the Greek verb `esti` with the English verb `to be`.  `Esti`, like `is` in English, admits of at least three different interpretations here:  existential, predicative, and veridical.  This allows three different readings of (1).

(1e) knowledge is set over what exists; ignorance over what does not exist; and belief over what exists and what does not exist.
(1p) knowledge is set over what is F; ignorance over what is not F; and belief over what is F and ?F.
(1v) knowledge is set over what is true; ignorance over what is false; and belief over what is true and false.

THE RANGE OF NONSENSIBLE FORMS
The restriction to D-properties (e.g. relational, moral, and aesthetic properties) is necessary to accommodate the "finger passage" (523-5), which says that appeal to forms is necessary only where sensory accounts lead to compresence,as they do with moral properties, but not with the property of being a finger.

THE OBJECTS OF KNOWLEDGE AND BELIEF
Premise (1) together with Plato's association of knowledge with Forms and mere belief with sensibles lead commentators to think that he is endorsing a Two Worlds theory (TW) according to which there is knowledge only of forms not sensibles, and there is belief only of sensibles.  However, TW would be an unwelcome Platonic commitment.

• TW would seem to undermine the common view that we can progress from belief to knowledge about one and the same thing by tying the original belief down with an explanatory account (Meno 97e 98a).
• TW would underwrite skepticism about the sensible world, which would make it very difficult to see why we should want philosophers to rule.
• Plato clearly contradicts TW; he claims that he and others have only beliefs about the various Forms, including the Form of the Good (505e 506a, 506c, 515e 516a, 517bc) and that philosophers do have knowledge of the sensible world (520c).

The image of the Cave is isomorphic to that of the Divided Line: C1 (the prisoners looking at shadows) corresponds to L1; C2 (the prisoners looking at models that cause shadows) to L2; C3 (ex-cons studying real objects through reflections and representations) to L3; and C4 (ex-cons studying objects themselves) to L4 (517b).  After achieving C4/L4 philosophers return to the cave (520c).

PLATO'S AUTHORITARIANISM
The guardians of the ideal state regulate the behavior of citizens so as to preserve NDL.  They are free to implement NDL by whatever means they deem necessary.  They have discretion to lie if this promotes social stability or happiness (389c, 459c).  They should censor literary and artistic works that do not adequately represent the virtuous life or promote unity (377e-, 399e, 401b).  They should install eugenics policies that will produce the best possible offspring (456e, 459d e, 460c, 536a).  They should promulgate the myth that people are born with various mixtures of the metals gold, silver, and bronze in their souls.  Predominantly gold souls will be rulers; predominantly silver souls will be auxiliaries; and predominantly bronze souls will be artisans (415a-c).  Rulers ascertain the proportion of the metals in anyone's soul.  They are politically unaccountable; the other classes cannot hold public office, and they have have no voting rights.

In assessing Plato's authoritarianism, we must remember that the political institutions of the Republic are for ideal theory.  Plato says little about non-ideal theory in the Republic.  He does not think that the ideal state is impossible (499bc, 502c), but he does think that genuine philosophers are rare (428e, 491ab, 503b) and that popular prejudice will make it difficult to realize the ideal state (488a-489a, 502c, 516e-517a; cf. 592ab).  At one point, he suggests that we might try to approximate the ideal (437ab), but we should not assume that he thinks that we should accept authoritarian rule by non-philosophers.  Indeed, in the Statesman
Plato claims that whereas monarchy is the best ideal constitution, democracy is the best non-ideal constitution (302e6-8, 303a-b).  If so, we should assess aristocratic rule for the circumstances for which it is defended, namely, ideal circumstances in which philosophers can be found and convinced to rule.

PLATO AND MILL
It is tempting to suppose that Plato simply fails to see the moral importance of individual liberty.  But it is surprising to see the extent to which Plato can accept many of the claims of liberals, such as J.S. Mill, about the importance of individual liberties.

In On Liberty Mill distinguishes paternalistic restrictions of liberty from restrictions of liberty based upon the harm principle (i 9, iv 1-4, v 2).  A's restriction of B's liberty is paternalistic if it is done for B's own sake; it is an application of the harm principle when its objective is to prevent harm to someone other than B.  At some points, Mill appears to say that liberty may be restricted iff the restriction is an application of the harm principle (e.g. i 9).

PATERNALISM
If we could justify a blanket prohibition on paternalism, we would seem to have a good argument against Plato.  Mill offers two explicit arguments against paternalism.

1. State power is liable to abuse (v 20-3).
2. Even well intentioned rulers will misidentify the good of citizens; because an agent is a more reliable judge of his own good, even well intentioned rulers will promote the good of their citizens less well than would the citizens themselves (iv 4, 12).

So far, Mill's arguments provide no principled rejection of paternalism, because they contain no case against successful paternalism (i.e. restrictions on B's liberty that do make B better-off).  But Plato has attempted to ensure that all paternalism in the ideal state will be successful.

CENSORSHIP
As books ii, iii, and x show, many of the restrictions on liberty in the ideal state take the form of censorship.  Mill mentions four reasons for ensuring free speech.

1. A censored opinion might be true (ii 1-20, 41).
2. Even if literally false, a censored opinion might contain part of the truth (ii 34-39, 42).
3. Even if wholly false, a censored opinion would prevent true opinions from becoming dogma (ii 1-2, 7, 20-33, 43).
4. As a dogma, an unchallenged opinion will lose its meaning (ii 26, 43).

(1)-(2) present no challenge to reliable censors who censor all and only false beliefs.  But Mill also believes that there would be something wrong with censorship even if the censors were perfectly reliable (ii 2).  Here is where (3)-(4) are relevant.  Mill appears to believe that freedom of expression is important because it is essential to the proper exercise of our deliberative capacities, in particular, our capacities for theoretical and practical reason.  As cognitively limited beings, we require free exchange of ideas in order to expand our menu of options and to better understand the comparative merits of the items on this menu.

LIBERTIES AND DELIBERATIVE CAPACITIES
This argument for free speech is really just a special case of Mill's more general argument for the importance of various liberties of thought and action that appeals to the role of liberties in the realization of progressive -- deliberative -- natures.  If we ground the importance of individual liberties in the exercise of deliberative capacities, what basis do we have for rejecting Plato's authoritarianism?  Plato does not recognize liberty or self-expression as a part or constituent of happiness; happiness is being ruled by one's rational part or by the reason of another if one's own rational capacities are too weak (590cd).  Because, on his view, non-philosophers are incapable of justifying their beliefs, (3) and (4) do not show freedom of speech to be part of their good.  Perhaps surprisingly, Mill accepts similar restrictions on the scope of his defense of individual liberty.

It is, perhaps, hardly necessary to say that this doctrine is meant to apply only to human beings in the maturity of their faculties.  We are not speaking of children or of young persons below the age which the law may fix as that of manhood or womanhood.  Those who are still in a state to require being taken care of by others must be protected against their own actions as well as against external injury.  For the same reason we may leave out of consideration those backward states of society in which the race itself may be considered in its nonage.  ...  Liberty, as a principle, has no application to any state of things anterior to the time when mankind have become capable of being improved by free and equal discussion [i 10].

Plato can accept this account of the importance of individual liberties.  His authoritarianism rests less on illiberal moral and political commitments than on his conception of human nature and the distribution of deliberative capacities.   But even if rational capacities are not distributed equally, artisans and auxiliaries are neither children nor mental defectives; their good consists in their developing their rational capacities and assuming as much directive control of their own souls as is possible.  Plato himself recognizes that everyone has a rational part (518c).  If so, and if, as he believes, political institutions should be designed to promote the interests of each citizen (F1), then political institutions should develop everyone's capacities for deliberative control of their lives.  But deliberative control of this sort presumably involves participation in public deliberations, a significant sphere for self-control and private deliberation, and informed and reflective assessment of alternatives in both public and private deliberation.

We can now see more clearly what is wrong with Platonic love and Plato's conception of the way in which guardians should express their concern for the welfare of their citizens.  If, as Plato believes, my own good consists in my exercising directive control over my appetites and passions, then another cannot express proper concern for me for my own sake by trying to improve me in ways in which my own deliberations play no essential role.  Concern for another must attempt to improve an agent's deliberative capacities in ways that engage those capacities.