Thumos and doxa as intermediates in the Republic

Broadly speaking, something can be called intermediate for Plato insofar as it occupies a place between two objects, poles, places, time, or principles. But this broad meaning of the intermediate has been eclipsed by the Aristotelian critique of the intermediate objects of the dianoia, so that it has become more difficult to think of the intermediates as functions of the soul. The aim of this paper is to show how, in the Republic, thumos is analogously treated as an intermediate with other kinds of intermediate objects, and tentatively to relate this psychological intermediate in a broader theory with doxa, as its epistemological ground in the course of action.


INTRODUCTION
When it comes to "intermediates" in Plato, one is tempted to think solely of mathematical intermediates, the objects of dianoia in book 6 of the Republic. Whether or not it comes from Aristotle's critique on such intermediates, one must admit that he himself forgets to describe as intermediates some of the most important aspects of Plato's psychology and ethical the ory 1 . But the word "metaxu" has undoubtedly a broader meaning in Plato's dialogues. As Joseph Souilhé already noted in his thesis in 1919 2 , Plato can be called a "philosopher of the intermediates" insofar as the aim of the whole of his philosophy is to bridge the gaps between what is taken to be two poles or two kinds of reality. Souilhé's first aim was to classify the wide range of intermediates into categories: "psychological" (thumos, erōs, doxa, dianoia), "ethical" (sophrōsynē, dikaoisunē, bios meson), "political", "cosmological", and "metaphysical"; a second consideration was to question whether there was a more systematic link between those intermediates. This paper addresses the following ques tion, which is crucial for the meaning we are to give to Plato's moral psychology in the Republic: is there a link between thumos as intermediate and its epistemological counterpart, doxa? Let us recall briefly what thumos and doxa stand for. In book 4 of the Republic (436b 441c), So crates argues that the soul is composed of three so called "parts" or rather "functions": the rea soning part (to logistikon), the desiring part (to epithumetikon), and an intermediary part (to thumoeides or thumos), which is often trans lated as "spirit" 3 . This intermediary function is presented as having a key role -in the best case -to mediate reason's commands, against the power of the desiring part. For whatever reason Plato shifts from a bipartition of the soul (with reason and desire) to a tripartition, thumos introduces a new way of thinking of the relation between reason and desire: thumos is immediately thought to be a metaxu. As for doxa, which we take to mean opinion, but also belief, this is certainly an ubiquitous concept in Plato's dialogues which is not tied with any systematic presentation; nevertheless, as it will be shown, doxa is presented in the Republic as a metaxu too, between knowledge and ignorance, having its object somewhere between what is and what is not.
The aim of this paper is certainly not to look for a system of intermediates. It will thus not be argued that thumos is the "seat" of doxa. As Sylvain Delcominette already showed convinc ingly, one should not conflate what appears to be a theory of the "parts" or "functions" of the soul, and what we could call a theory of "facul ties" or "capacities" 4 . Indeed, insofar as doxa is concerned, it is quite clear from the Republic that doxa can be at least ascribed to different "structures" or characters or, broadly speaking, to the "agent"; it seems conversely impossible to ascribe the faculty of doxa to a specific "part" of the soul, and even less so to ascribe different "kinds" of doxa to different parts 5 . Neverthe less, the issue persists: if we are to accept the idea that there is an ethical function of thumos in the tripartite soul, which is manifest for the auxiliaries, for example in the form of what has been called an "imperfect virtue" 6 , one has to look for the epistemological grounds of such virtues or dispositions.
The question I want to raise is slightly dif ferent from the ones which try to ascribe sys tematically doxa (or whatever function) as an intermediate faculty to thumos as an intermedi ate part; my question would rather be: is there a reason why we would ascribe doxa to thumos because these two are both intermediates? In what follows, I will try to show that there is a homogenous theory of the functionings of the intermediates, that leads us to ascribe in a privileged way doxa to thumos in certain ethi cal situations.

THUMOS AS INTERMEDIATE
The argument for positing thumos as in termediate is found in book 4 of the Republic (439e 441c), it is not the place here to recall the precise argument that leads to the discovery of the tripartite soul 7 , but it is interesting to note that the "intermediate" dimension of thumos can be understood in a polysemic way. 1) Meaning 1: Thumos is found out by contrasting its function first with desire (439e6 440e6), then with rea son (441a5 c2); it is neither desire nor reason even if some of its features seem identical. Thumos is then first described as a kind of "interval" cov ering a variety of ambivalent actions and passions: being angry or ashamed, resisting desires or fighting for some values, etc., all of them being best described as in between reason and desire. 2) Meaning 2: Thumos is nevertheless a "median position" between the two extremes regarding virtue; if thumos is first thought as an interval, it rep resents at the end of the argument an autonomous function (eidos, genē) of the soul in between the two other poles, the range of actions and pas sions being unified by a single class term (439e4; 440e8; 441c6). In this respect, the thumoeidic person, as it is clearly showed by the examples of Leontius, the honest man (who does not seem to be a "wise man" though), children, animals and Ulysses, are not paragons of virtue, but they may nev ertheless embody an honest behavior without being completely virtuous. 3) Meaning 3: According to Socrates, thumos helps reason to fight desires whenever it is possible and provided that it is well educated; thumos is an auxiliary (epikouros) for reason (441a2 3). A third meaning of "in termediate" emerges here, insofar as thumos is not only an interval and a median position, but also tran scribes reason's recommendation in the whole agent. In other words, thumos "mediates" reason's rule in a positive way 8 .
The polysemy of "intermediate" in our passage may explain how difficult it is to as cribe a clear cut theory of the cognitive power of thumos. Examples of conflicting situations (thirst, and then the example of Leontius) show that a complex epistemic process is go ing on in the agent, which relies on different understandings, depending on the function of the soul that leads the course of the ac tion 9 . The action is morally distinct whether we rely on a) what is pleasant and painful, b) or on what is worthy or valued by others, c) or on what is reasonable and/or rational. There is a supplementary difference, which has been notoriously described through the distinction between good independent and good dependent principles 10 , whether we act out of mere compulsion, or out of knowledge, or out of a doxa which is potentially right or wrong.
If we take the example of Leontius (439e6 440a6), his desire to see the corpses refers to a cognitive understanding that confers pleasure to this kind of morbid desire, whereas his anger against his same desire relies on a internalized judgement according to which taking pleasure in the misfortune of others is morally bad. By contrast, Ulysses (441b2 c2) may well be driven by his revengeful anger to kill the suitors out of a judgement that condemns such a vile behav ior, but he forms a rational (yet not necessarily morally just) judgement that prevents him to do so right away, probably to make his revenge more efficient.
The question is: in the course of action, what kind of activity does the agent enact, and by which part of the soul? A straightforward answer is that the desiring part desires, needs, craves, pushes and pulls, that the reasoning part reasons, learns, contemplates, and finally that the thumoeides affects a state of mind in the agent that is anger, shame, and other emotions that are precisely intermediate in being neither a desire nor a reasoning (meaning 1). How can we characterize thumos's function in this tripartite model? Following Angela Hobbs's analysis 11 , we can say that the thumos "values", that is: gives personal importance to principles or objects, and leads the agent to commit himself in what he finds good, beautiful and just because that's what he values most. Anger and shame are thus intermediate behaviors that exemplify the in termediate position of the character regarding virtue (meaning 2). As a motivational principle, thumos has then a role to play in each action we make, regardless of whether we are philosopher or not, depraved or not, philotimos or not. For in the course of action, there is a desire, and ei ther knowledge proper, or doxa, right or wrong; and in the last case, thumos gives the content of the doxa a value that commits the agent in his action, all the more so if reason pervades or produces this doxa. What is at stake here is the way thumos as a part of the soul might be able to grasp something as a form or appearance of the good, insofar as moral judgments bear upon something which is potentially related to a kind of knowledge (meaning 3). It is all the more important for our topic, for if a kind of virtue -an imperfect one 12 -is related to thumos, notably for the auxiliaries, it has to do with their ability to acquire some intermedi ary disposition between knowledge and mere obedience and compulsion.
Taking the same previous examples of Le ontius and Ulysses, one thing is to say that there is, in every situation, a judgement that relies on grounds that can be pleasure and pain, values and reason, and another thing to ascribe to each function of the soul a definite cognitive power. Leontius and Ulysses have, to say the least, a conflicting behavior; one way to understand this conflict is to posit opposite judgements on what is actually good and bad -each judgement coming from a general cogni tive apprehension of the situation depending on different criteria. It is not necessary for our present purpose to claim that desire and thumos have their own cognitive capacity; let us just say, in a more economical manner, that the rational part reasons; thumos listens to reason and acts according to a doxa that comes from reason giving it some value; and the desiring part desires, but might infect the doxa with its own criterion of appraisal, that is pleasure and pain. Thumos is intermediary because its function is to give value to a doxa, wherever it may come from, committing the agent into this system of values.

THUMOS AS PSYCHOLOGICAL SUPPORT TO DOXA
How can we explain the relative privileged link between thumos and doxa in the Republic? Instead of focusing on a putative theory be tween faculties and parts of the soul, we may try to follow another path to link the two func tions, in positing an analogy between thumos and doxa as intermediates. For as a matter of fact, the way Socrates describes doxa as an in termediate in book 5 can be well compared with the way thumos is discovered in book 4. Again, three meanings could be found of metaxu when applied to doxa.
1) Meaning 1. Given that there is a wide range of objects and discourses be tween authentic knowledge, and ig norance, there must be an "interval" between these two poles (477a9 b1). Doxa is the name given to what is "in between" being first contrasted with knowledge (477e8 478a1), and then with ignorance (478b6 c5). Its object being between "what is" and "what is not", doxa refers to this inter val, which we know to be very wide: from an ordinary perceptive opin ion on what is beautiful to a strong judgement on what is just and good, the reign of the opinion is potentially infinite. This first meaning of doxa relies, so to say, on its extension. So crates ends his argument by insisting on the operation of considering doxa precisely as an interval defined by these two poles, even if it is in a nega tive way, in order to mark the limits of this capacity. As for thumos, "setting the intermedi ate over the intermediate" is already giving doxa its place, and preventing it from overflowing reason's function. 2) Meaning 2. Doxa has power; where does it comes from? Doxa is not only a vague interval but also a "position" between knowledge and ignorance. As a judgement, a belief, or even as a per ceptual image, doxa gives the illusion to maintain something steady, even if plural and wrong. This is the case of the "lovers of sights" and "sounds" who claim to be experts in beauty (475d1 e1; 479d3 e5). Because those who do have a doxa act and speak as if they possess a real knowledge -and precisely because they do not abstain or claim their ignorance, doxa is a metaxu between knowledge and igno rance insofar as they assert something which, even if false, pretends to be real and true. It is then not sufficient to demarcate doxa in its extension, in be tween what is and what is not; Socrates has to define doxa as a metaxu in a hierarchical way, as a median position between what is truly known and what is simply ignored. In giving its right place as median position, Socrates makes doxa a class of judgement in regard to true knowledge, accounting for its inconsistency and nevertheless its psychological power. 3) Meaning 3. But why does Socrates then admit that doxa is a capacity, rather than a non capacity, as ignorance is? 13 Last, doxa is described as a possible "mediation" through the two other poles (knowledge as a capacity, and ignorance as a non capacity), insofar as people would accept, in the best case scenario, that there is a differ ence between the philosopher and the philodoxos. We know, from the Meno (97b9), that doxa is as efficient in the action, if true, as knowledge. Here in the Republic, Socrates makes a further step. In forming a true opinion, in be ing persuaded by the philosopher that there is indeed a difference between knowledge and opinion (476d8 e3), one might expect, at best, that one can hold a doxa knowing that it is a doxa and not knowledge. This is not to say that doxa could, if true, be as valuable as knowledge; but in succeeding the refutation and persuasion, Socrates could make doxa a (non rational) mean to assert the superiority of rea son 14 . I will come to this point in my third section.
This analogy between the functioning of both intermediates, doxa and thumos, does not necessarily entail that there is a privileged link between ethical and epistemological intermedi ates. And there is no hint in the description of doxa which is explicitly said about its ethical counterpart, thumos. Nevertheless, it is inter esting to show how these two intermediates in the Republic are associated to give a full ac count on what it is to have an opinion, a belief, a representation of a value, as the experience of the agent 15 .

The doxastic object of thumos
First of all, thumos seems to have a privi leged range of objects, all of them reducible to timē (honor and esteem) and nikē (victory) ac cording to book 9 (581a9 10). The philotimos (lover of honor) behaves according to doxai that refer to these two objects. Now, these two terms could apply to many other objects, persons or actions, insofar as they contribute to acquire some timē or nikē; for example, public honors or presents are thought to be necessary to acquire more timē (social honor) in general. It goes the same way with victory, beauty, courage and manliness, love for action rather than love for discourse and knowledge, power, love for gymnastics rather than love for music, etc. All these objects are valued by the philotimos with the view to acquire more timē. In theory, one can "value" anything, so as to become a privileged object for his thumos, but the philotimos selects what he values for the sake of timē. To put it in a nutshell, the kind of action attributed to thumos in book 4 (to esteem and to value) is generally (though not systematically) equivalently understood as a special kind of desire: "to love timē" as an object 16 . Timē and nikē, which refer to relative status are best described as doxai, thought as reputation and all the kinds of judgements that refer to this very reputation: what people say, praise and blame, rumor, and further mode of appearances such as glory, shame, etc.
We are now in a position to have a better understanding of this meaning of intermediate as "interval", both for doxa and thumos. There is an intimate connection between those two intermediates not so much because of a so called cognitive ability of thumos, but because both the function of the soul and the capacity pervades a very wide range of objects, some of which are not easy to refer to desire or reason only, as book 9 recalls: Won't a money maker say that the pleas ure of being honored (τὴν τοῦ τιμᾶσθαι ἡδονὴν) and that of learning are worth less compared to that of making a profit, if he gets no money from them? -He will. -What about an honor lover (ὁ φιλότιμος)? Doesn't he think that the pleasure of making money is vulgar (φορτικήν τινα ἡγεῖται) and that the pleasure of learning -except insofar as it brings him honor (μὴ μάθημα τιμὴν φέρει) -is smoke and nonsense? -He does. -And as for a philosopher, what do you suppose he thinks the other pleasures are worth compared to that of knowing where the truth lies and always being in some such pleasant condition while learning? Won't he think that they are far behind? (Resp. IX, 581c10 e3, trans. Grube, rev. C.D.C. Reeve).
Objects of thumos are always doxai in the sense that they are social and political constructions of what people value most in a given city. No wonder then, that the objects of thumos are potentially instable, inconsistent, and rest all the more so on sensible particulars and situations.

Thumos gives power to doxa to overcome desires
A second important aspect of the analogous functioning between thumos and doxa is the way the first gives strength to the latter, and especially over pleasure and pain.
But what happens if, instead, he believes (ἡγῆται) that someone has been unjust to him? Isn't the spirit within him boiling and angry, fighting for what he believes to be just (συμμαχεῖ τῷ δοκοῦντι δικαίῳ)? Won't it endure hunger, cold, and the like and keep on till it is victorious, not ceas ing from noble actions until it either wins, dies, or calms down, called to heel by the reason within him, like a dog by a shep herd? (Resp. IV, 440c7 d3, trans. Grube, rev. C.D.C. Reeve).
The honest man holds a doxa on what is just and unjust. It is not said how the agent (in our case an honest person, but not neces sarily a "virtuous" one) forms its belief on justice, but it appears that this belief gains its force through his spirited part, through bodily symptoms and anger. Thumos is not the function through which a doxa is formed, but it is, for sure, that through which it gains its force and value in the course of action. As it has been often pointed out, there are many desires that are supported by a doxa, especially in the case of the characters in book 8 and 9, for example the oligarch 17 . It may even be the case that an acratic person is best understood as an agent whose doxai follow opposite di rections 18 . So again thumos is certainly not the only function in the soul that deals with doxa; rather, thumos is an auxiliary power ful enough to overcome natural pleasure and pain ("hunger, cold, and the like") or even life ("either wins, dies, or calms down"), in giving doxa a sufficient value against com peting desires that would follow a pleasure/ pain criterion.

Thumos gives doxa a relative stability
A third feature of the thumos/doxa relation is made explicit in book 4 through the descrip tion of civic courage of the auxiliaries. Even if doxa is volatile, not being grounded on reason, thumos is capable to transform a doxa into a quasi permanent disposition.
Then, you should understand that, as far as we could, we were doing some thing similar when we selected our sol diers and educated them in music and physical training. What we were con triving was nothing other than this: that because they had the proper nature and upbringing, they would absorb the laws in the finest possible way (ὅτι κάλλιστα τοὺς νόμους πεισθέντες δέξοιντο), just like a dye (ὥσπερ βαφήν), so that their belief (δόξα) about what they should fear and all the rest would become so fast (δευσοποιὸς) that even such ex tremely effective detergents as pleasure, pain, fear, and desire wouldn't wash it out -and pleasure is much more po tent than any powder, washing soda, or soap. This power to preserve (τὴν δὴ τοιαύτην δύναμιν καὶ σωτηρίαν) through everything the correct and law inculcated belief (δόξης ὀρθῆς τε καὶ νομίμου) about what is to be feared and what isn't is what I call courage, unless, of course, you say otherwise. (Resp. IV, 429e7 430e5. trans. Grube, rev. C.D.C. Reeve) In this passage, the origin of doxa is made clear enough: coming from law and reason, a series of beliefs are internalized by the auxiliaries through different means (music, gymnastic, and other kinds of training that have been depicted especially in book 3). Be cause the origin of doxa is reason and law, it is a just and correct one (orthē). But it is not because it is a right opinion or belief that it lasts in the face of pleasure, pains and other passions: thumos, which is known to be one of the tendencies that has been the attention of the educator in the prior education, has the power (dunamis) to preserve (sōtēria) the opinion against other desires. A difference then should be made between having an opinion, believing it is true and assenting to it on the one hand, and having an opinion that constitutes one's character on the other hand. Of course, this is an "imperfect" virtue which is described here, insofar as counterfactual situations may well destroy the power of thumos; but it remains true that only the power of thumos conveys the doxa to be steady in spite of its ontological and epistemological instability.
What is important then is not only the fact that the doxa is right or wrong, but also the way thumos (and the whole agent) considers it as a dynamic intermediate to perform a good or (imperfect) virtuous action. If we want to account for the epistemological processes of thumos, we should not properly say that it has or forms a doxa, but rather that it gives doxa some of the properties to become not only a judgement, either propositional, or perceptual or both, but a real valuable belief.

INTERMEDIATES IN THE EDUCATION
It is now possible to account for the impor tance of the notion of "intermediate" during education in the Republic, in giving doxa and thumos the role of mediation towards the posi tive pole from which they are defined as an in between position.
At the end of book 4, Socrates concludes, with the help of a metaphor, on what it is to harmonize our own soul in giving a last defi nition of "justice", after he gave definitions of the three other cardinal virtues.
And in truth justice is, it seems, some thing of this sort. However, it isn't con cerned with someone's doing his own externally, but with what is inside him, with what is truly himself and his own. One who is just does not allow any part of himself to do the work of another part or allow the various classes within him to meddle with each other. He regulates well what is really his own and rules him self. He puts himself in order, is his own friend, and harmonizes the three parts of himself (συναρμόσαντα τρία ὄντα) like three limiting notes in a musical scale (ὥσπερ ὅρους τρεῖς ἁρμονίας) -high, low, and middle (νεάτης τε καὶ ὑπάτης καὶ μέσης). He binds together those parts and any others there may be in between (εἰ ἄλλα ἄττα μεταξὺ τυγχάνει ὄντα), and from having been many things he becomes entirely one, moderate and harmonious. Only then does he act. And when he does anything, whether acquiring wealth, taking care of his body, engaging in politics, or in private contracts -in all of these, he believes (ἡγούμενον) that the action is just and fine that preserves this inner harmony and helps achieve it (ἣ ἂν ταύτην τὴν ἕξιν σῴζῃ τε καὶ συναπεργάζηται), and calls it so (ὀνομάζοντα), and regards as wisdom the knowledge that oversees such actions (σοφίαν δὲ τὴν ἐπιστατοῦσαν ταύτῃ τῇ πράξει ἐπιστήμην). And he believes that the action that destroys this harmony is unjust, and calls it so, and regards the be lief that oversees it as ignorance (ἀμαθίαν δὲ τὴν ταύτῃ αὖ ἐπιστατοῦσαν δόξαν). (Resp. 443c9 444a2; trans. Grube, rev. C.D.C. Reeve).
Socrates has already used the musical metaphor to convey an image of a harmonized soul in book 2 and 3, notably in 410c 412a, where the aim of the first education by music and gymnastics was to find a balance in the soul of the future guardian between her/his thumoeides and her/his philosophical nature, in "tuning" them. It is by this tuning that one could achieve a musical "chord", through the equilibrium between these dispositions (412a4 7) 19 . In book 4, the chord depends on the knowledge and sophia that comes from the law, in tuning the three strings of the harmony which correspond to each function of the soul. The harmonia is then not only a tuning between dispositions, but a hier archical ordering of the soul's three parts, so that reason should rule over the others (epistatousa), thumos should "preserve" (sōzē) reason's rule -as we have seen through the dying metaphor (429e7 430e5), and the de siring part should obey this disposition. So that from book 3 to book 4, thumos is not anymore a natural tendency in the soul, but should become the equivalent of the mesē in the musical instrument, that is the position through which the interval between reason and desire is made definite and virtuous.
One may wonder about the oddity of the sentence: "and any others there may be in be tween", as if there were other intermediates than thumos. Plato probably refers here to a tetrachord, and mentions only the principal fixed strings (horoi) -the hypate and nete be ing the lower and the highest string, the mese and paramese being in between -, the movable other strings depending on the type of harmonia wanted 20 . It is unlikely that Plato means that there are other "intermediates" between rea son and desire than thumos; rather, we should understand that, given the fixed position of these three strings, some variations may occur between the just persons, whether they belong to the ruling class, the auxiliary class, or the third class of the city, and whether their natural disposition is more akin to one of the natural tendencies that have been described in book 3. In other words, thumos as a position in between reason and desire in book 4 overlaps the "inter val" of actions and dispositions that it covers in book 3. Then, we must recognize that there is a certain "plasticity" of the thumos that can be molded, shaped and modeled during educa tion, in order, at least, to have an "imperfect" virtuous person.
The two last sentences are quite relevant as to the link between thumos and doxa as mediation.
Knowledge (epistēmē) is presented as the sole ground for virtue and sophia, whereas it is a doxa that is assimilated to disharmony, but also to ignorance (amathia). This strong dichotomy between knowledge and doxa does not seem to be coherent with what is said of doxa in book 5 where ignorance is distinct from it from an ontological point of view. A solution to this apparent paradox would be to refer to a distinction between a right and a wrong doxa, the latter being responsible of ignorance and vice. However, it is not a right doxa which is responsible for virtue either, but proper knowledge. So I would suggest, rather, that this description of this harmonious per son is not of a "real" virtuous man who would have the knowledge of it, but a mere ordinary man, who already has a doxa on what is the principle of the harmony or disharmony in his soul. Grube translates the "ἡγούμενον" as "he believes", as if it were another doxa whose ob ject itself is the difference between knowledge and doxa. What is maybe an over translation is getting to the point: what the honest man has is a "thinking", an ethical judgement, which, depending on the education of his thumos, val ues reason and law as the rule of his action. A similar situation occurs in book 5, when the philosopher finally persuades the other citizens that there is a difference between doxa and knowledge, even if the citizens do not have access to proper knowledge. We must then rec ognize that thumos's function here is to value reason and law as the proper origin of right doxa; this explains why the honest man finally "names" sophia the rules and recommenda tions that thumos is inclined to follow. Here, thumos has a crucial role to play in recogniz ing, through a right doxa, what falls within reason or the law's rule, and what falls within mere unjustified doxa. The median position of the intermediate is not enough to ascribe to one of the opposites a positive value; another function of the intermediate is to be a step for ward to the positive pole. In other words, the intermediate gives a meaning to the poles in being a mediation between them and positing them as extremes as it does, and valuing rea son, good, fine, noble, as positive poles rather than desire, pleasure and pain.

CONCLUSION
All these striking links and similarities be tween thumos and doxa should not lead us to posit that thumos is a seat of doxa, nor that, I contend, thumos is an epistemological faculty similar to doxa as far as ethical judgments are concerned. Plato never says that explicitly and has probably no reason to do so. We cannot go further then in positing a system of inter mediate faculties. However, thumos is said to be sensitive to reason in a way that compels us to make it an essential psychic function to give doxa a practical meaning. If there is no theory between ethical and epistemological intermediates, there are, indeed, analogous operational relations between them. Thumos and doxa are polarized intervals, but also po sitions between real and pre existent valued poles (what is valued as good), and finally dy namic starting points to access the positive pole (reason's rule).