David Schur, Plato's Wayward Path: Literary Form and the Republic
Part I. Literary Form and Classical Rhetoric. 1. The Problem of Literary Form
2. Philosophical Rhetoric
3. Literary Practice, Modality, and Distance
Part II. Concerning the Republic. 4. From Beginning to End and Back Again
5. Digressing toward a Possible Regime
6. Imagining Images in Chains
Glossary of Key Greek Words
3. Literary Practice, Modality, and Distance
The previous two chapters discussed hermeneutic problems and rarely touched on solutions. Positive goals conventionally associated with literary-rhetorical methods of reading were mentioned, but they remain to be explained more fully in the context of interpreting Plato’s dialogues. As I will continue to argue, a powerful teleological impulse directs readers to seek single-minded and persuasive arguments in works that are, nonetheless, manifestly not univocal. The drive to explain literary form is one of the main factors that compel readers to reconstruct a covert univocality from Plato’s dialogues. This chapter begins with some positive methodological suggestions that are eventually directed toward Plato’s Republic. After reviewing some hermeneutic concepts that distinguish the modern rhetorical analysis of tropes—concepts such as relative emphasis, overdetermination, and modality—I introduce some specific tropes of distancing that emphasize and express remoteness throughout the Republic. Literary interpretation generally treats texts as highly mediated forms of communication, allowing that authors are more or less remote from what their texts say. And this same kind of remoteness is, I argue, given considerable weight in the Republic, both as a linguistic phenomenon and as a methodological topic.
The general problem of the dialogue form’s digressive and indirect presentation of ideas is compounded by perplexing “ambiguities, obscurities, gaps, and contradictions” that pervade the characters’ arguments.  These are so prevalent in the dialogues, observes E. N. Tigerstedt, that Plato may seem to have introduced them intentionally, as though his aim were not to satisfy “our legitimate claim for clarity and coherence but rather to derive a malicious pleasure from eluding it.”  This is a strange thought, worth considering further: while making tantalizing gestures toward clarity and coherence, Plato may actually be methodically elusive. The strangeness can be erased by insisting that Plato’s elusiveness is always and only an apparent deviation from methodical communication—but genuinely methodical elusiveness is a surprising phenomenon. Although Plato’s writings are usually readily comprehensible, says Tigerstedt, “it is often difficult to be sure of what he really means” (14). I understand this basic difficulty in the interpretation of Plato as a constant struggle against the waywardness, the unpredictable and unexpected deviance, of his written texts. In other words, each elusive dialogue is pointing in multiple different and at times irreconcilable directions, while the interpreter typically seeks to find a path that leads from Plato the author to a univocal, promotional message—a philosophy of some sort—that supposedly must have inhered in the text from the start. Yet to carve such a straight path out of Plato’s elusiveness is to disregard integral aspects of the text.
If we are to avoid assuming that the dialogues are covertly univocal tracts—which is entirely different from assuming, as I do, that each text is a fixed communication composed in words by a single author—then it will be best to treat each dialogue as a separate text. In the current chapter, I look to Plato’s Republic in an effort to account for one dialogue’s literary features, without trying to circumvent them or otherwise explain them away. I choose to focus on the Republic for several reasons: its current popularity and influence make it an exemplary object of study, its length allows broad patterns as well as details to emerge as demonstrably significant, and its self-conscious procedural discourse (seen in characters’ frequent references to their own conversation) highlights waywardness as a thematic concern. This self-consciousness, or reflexivity, is itself a salient example of the now modern-seeming literariness that was already present in Plato’s writing.
A Literary Practice of Interpretation
If we are to address the literary form of the Republic, and if we are to do so in a manner that accepts the work’s literariness, then a fuller account of some literary-rhetorical working principles is needed, regardless of how different each individual interpreter’s practice will be. While scholars working on literary form in Plato have been faulted for having less success with interpretive close readings than with metaphilosophical speculation, here I have been raising concerns about a more basic and metatextual stumbling block: namely, a fundamental difference between expository and literary conceptions of how texts function.  (Guilty of dwelling extensively on a metatextual question, I myself can only protest that the question has blocked my way and consumed my interest as a reader of Plato’s text.)
Those on one side of this methodological divide maintain a traditional view of rhetoric, where many factors (including the historicizing goals long associated with philology and philosophy) foster a linear, teleological focus on literary elements as building blocks of argumentation and rhetorical persuasion. This, I have contended, is a serious hermeneutic limitation with regard to literary form. Modern rhetorical analysis, on the other side, brings with it a degree of freedom in this respect, by not requiring interpreters to derive a single persuasive agenda (a programmatic ideology or moral) from a writer’s verbal maneuvers. Simply put, for a literary interpretation broadly conceived, all texts have many functions; perform many acts, both pragmatic and aesthetic; and advance many ideologies simultaneously.  Which raises a consequential concern: how are we to make discerning choices? Such a question can hardly receive a full treatment here. For clarifying my approach to the Republic, however, the modern interpretation of verbal rhetoric is perhaps best understood as a matter of emphasis. The following rationale for describing and interpreting language owes much to linguistics, but it goes far beyond that discipline’s boundaries. Instead, it undergirds a method of humanistic cultural self-understanding. 
For the sake of clarity, the kind of modern and literary rhetorical analysis I have in mind can be defined as the description of tropes and the interpretation of their functions. As in ancient rhetoric, the tropes described by modern rhetorical analysis are identifiable uses of language that deviate (turn away) from standards of literal, prosaic expression. They exceed and enrich the norm of literal, denotative expression. But if the effects created by the tactics of ancient oratory are measured by the extent to which they support a speaker’s ultimate, single-minded, pragmatic strategy of persuasion, the modern interpretation of tropes allows for their participation in multiple, simultaneous patterns of written language, so that currents and countercurrents of thought emerge through themes-as-expressed-in-language.
In practice, two basic policies may serve, when combined, as a schematic antidote to expository approaches that limit Plato’s rhetoric to a univocal purpose. First, a literary hermeneutic will recognize patterns essentially by allocating attention and interest, which can be quite different from trying to trace a linear argument through a text. Formal-semantic patterns are construed in different ways by different readers, but each reading or interpretation is anchored in specific textual evidence. Careful analysis of a text requires attending to the interplay of language occasioned by specific wording. In terms devised by Roman Jakobson and associated with Russian formalism, wording consists of the text’s syntagmatic arrangement and its paradigmatic selection of diction, which are interrelated in complex patterns.  When observing and describing these patterns—accounting for local details and global structures in complex systems of wording—readers allocate more attention to some details and less to others, depending on their perception of emphasis in the text. For this reason, guided by emphasis and interest, reading becomes a dynamic balance of attention and inattention, rather than a search for hidden meanings and covert authorial agendas. This aspect of critical practice is well served by the formalist principle of foregrounding, which identifies significance by noting textural patterns of similarity (parallelism, repetition) and contrast (deviation). 
The second policy, which concerns polysemy, provides an inseparable semantic complement to the formalist description of emphasis and patterning. Rather than suppress plurisignification for the sake of univocality, an operative literary approach to texture will give ample consideration to multiple, concurrent meanings when deciding what is important. In the examination of details and when drawing together broad thematic trends, emphasis (and therefore significance) attaches to elements that exhibit redundancy, overlap, and semantic plenitude. Here we have the principle of overdetermination, which accords greater relative significance to elements that participate in multiple patterns of meaning simultaneously. Closely related to overdetermination is the idea of mutual reinforcement.  Constituents of all types may share mutual connections, each exerting more or less gravity, so to speak, depending on whether it is more or less polyvalent and on its proximity to others. Each one can gain weight from multiple directions and thereby become an increasingly relevant part of the patterns to which it belongs. To indulge in a metaphor of my own: a polyvalent detail with a dance card full of other, related elements is a more meaningful one; a well-attended dance is a more inviting one; and each text is hosting multiple dances.
Within the parameters just described, literary emphasis—not bound to the reconstruction of an ultimate, single argument—is always construed as relative; and it is mainly in this sense that interpretation may become holistic and contextual.  Furthermore, literary interpretation is obliged to be pluralistic: differentiations in texture display concurrent patterns of emphasis, and no text is an undifferentiated block of argument. In academic practice, it is de rigueur to present one’s literary reading as though it were the correct one; in practice, though, success is achieved by explaining how the evidence functions in interesting, important, significant ways.  Interpreters of Plato’s writing who do not acknowledge the value of relative emphasis in form, and the validity of pluralism in meaning, will continue to convert the dialogues into expository treatises. 
If not quite the enemy, overdetermination is at least the antithesis of univocality in rational argument and in figurative representation. The linear focus needed to pursue logical argument through the variables introduced by literary form must be narrow, and must remain insensitive to the potential for words to interact in unexpected and unpredictable ways. What Helen Vendler calls the “unifying play of mind and language” is inevitably the variable interplay between words, their meanings, and the constellations that emerge into the foreground between text and reader (1997:5). Similarly, as long as each element, combination, and text (like a word, a phrase, and a sentence) participates in different constellations, then forceful attempts to regularize the text’s combinatorial universe will result in nearsighted underinterpretation. Thus underinterpretation puts just as much of a strain on Plato’s language as does overinterpretation. And allegorizing treatments of figurative expression, by assigning a fixed group of meanings to a fixed grouping of figures, ultimately transform the figurative potential of manifestly nonliteral features into covert, univocal, and literal messages.
Even from this abstract overview, it should be fairly clear that I am describing literary practice as a dynamic process. While navigating through strikingly different types of evidence, one’s view of some evidence changes in view of other evidence. But it may not yet be clear where such an interpretation might lead, given my criticisms of the telos that guides expository approaches to the literary Plato. The task here is at any one time to reach a perspective on a theme as articulated in language by Plato, as opposed to a thesis that Plato supposedly held but left undeclared. In this regard, it is crucial to define some terms associated with theme. I take theme itself to denote an abstract topic that permeates a text, whether explicitly or implicitly. A theme is a topic of interest, a preoccupation that can emerge in different contexts and in different forms throughout a work—and which need not be addressed by the text in direct, linear discussion. So justice is a prominent theme in the Republic. A thesis is a proposition asserted as true by an author; this is the sort of claim typically advanced in expository writing. One thesis of my book is that the rhetorical patterning of Plato’s Republic expresses, among many other things, a critical or questioning perspective on the conventional methods of investigation pursued by Socrates and his interlocutors.
A thematic statement, in contrast with a thesis, communicates a proposition that is not asserted by the author.  Such statements are propositions made in the text but not endorsed by the author. A statement about an important theme will naturally attract special attention. (This sort of unasserted statement is common in traditional forms of literature, being made by characters and narrators.) When, in Plato’s Republic, the character Socrates is quoted (by Socrates the narrator) as saying “virtue has one form [eidos], but evil has countless forms” (ἓν μὲν εἶναι εἶδος τῆς ἀρετῆς, ἄπειρα δὲ τῆς κακίας, 445c), we have a thematic statement; one among countless others that may be found in the text.  In order to treat this as a thesis being asserted by Plato, interpreters must perform extra acts of literalization to establish Socrates as an allegorical equivalent of Plato—even when this equivalence will not consistently support the interpreter’s reconstruction of the argument. On the one hand, Socrates, according to this procedure, is not really Socrates; on the other hand, what Socrates says as a figure is what Plato is really saying. According to the metaphorical metalanguage of literary criticism, this is an attempt to straighten out or regularize the twists of figurative language. Taking a literary speaker’s thematic statement for an authorial thesis is typically considered a mistake in the literary practice of interpretation, but too much handwringing over this sort of intentional fallacy could obscure the prior necessity of making a decision between literary and nonliterary approaches.
Modality and the Qualification of Discourse
Born into a world shaped by Platonism, most readers of Plato will find it hard even to entertain the possibility that he may not have written simply and solely to persuade readers to adopt his own practices or beliefs. Despite the popular reverence for Socrates’ professions of ignorance, he is generally viewed as a know-it-all, a character in the dialogues whose wisdom is exceeded only by that of his author Plato. Thus, as already described, we take Platonism to be a body of beliefs, constituting a philosophy that itself largely makes up philosophy as we know it. Equally relevant is the fact that many readers regularly and habitually aim to extract a moral (a main thesis or proposition endorsed and promoted by the author) not just from expository texts but also from decidedly literary ones as well. Scholars struggle over the moral of Antigone or Hamlet no less than they fret over the message of the Republic.  Effectively blinkered, readers may resolutely pursue such expository and literalistic goals, gathering and assessing information while shutting off multiple avenues of ongoing exploration. I am not arguing that readers cannot or should not read this way. I have, however, sought to explain why such a thesis-seeking approach must remain blind to the kind of textural complexity that is valued in the modern practice of literary criticism.
Intertwined with the conviction that Plato is trying to teach us lessons in a rhetorically persuasive manner is the assumption that he is endorsing and promoting the propositions (theses, points, and claims) expressed by his characters, especially the ideas introduced and advocated by his protagonist Socrates.  When an argument is brought to some kind of conclusion during a dialogue, this conclusion is viewed as a goal reached by Plato’s composition. As a matter of course, many commentators do not hesitate to use tags such as “Plato thinks,” “Plato believes,” and “Plato says” to ascribe authorial arguments and assertions to the dialogues. Yet when Nicholas White, for instance, declares that “the aim of the Republic is very simple” (1979:13), he also recognizes substantial complications concerning the interpretation of “Plato’s intentions” (12), pausing at one point during his summary of the dialogue’s argument to observe that “another complication is the fact that the gap between Plato’s actual view and what is generally attributed to him is even greater than what I have just said would lead one to suspect” (45).  White’s commentary is a sensible and practical contribution to the study of Platonism. I am simply noting that such gaps—between Plato’s individually held views, the views expressed in his dialogues, and the views attributed to him by his interpreters—reflect profound complications.  As I have been suggesting, contemporary attempts to unify and neutralize Plato’s dialogues under Plato’s univocal control are not consonant with the composition of Plato’s texts.
That said, the question of attribution is arguably less important than is the degree of Plato’s commitment to the propositions expressed by his characters. Plato wrote many things, which is to say that he authored many statements. Yet his commitment to these utterances, I will argue, is consistently and globally characterized by ongoing circumspection. When dealing with propositions, philosophers and linguists often focus on a statement’s truth value—whether the proposition expressed by a statement is true or false. And this is similar to the way professional philosophers are apt to judge arguments from Plato’s dialogues, as if abducting these arguments from textual slumber in order to interrogate their independent integrity according to absolutist standards of logical reasoning. But while propositions themselves may be judged true or false, the way a proposition is specifically formulated in language is also an expression of the speaker’s attitude. Through language, that is, “speakers commit themselves, to a greater or lesser degree, to what they are saying” (Matthews 1997, s.v. “assertive”). In linguistic terms, this degree of commitment, which contributes on a larger scale to our sense of the speaker’s or author’s cumulative stance and of the text’s overall status, is signaled by grammatical forms and constructions that are categorized under the heading of modality.
Although modality continues to be debated by linguists and further complicated by developments in modal logic, a nontechnical version of the concept should be perfectly adequate for our purposes. When dealing with sentences, which is what most linguists do by trade, modality may be considered a semantic notion that comprises “the speaker’s opinion or attitude towards the proposition that the sentence expresses or the situation that the proposition describes.”  Just as most languages have a distinct and seemingly neutral or unmarked procedure for making declarative statements that the speaker “believes to be true” (Palmer 1986:26), modality usually refers to a hierarchical range of attitudes that diverge from the directness of certainty about facts and reality.  The full range of modality or modalities is extraordinarily wide, covering territory from “possibility, probability, and necessity” to “permission, obligation, and requirement” (Palmer 1986:19). In general, though, modality is “the linguistic phenomenon whereby grammar allows one to say things about, or on the basis of, situations which need not be real” (Portner 2009:1). Among the many facets of modality, I wish to single out uses of language that address “situations which need not be real.” In this general sense of modality, which I will be using henceforth, to modalize a statement (or a stretch of discourse) is to qualify it as more or less remote from certainty (or alternatively, possible or probable to some greater or lesser degree).
Although they have other important functions, the subjunctive and optative moods of the verb in ancient Greek are commonly used to signal that something may or might be the case. These moods are marked morphologically in verb forms, but semantic modality can also be expressed lexically, by specific verbs, adjectives, adverbs, and particles; and its presence can be determined by contextual cues, which takes us beyond semantics proper into the domain of pragmatics.  Linguistics has traditionally limited the scope of modality to the study of sentences, and to grammatical forms in particular, yet “many of the features associated with modality may not be marked grammatically” (Palmer 1986:5), and this is certainly the case in ancient Greek. Linguists have taken an interest in discourse modality and in the related phenomenon of hedging, but the idea of modalizing tropes remains an open area for research.  Of course, grammatical moods in Greek have a variety of conventional, as opposed to strictly logical or argumentative, functions. Politeness is a primary one. My point, however, is not that Socrates frequently hedges in the Republic (although he does); instead, such hedging gains in prevalence by being part of an even broader pattern in the text. Thus even when Socrates is simply being polite, Plato’s composition—which operates far from the naïve simplicity of casual conversation—exhibits this politeness in a more self-aware and less socially conventional manner. Here I am proposing that modality offers a useful way to understand a major function of literary form in Plato’s Republic.
As already discussed, defining literary discourse per se is impracticable, so that the operative issue is whether a stretch of discourse is construed as literary by the interpreter. Lamarque and Olsen, whose definition of thematic statement I have adopted, stress the special lack of assertion (authorial endorsement or commitment) found in literary discourse. “A literary work,” they write, “does indeed have themes or meanings which can be stated as propositions, but … these propositions have a unique role in literary appreciation, which is not reducible to the role they might have in philosophy, religious discourse, or the social sciences” (1994:323).  The special status of these propositions in literary discourse is that they are nonasserted; or, in the terminology I am using, they are modalized. I should stress that statements of this sort are not empty gestures; thematic statements within recognized literary contexts “can be assigned significance and thus be understood without being construed as asserted” (328). As the discussions and arguments shared in Plato’s writings should make clear, such tentative exploration need not be a superficial dip into shallow ideas. This is why I will not shy away from suggesting that literary form in the Republic allows for a positive and substantial depiction of methods that are nevertheless inconclusive in principle as well as practice.
Several scholars have described Plato’s elusive stance in terms of distance, but their accounts have tended to fit it into their own pedagogical agendas, suggesting that Plato (covertly) uses the distance between himself and his audience to rhetorically persuasive effect.  The modalizing distance that I discern in the text of the Republic is more akin to the narratological category of distance described by Genette under the heading of narrative “mood,” and to the mediacy that Stanzel sees as a defining feature of narrative itself.  Although I see no need to treat the author as a narrator, Plato is like a narrator (and, in the Republic, Socrates is one) who controls information in ways that both deny and fuel our desire for complete and ultimate knowledge.  The very fact that Plato’s distance from speech in the dialogues needs somehow to be explained indicates its modalizing force; if, that is, I am right that interpreters must explain this distance away in order to reconstruct a subtext of univocal Platonic assertions.
Local Qualification in Socratic Dialogue
Most of what interests me about modality in the current study is a matter of semantic interpretation that extends well beyond the sentence. Some of the most important modalizing structures, such as fictionality and narrativity, are discursive and global rather than sentential and local, lying far outside the usual scope of linguistics but lending themselves to literary-rhetorical analysis. Yet the Republic, and the dialogues generally, are permeated by modalizing expressions at all levels of description, so much so that modalization might well be called a defining feature of Plato’s depictions of Socratic dialogue.  A promising way to understand modalization in Plato might be to focus on the amorphous phenomenon known as Socratic irony. I think my argument here, however, will be better served by focusing on linguistic features that are relatively unambiguous. Continuing to draw on concepts from the scientific realm of linguistics for the nonscientific purposes of humanistic interpretation, we may identify specific as well as broad indicators of discursive modality in the language of Plato’s text.
As a rule, conversation lends itself to specific grammatical constructions, and it becomes even more idiosyncratic in Plato’s portrayal of exchanges dominated by Socrates. Some general and commonplace modalizing expressions are ubiquitous and accentuated enough to be eminently familiar to readers of Plato’s dialogues. These include statements explicitly qualified by various lexical elements; questions, commands, and exhortations; and hypothetical conjectures, proposals, conditions, and forecasts. All of these constructions may be distinguished from straightforward declarative statements made in the indicative mood: they are forms of expression that do not communicate ideas, things, and situations as they are, but instead tender possibilities for consideration, describing what could or should be. Before undertaking a closer analysis of textual details in subsequent chapters, it will be worthwhile to survey briefly some of the ways that Plato’s dialogues, and the Republic in particular, are unusually saturated with expressions of possibility. As in English, modal expressions in Greek are often genteel and laden with potential irony. But the main function under consideration here is the sense of remoteness—the perspectival mediation of certainty, fact, and reality—that can accumulate under the weight of nondeclarative forms of speech.
The conversation between Socrates and his companions is predominantly clever, urbane, teasing, and polite, and its casual tone creates another dimension of constant qualification. Elements of this tone, such as wordplay, contribute to a small-scale tendency toward digression that resembles the distractibility of real-life speakers in casual conversation. Along with particles, which are used in myriad and virtually untranslatable ways (and many of which could be labeled modalizing particles), the text is thoroughly peppered with polite expressions of conversational modesty.  Of these, I would single out some verbs for special notice: horaō ‘to see’ (used of following the conversation, as in “I see [what you mean]”); dokeō ‘to think, seem’ (often used with the dative case to express an opinion, as in “it seems to me”); and phainō ‘to make appear’ (often used in the passive, conceding likelihood in terms of appearance rather than reality, as in “it appears”). These commonplace words, which act as incidental stitches holding the conversation together, reappear in consciously heuristic discussion as crucial topics: appearance, opinion, likeness, and reality. Such verbs are liberally supplemented by conversational expressions of likelihood, necessity, and truth—including isōs ‘perhaps’; eikos ‘likely, probable’; dei ‘it is necessary’; khrē ‘it must’; anankē ‘necessarily’; alēthē ‘true’.
Given Socrates’ obsession with questioning through conversation, a text such as the Republic naturally contains an extraordinary quantity of questions. Questioning is thus a foregrounded linguistic phenomenon in the text. It is also sometimes a debated topic in the conversation, as when Thrasymachus accuses Socrates of being an avid questioner but an evasive answerer (336c–338a). Some questions in the dialogue function as statements, others are really responses to questions, but the tone of questioning is often relentless. Imperatives of various sorts are also obviously much more prevalent in the conversation of the Republic than in, say, the nonfictional, historical prose of Thucydides. Characters urge themselves and the conversation forward in various indirect ways. “Is it resolved [dokei] that we must [khrēnai] try to carry this out?” asks Socrates at one point, questioning whether the group’s inquiry should continue, adding “I suppose [oimai] it’s no small job, so consider it” (δοκεῖ οὖν χρῆναι ἐπιχειρῆσαι περαίνειν; οἶμαι μὲν γὰρ οὐκ ὀλίγον ἔργον αὐτὸ εἶναι: σκοπεῖτε οὖν, 369b). While the legislative “is it resolved?” (dokeō) has a ring of decisiveness, the obligation of “must” (khrē) is softened by “try,” which prioritizes effort over success. Similarly, the resolution to accept the obligation is proposed in a question rather than declared outright. Then “I suppose” qualifies the litotes of “no small job” with an ironic recognition of the difficulty involved, and the imperative “consider it” offers the interlocutor Adeimantus a chance to back out. Along with commands (which are usually in the imperative mood), we have many hortatory uses of the subjunctive serving a similar function, particularly in the first-person plural. “Let us make [poiōmen] a city-state in discourse from the beginning” is a typically phrased Socratic exhortation (τῷ λόγῳ ἐξ ἀρχῆς ποιῶμεν πόλιν, 369c).
In their discussion of ideas and methods, the characters in the Republic spin out all manner of speculation: conjectures, provisional plans, and hypothetical scenarios (sequences of imaginary events constructed from nonasserted premises or conditions). Grammatically, these speculative ventures rely heavily on the optative mood. This question asked by Socrates is an interesting example:
εἰ γιγνομένην πόλιν θεασαίμεθα λόγῳ, καὶ τὴν δικαιοσύνην αὐτῆς ἴδοιμεν ἂν γιγνομένην καὶ τὴν ἀδικίαν;
“If we should watch [theasaimetha] a city-state coming into being in discourse, … would we also see [idoimen] its justice coming into being, and its injustice?”
369aFor now, suffice it to observe that this overdetermined sentence may be read as a methodological, conditional, speculative, and rhetorical question. Having the force of a conditional statement, it describes a possible state of affairs: if one should do this, one would do that. At the same time, with the force of a question, the sentence outlines a possible course of action, taking on the guise of a proposal, an exhortation, and a supposition that could be accepted. The hypothesis is a theoretical possibility that could, in turn, be accepted as a practical possibility.
In traditional grammars of ancient Greek, the should–would construction just observed is known as the future-less-vivid and is classified as a future or potential condition. There is a sense in which the lack of commitment in such conditional sentences already makes them function like questions; they are limited to proposing possibilities. As Goodwin writes in his standard reference work on Greek Moods and Tenses, “the question as to the fulfillment of the future condition is still undecided.”  Hence, in analyzing the example just discussed, we may refer to the possibility raised by the conditional if-clause (we might watch); the possibility of fulfillment in the then-clause (we might see); and the possibility, raised by the question, of accepting the entire statement as a supposition for further reasoning (we might agree). And let us not forget that the future is, strictly speaking, always unknown. As a modern linguist puts it, “in general, everything which relates to the future is qualitate qua hypothetical, since a possible rather than the actual world is involved.”  The future-less-vivid conditional sentence in ancient Greek is thus named because it emphasizes possibility over likelihood. But even expressions of likelihood (or plausibility) remain modalized (remote from certainty) to a significant degree. 
Indeed, the modality and temporality of making a city-state in discourse is more complex than it might at first appear. When planning city-states and their governments in the Republic, Socrates and his companions often use predictive, future-oriented constructions in ways that signal the theoretical force of their conversational practice. For example: “The farmer, as it seems, won’t make [poiēsetai] his own plow himself,” says Socrates at one point (ὁ γὰρ γεωργός, ὡς ἔοικεν, οὐκ αὐτὸς ποιήσεται ἑαυτῷ τὸ ἄροτρον, 370c). Even this apparently simple statement about an imaginary farmer is both a proposal for the city-state and a suppositional description of what would happen in the proposed city-state if it were ever made. Because Socrates’ proposal is looking for agreement, the sentence is also questioning; declarative rather than interrogative or conditional in form, but questioning and hypothetical in function. The sentence also happens to be qualified (“as it seems”) and negative (“won’t make”), features that further modalize its force. Meanwhile, because the speakers are making a plan rather than an actual city-state, what they are doing in the present (in the practice of actual conversation) keeps getting thrown into the future (toward the possibility of action). “We will also need traders,” remarks Socrates, for instance (καὶ ἐμπόρων δὴ δεησόμεθα, 371a). The plan needs traders in the present, but the planners would need traders if they actually were to make their city-state in the future.
Conditional constructions, which present one proposition as being dependent on (and thus subordinated under) another, are especially important for the interpretation elaborated in the second half of this book. Plato’s Socrates is fond of building elaborate sequences of argumentation that rest on hypothetical foundations. In grammar, where “a declarative main clause typically conveys the speaker’s commitment to the truth of the proposition expressed, such a commitment is often lost under subordination” (CGEL 174); even a statement like I know he is ill does not have the same force as he is ill. If there is indeed “a significant association between subordination and markers of modality” (CGEL 174), this is only compounded when an introductory main verb is itself explicitly modal in meaning (I doubt he is ill). My point here, though, is that the modalizing role of grammatical subordination can shed light on a related phenomenon in Socratic discourse, seen in theoretical experiments performed by way of words that accrete to form expansive chains of interdependent argument. In one basic pattern found in the Republic, interdependent claims (themselves often put forth as proposals, rather than firm assertions) derive from and refer back to a suppositional ground that is decidedly unasserted. When this happens, the hypothetical modality of the initial supposition can cast a persistent attitude of uncertainty over subsequent claims and conclusions.
Global Modality in the Republic
All of the modalizing features just described, which are common enough in ancient Greek but endemic to Plato’s dialogues, accumulate in the back-and-forth of conversation, forming an important aspect of Platonic style by saturating the text with modal expressions. If the Republic as a whole does not come across as modalized by these local instances of qualified speech, this is largely because the conversation in the book has a strong heuristic pulse that propels it through compelling arguments and thematic statements. No matter how modalized it may be, these statements of opinion undoubtedly whet readers’ desire to possess a linear Socratic-Platonic message. For the Republic portrays the evolution of a conversation from disorientation and aimless small talk into a self-conscious project, in which the speakers adopt methods for the purpose of understanding justice and injustice, and these topics are treated as objects to be pursued, seen, and found. At the same time, however, the discourse of this heuristic project is globally modalized by tropes that rhetorically establish distance between the seekers and their goals. Where an interpretation focused on expository, linear argumentation in the text might insist that this is Plato’s real meaning—that the book’s pattern of distancing is meant to replace the conversation’s pattern of methodical seeking—the second half of my study observes that these are major dimensions of the Republic that both exist concurrently and have contradictory functions. In other words, a strong current of methodological self-examination (a countercurrent, if you like) accompanies the heuristic arguments found in the book, articulating a perspective in which fixed conclusions keep turning into ongoing questions.
Recursion is the master trope, if you will, driving this thematic countercurrent. For instance, the way that the Republic’s conversational investigation of justice keeps doubling back on itself, becoming a conversation about the investigation of justice, is a major example of recursion in the text. We see recursion when a version of something occurs within itself, a phenomenon that can reach the extreme of infinite regression known as a mise-en-abîme. Here is how Jorge Luis Borges describes the famous literary example of the Thousand and One Nights, in which virtually countless narratives are embedded within each other: “This collection of fantastic tales duplicates and reduplicates to the point of vertigo the ramifications of the central story in later and subordinate stories” (1964:195). The Republic, I would stress, is permeated by similarly embedded structures of regressive self-reflection. 
Thematically, recursion in the Republic tends to emphasize the difficulties faced by human beings who are asking questions about the highest forms of knowledge. The closer the inquiry gets to talking about its highest goals, the more those goals are seen to recede from the seekers’ grasp. My analysis of the text will give special attention to interrelated varieties of recursion, notably fictionality, digression, reported speech, and hypothetical speculation. These are large-scale tropes that work in sequences of subordination—so subsequent events in a fictional narrative remain under the auspices of fictional premises; a digression is subordinate to a main topic; a report is subordinate to an original; and a speculative proposal is subordinate to a supposition. As recursive sequences (a digression in a digression in a digression, for instance), these tropes amplify modality by extending and persisting through longer stretches of text. The next three chapters consider different sequences: the Republic as a fictional narrative, the conversational inquiry into the ideal state, and the discussion of the Good as an ideal object of knowledge. In these sequences I discern a recursive, self-reflective, and modalizing perspective on methodical investigation that pervades the book as a whole.
[ back ] 1. Tigerstedt 1977:15, cf. 24, 92.
[ back ] 2. Tigerstedt, in a chapter called “The Problem,” is here paraphrasing Heinrich von Stein’s description of the hapless, uninitiated reader’s plight on encountering Plato’s dialogues (1977:15); the problem is the interpretation of Plato in the face of how Plato wrote his dialogues.
[ back ] 3. For the suggestion that theories have fared better than interpretations in the study of Plato’s literary form, see e.g. Cohn 2001:485, Rutherford 1995:26, Irwin 2002:199.
[ back ] 4. In moving from sentence to discourse, linguistic description encounters a similar proliferation of functions in English, since even basic “clause types” (declarative, interrogative, exclamative, and imperative) are used in many ways, while combinations of sentences in discourse can be interpreted as speech acts having many direct and indirect pragmatic functions (illocutionary meanings). See CGEL 36, 61–62.
[ back ] 5. Contrast the different project described by Irwin 2002, in which science may arrive at “true and historically accurate accounts of Plato” (199). See also Jeffries and McIntyre 2010:22–24 on objectivity in stylistics.
[ back ] 6. See Jakobson, “Linguistics and Poetics” (1987:62–94); Lyons 1968:70–81.
[ back ] 7. Jeffries and McIntyre 2010:30–34; Hawkes 2003:44–99; Fowler 1996:92–109; see Adams 1992 for Shklovsky, Mukarovsky.
[ back ] 8. Cf. the approach to language generally in CGEL: “The full support for a decision in grammatical description consists of confirmation from hundreds of mutually supportive pieces of evidence” (2002:21).
[ back ] 9. Cf. Hirsch 1976, esp. 41–44 on “patterns of emphasis” and “relative emphasis.”
[ back ] 10. See Heath 2002:39–57. A simple example: Hirsch, even when arguing that we must judge between two incompatible interpretations of a poem (Brooks versus Bateson on Wordsworth), nevertheless gives the impression that both readings are quite successful (1976:41–51). Certainly, neither is deemed a failure.
[ back ] 11. E.g. Crombie on the “unifying theme—of which indeed Plato gives no plain statement” in the Republic (1962:74; see also 80). See below on possible differences, not observed in Crombie’s terminology, between theme, thesis, and thematic statement.
[ back ] 12. See Lamarque and Olsen 1994:287, 321–338; along with Beardsley 1981:404.
[ back ] 13. Bloom’s translation of the Republic (1968), frequently modified, has been used throughout because of its literal accuracy. The Greek edition used is that of Burnet (1903). Transliterated Greek words cited throughout the text may be found in the glossary.
[ back ] 14. Cf. Heath 2002:11–16.
[ back ] 15. For extended discussions of Socrates as mouthpiece for Plato, see Corlett 2005; Cohn 2001; Press 2000; Wolfsdorf 1999; Edelstein 1998; Mulhern 1971.
[ back ] 16. According to White, the aim of the Republic is “to discover what justice (dikaiosunē) is, and to show that it is more beneficial, in a certain sense of that word, than its contrary, injustice (adikia)” (1979:13). Emlyn-Jones and Preddy: “Ostensibly the construction of a ‘Kallipolis’ (‘beautiful,’ or ‘fine city’), the purpose and aims of Republic are nevertheless far from straightforwardly expressed” (2013:1.ix). Cf. Reeve 1988:41–42; Crombie 1962:73–74.
[ back ] 17. Complaints like that made by Burnyeat, who says that for Leo Strauss the Republic “means the opposite of what it says” (1999:300–301), ignore the difference between the conversation, in which characters say things, and the book, which does not state or assert anything.
[ back ] 18. Lyons 1977:452; see Palmer 1986:1–23.
[ back ] 19. For considerations of the indicative as “unmodalized,” see Lyons 1968:307–308; Palmer 1986:26–29; Duhoux 2000:179; CGEL 173.
[ back ] 20. As Huddleston and Pullum put it, “mood is a category of grammar, modality a category of meaning. Mood is a grammaticalisation of modality within the verbal system” (CGEL 172; see also 33–43, 172–177). For a concise statement of reservations concerning such an account of mood, see Salkie 1988. On the modalities of the Greek verb, see Duhoux 2000:176–186.
[ back ] 21. On discourse modality in linguistics, see Palmer 1986:91–94; Portner 2009:3–8. See Enkvist 1985 for a discussion exploring the place of rhetoric in stylistics.
[ back ] 22. See also Lamarque and Olsen 1994:321–338; cf. Cohn 2001:489.
[ back ] 23. Nightingale 2002 (epistemological distance) and 2004:96 (rhetoric of estrangement); Rowe 2007:25, 28–32; McCabe 2008:103–106; Clay 2000:28, 37; see also Tigerstedt 1977:94.
[ back ] 24. Genette 1980:162–185; Stanzel 1984:6.
[ back ] 25. For other views of narrative in Plato, see Morgan 2004 and Halliwell 2009.
[ back ] 26. On the model of linguistics, levels of description here correspond to categories or dimensions of constituents, as in the level of words, phrases, or sentences.
[ back ] 27. On particles and qualification, see Cook 1996:139–155.
[ back ] 28. Goodwin 1890, sec. 392.
[ back ] 29. Wakker 1994:21. So Comrie 1985 notes that the future tense is “necessarily more speculative” than the past, “in that any prediction we make about the future might be changed by intervening events” (43). For more on tenses and modality see Fleischman 1989. Quirk et al. suggest that the future tense does not really exist in English, and see the verb form will as essentially a modal auxiliary (1972:84–90).
[ back ] 30. Scalar models of linguistic modality that measure degrees of remoteness from actuality are comparable with some philosophical theories of conditionals: “Conditionals can be accepted with different degrees of closeness to certainty” in “a range of epistemic attitudes” (Edgington 2008, sec. 3.1); see Ramsey 1990:145–163.
[ back ] 31. For more on the mise-en-abîme (a phrase drawn from heraldry by André Gide), see Dällenbach 1989; see also Bal 2009:62–64, who prefers to call such narratives “mirror-texts.”