Free Indirect: Timothy Bewes

Kazimir Malevich / Suprematism No. 58
Kazimir Malevich / Suprematism No. 58

Free Indirect : Timothy Bewes

“Free indirect discourse” and “free indirect style” are familiar terms in narrative theory, where they designate a mode of representing the speech or thoughts of a fictional character in the third person—directly, but without using quotation. Contrary to what is sometimes supposed, free indirect discourse is not in itself a technique of ambiguity. When Virginia Woolf represents the thoughts of Clarissa Dalloway in the opening of her 1925 novel—

Mrs Dalloway said she would buy the flowers herself.

For Lucy had her work cut out for her. The doors would be taken off their hinges; Rumpelmayer’s men were coming. And then, thought Clarissa Dalloway, what a morning—fresh as if issued to children on a beach.

—there is no doubt about the point(s) of view being presented. Gérard Genette coined the term “focalization” to denote the point of view of a passage of free indirect discourse.1 Between the first and second sentences, marked by a paragraph break, the focalization of the passage shifts from an apparently omniscient third person narrator to Clarissa Dalloway. Such an observation is not contentious, despite the fact that we are never informed of this shift by narrative commentary or by any grammatical indicator.2 Nevertheless, free-indirect discourse has provoked a rich body of reflection by narratologists and literary critics due to its supposed ability to give us the point of view of a character even while retaining, in some sense, the perspective of the observer. In one seam of thinking, free indirect discourse is held to be a “mixture” or “composite” of subjective and objective perspectives, of direct (spoken) discourse and indirect (reported) discourse. The literary critic Franco Moretti sums up the innovation of free indirect style with the formula “Emotions, plus distance.” Free indirect style, in Moretti’s words, achieves “the genesis of an unprecedented ‘third’ voice, intermediate and almost neutral in tone between character and narrator.” The individual voice thereby retains “a certain amount of freedom,” alloyed “with the impersonal stance of the narrator.”3

In an influential recent account of its presence in Jane Austen’s novels, D. A. Miller describes free-indirect style as a “radically cloven” practice, a mode of broaching “an impossible identification,” and suggests that it emanates as much from the narrator’s “persistence in detachment from character” as from a desire to merge with it. The initial separation of narrator from character is thus the precondition for the exquisite paradoxes that result, in Miller’s analysis, from their ambiguation. Free indirect style is a mode of “close writing” in which narration “comes as near to a character’s psychic and linguistic reality as it can get without collapsing into it . . . [T]he character does as much of the work of narration as she may without acquiring its authority.”4 The spatiality of Miller’s formulations is significant, even as it posits the possibility of nullifying all spatial differentiation.  Miller’s conceptualization of free indirect style, while reducing the disjunction between narrator and character to a mathematically minimal degree, preserves the place of the character in the field of vision of the author.5

Coexisting with this sense of a combination of perspectives, however, is an almost utopian tendency that celebrates in free indirect discourse something closer to the abandoning of perspective, that is, the emancipation of the character from the author’s (or narrator’s, or theorist’s) field of vision. This is the tendency that I wish to explore in the course of this essay, and to do so in the context of a specific definition of the political associated with the French philosopher Jacques Rancière. Rancière defines politics not as a question of the representation of subject positions but as a question of “the distribution of perception,” of “what is seen and what can be said about it”; of who, in a particular formation, is judged to have “the ability to see and the talent to speak”;6 a question, in other words, of what counts as a viable point of view and what does not. According to Rancière, the politics of literature involves the historical emergence of a “regime” of artistic production in which, for the first time, the subject matter could be anything at all, and the point of view that of anyone at all. “There are no longer rules of appropriateness between a particular subject and a particular form, but a general availability of all subjects for any artistic form whatsoever.”7 Rancière pays close attention to specific works (including Madame Bovary, À la recherche du temps perdu, and The Waves) but discusses them primarily insofar as they exemplify this turn-of-the-nineteenth-century formation and support his claim (contrary to theorists of a “postmodern” break, for example) of the continuation of this regime into the present. He identifies three key features of this development: “a way of writing that tends to remove meanings; a way of reading that sees this withdrawal of sense as a symptom; and finally, the possibility of interpreting the political significance of such a symptom in ways that are opposed.”8 Rancière pays attention to literature, then, not for the singularity of vision of particular works but as a large historical formation with “its own” politics, which, by implication, are not transferable outside it.9 Nor is Rancière conspicuously interested in theorizing the possibility of a practice that might escape the literary regime.

What happens, however, if we suspend our commitment to a historical, ideological understanding of the category of literature in order to ask not about the institution of literature, and the kinds of people who have access to it, but about particular literary works and how they may effect a certain redistribution of perception, irrespective of the historical formation of “literature”?

There are at least five different subject positions involved in the production and reception of novelistic literary works, which we can provisionally name as follows: author, narrator, reader, professional scholar and character. Each implies a different, relatively stable position in the distribution of perception; each is implicated in a different quotient of authority associated with the institution of literature. When I include “character” among these inhabitable subject positions, however, I am already forging a departure from the institution, which treats characters as projections or instances of subjectivities and social positionalities outside the work. That prior framing of character is itself part of the “distribution of the sensible” of the literary regime. But again, if we suspend our habitual dependence on the framing categories of “literature” or “fiction,” a totally new question comes into view: What does it mean to be a character? That is, how can we conceive the character not as an object but as a distribution of perception? Answer: To be a character is to exist as a term in a visual (or textual) field in which the conditions of visibility and expression are entirely determined by an authority external to oneself.

A second question becomes askable in the late twentieth- and early twenty-first century context: what possibilities exist, as authors, narrators, readers, scholars, or characters, for resisting, refusing or dissolving the conventional subject positions associated with these identities? To ask this is to open up the possibility of an escape from the literary regime in contemporary literature. It is also to implicate the literary theorist him- or herself in the formation of that regime. To what extent does the use of the term “regime” to talk about literature contribute to the consolidation of the literary institution as a particular distribution of the sensible?

The case I will try to make is that, following Rancière’s work, the question of who is entitled to speak is not only formulatable as a historical question of extending the rights of access to literature to previously marginalized subjects; it may also be formulated as a question of problematizing, even dismantling the expressive dimension of the literary utterance as such; that is, of producing an utterance that is devoid of hierarchy, any privileging of one sense or perspective over another; an utterance in which all normative values or orienting ideologies are merely latent or ineffective. My hypothesis is, first, that such a dismantling of perspective is a way—perhaps the only way—to construct the category and practice of literature as irreducible to any regime; i.e., to challenge the dominant sense that every element of literature is ideologically inflected; and second, that to do so is precisely the project of certain contemporary writers who are working without a particular commitment to literature as such. By reformulating the free indirect mode no longer as a “style” or a “discourse” but as a distinct, non-anchored subjectivity, we may begin to understand how contemporary fiction proposes a solution to the ideological problem of literature, a problem that, as we have known at least since Friedrich Engels, is a problem of perspective, not of content.10 Instead of seeing the politics of literature as a negotiation of points of view within a particular text, literature would become legible as a form that, while representing a plurality of perspectives, does not itself adopt, affirm, inhabit or express any perspective at all. This would be a literature defined by what Mikhail Bakhtin called a “zone of maximally close contact between the represented object and contemporary reality in all its inconclusiveness. . .”11

When it comes to abandoning perspective, we might draw a visual analogy with the development of the Russian painting avant-garde through the 1910s—from the fragmentation of point of view in a work like Malevich’s Woman at the Tram Stop (1913-14) to the complete obliteration of perspective in works of so-called abstraction such as Dynamic Suprematism (1915-1916) or Black Square (1915). When Bakhtin, writing about the element of “address” in the works of Dostoevsky, says that “there is nothing merely thing-like, no mere matter, no object—there are only subjects,” he is not talking about the complexity or the proliferation of points of view, but about the disappearance of perspective.12 “Narration in Dostoevsky,” writes Bakhtin, “is always narration without perspective.”13 Bakhtin himself draws an analogy with art criticism when he talks about the absence of any “distance perspective” on the heroes of Dostoevsky’s narratives. “The narrator is literally fettered to his hero,” he writes.

He cannot back off from him sufficiently to give a summarizing and integrated image of his deeds and actions. Such a generalizing image would already lie outside the hero’s own field of vision, and on the whole such images presume some stable position on the outside. [Dostoevsky’s] narrator does not have access to such a position, he has none of the perspective necessary for an artistically finalizing summation of the hero’s image or of his acts as a whole.

I want to illustrate this point not with a passage from Dostoevsky but with a number of sentence fragments and longer passages from the work of the contemporary Scottish writer James Kelman, author of the 1984 Booker Prize-winning novel How Late it Was, How Late. Kelman’s novel is the story of Sammy Samuels, an unemployed construction worker who, after being arrested and beaten by the police, wakes up blind in a police station cell. Sammy’s story is narrated in almost uninterrupted free indirect discourse. Both Sammy and the narrator appear to speak in the same Glaswegian dialect. More significantly, the narrator refuses to make Sammy an object of humor, reflection or pontification. Regarded in conventional literary-critical terms, Sammy is an exemplary subject of power, whose blindness consigns him metaphorically to a position of ultimate powerlessness. However, Kelman’s narration does not collude in that powerlessness; the way it avoids this is by refusing to make Sammy the object of an authorial gaze. This principle is never violated—even when the narrator’s distinctness from Sammy is registered grammatically.

Whenever Sammy walked it into town—which was usually always—then he took the road to the bridge…

[Sammy] glanced at his wrist but had fuck all watch on and he couldnay have seen it even if he had.

One day Sammy was doing a bit of shifting for a female that lodged in the same house as himself. Struggling along the road with a big fucking bundle of her suitcases and fucking poly bags man a million of the fuckers! So the guy I’m talking about, he came up and gave Sammy a hand. So one thing and another, Sammy wound up taking the guy for a drink—no just once but a few times; now and again, depending how he was fixed. The thing is but the guy didnay like drinking in pubs. He just wasnay a pub drinker. Ye meet guys like that. Even if they’re holding a few quid, they still prefer hitting the off sales. That was this guy, a real outdoor fucking person.

Okay, cutting a long story short here cause Sammy’s head was getting into a state and what was coming out wasnay always very good. The guy was fuckt I mean put it that way, he was fuckt, so there’s nay sense prolonging it. If ye’re wanting to play fair: alright? Let it go, fucking let it go, just let it go, a wee bit of privacy, know what I’m talking about, ye give a guy a break, fuck sake, sometimes it’s best just accepting that.

Sammy’s hand was on his forehead. He felt bad. He felt fucking awful man. It wasnay things closing in on him, cause it had already happened, it had happened; they had fucking closed in. He was beat. They had beat him. It wasnay his body. His fucking body man it wasnay his fucking body. It wasnay his body.14

The point of view of the narrator in such passages is one of “immediate proximity” to the hero (as Bakhtin says of Dostoevsky): all representation is structured from a “maximally close” and “aperspectival” perspective. Any notion of an economy between the “freedom” of the character and the “impersonality” of the narration misses the point—and the radicalism—of Kelman’s work. His use of free indirect style is no “composite” of direct and indirect discourse but a distinct formation that does not accommodate notions of balance, economy, spatiality, or perspective. There is no spatial differentiation between the point of view of the narrator and that of the hero. This means that there is no characterizable relation between them whatsoever, whether of irony or satire, empathy or sympathy.

This, I would suggest, is how we should read Malevich’s or Olga Rozanova’s “non-objective compositions” also—as the last stage in a project to democratize point of view, a project whose endpoint is the abandonment of perspective altogether.

It is on such grounds that Bakhtin, in a late text, conceives of free indirect discourse as existing at “the frontier of the philosophy of language.”15 Free indirect discourse (or as Bakhtin calls it, following his colleague and associate Valentin Vološinov, “quasi-direct discourse”) is predicated not only on dialogical relations between speakers or between utterances but on the dialogism of the word itself, a quality of sociality that arises from the appearance of the word in a particular social context. The word, says Bakhtin, like any sign, is “interindividual”:

Everything that is said, expressed, is located outside the “soul” of the speaker and does not belong only to him. The word cannot be assigned to a single speaker. The author (speaker) has his own inalienable right to the word, but the listener also has his rights, and those whose voices are heard in the word before the author comes upon it also have their rights (after all, there are no words that belong to no one).16

All discourse is, at its origin, indirect, a quality that carries language itself “beyond the boundaries of linguistics,” since linguistics, at least in the Saussurean tradition, considers language primarily as a signifying system, abstracted from the social conditions in which speech acts actually take place.17 By contrast with such linguistic approaches to meaning, for Bakhtin any relation to meaning is dialogic—produced only in real speech situations. “Understanding itself is dialogic,” which is to say, “unfinalizable.”18 Free indirect discourse enables, in Bakhtin’s words, the “represent[ation of] someone else’s idea, preserving its full capacity to signify as an idea, while at the same time preserving a distance, neither confirming the idea nor merging it with [the author’s] own expressed ideology.”19 That reference to “distance” is misleading, for in free indirect discourse the distance between author and idea is infinite, or we might say, vanishes to nothing.

By “full capacity to signify as an idea,” Bakhtin means that the idea of a fictional character is not an “object of representation,” nor merely a “characterizing feature,” nor a preexisting idea of the author projected into a character—one that might equally be embodied in another character or another work. The idea we encounter in Dostoevsky is fully independent, “fused” with the person in whom it is born, whose idea it is; as such it remains internally dialogic and therefore unfinalizable. The character, this is to say, is no longer caught in a field of vision, either of the author or of the narrator. Dostoevsky’s characters are “free people, capable of standing alongside their creator, capable of not agreeing with him and even of rebelling against him.”20 Here we touch on the real ambiguity of free indirect discourse; for in comprehending all these qualities, free indirect discourse is less a poetic practice than a model for understanding the true nature of language.

The logic of “free indirect” operates, then, at two levels. The first is inter-subjective or inter-discursive: the level of the third-person reporting of speech or thought. We might call this level rhetorical, since it leaves intact the boundaries of the entities that are thereby put into dialogue, and presupposes the determinability of who is speaking and what is intended in that person’s speech. The second level is intra-subjective or intra-discursive, that of the de facto functioning of language. Certainly free indirect, in Bakhtin’s understanding, is a poetics that operates rhetorically; but it is also a philosophy of language, the proposition of a dialogicality internal to language. The most celebrated work of dialogical literary criticism is Bakhtin’s study of Dostoevsky (Problems of Dostoevsky’s Poetics); but it is Vološinov who brings out the free indirect as a theory of language itself.

For Vološinov, in Marxism and the Philosophy of Language, language, in its essence, is not an abstract system of impersonal significations but a feature of everyday life that includes many “extraverbal” or “situational” factors to do with the “historical instant to which the utterance belongs.” Thus, language is not merely an interaction of speaker with speaker but of word with word; language is made up not only of “normatively identical forms” but of pre-existing utterances, anticipations of a response, and local inflections of meaning: an entire fabric of “social orientation” that is inseparable from the expression itself. Understanding is not a task of recognizing a familiar linguistic form but, on the contrary, of responding to its novelty.21

Language itself, then, has a “free indirect” quality, for the meaning of an utterance is inaccessible to a purely linguistic analysis undertaken from outside the concrete context. According to Vološinov, to attempt even “to delimit the object of investigation” is to “forfeit the very essence of the thing we are studying.” When free indirect discourse, in Vološinov’s words, “obliterate[s] the precise, external contours of reported speech,” it formalizes an asyntacticality that is inherent to all social discourse.22


The question I want to ask here is whether the notion of “free indirect”—liberated from the poetic or rhetorical notion of free indirect discourse or free indirect style—offers an answer to the question of the possibility of a nonregime mode of thought within art and literature. That is to say, can a concept of “free indirect”—a concept that would be at odds with its very conceptualization—provide the basis of an authentically political thought, the location of a political subjectivity?

We can also state this proposition negatively, from the other side: Is the “revolution” of free indirect discourse (as Rancière has called it) anything more than an “aesthetic” revolution?23 Can “free indirect” as a logic of perception rather than of discourse, shorn of its “aesthetic” qualities (which are also its qualities as a practice), offer anything more than a poetics—“an absolute manner of seeing things”?24 Can literature thereby escape the literary “regime”?

The essence of politics, we read in Rancière, is “dissensus.” And dissensus “is not a confrontation between interests or opinions. It is the demonstration of a gap in the sensible itself.”25 Thus Rancière contrasts two possible understandings of dissensus, analogous to the two notions of “free indirect” described above: on one hand, a novelistic discourse, a poetic practice, a technique for representing speech or thought, predicated on the stability of subject positions—both within the diegesis and without; on the other hand, a non-subjective quality that is intrinsic to language as such; a mode of reading as much as of writing—in relation to which the practice of free indirect discourse, in a vivid image from Vološinov, represents nothing so much as the rupturing of a “dike.” With the removal of the attributing clauses (“thought Clarissa Dalloway”) and subordinating conjunctions that respectively determine the direct and indirect reporting of speech, in free indirect discourse “authorial intonations freely stream into the reported speech.”26

The filmmaker Pier Paolo Pasolini, playing with the term used in Italian film theory for point-of-view shot (soggettiva), refers to the point of view of the camera as a soggettiva libera indiretta: a “free indirect subjective.” Pasolini is explicitly differentiating between the forms of literature and cinema, pointing out that “cinema does not have the possibilities of interiorization and abstraction that the word has”; that cinema “lacks the entire abstract and theoretical dimension which is explicitly involved in the evocative and cognitive act of the character’s monologue.”27 But Vološinov’s theory of language suggests that those literary qualities of “interiorization and abstraction” are not centrally part of language either—considered in its full possibilities as language. In this regard, free indirect discourse expresses the secret propensity of discourse to transcend the limitations of merely linguistic sense and enter into its fully “dialogic” existence. When Gilles Deleuze, drawing on Pasolini’s essay, talks of the cinema of Orson Welles, Michelangelo Antonioni, Jean-Luc Godard and Pasolini himself as putting forward a “free indirect vision,” a “free indirect subjective,” it is Bakhtin’s theory of the novel that provides him with the standard for what cinema is capable of: dispensing with organizational unity; bringing the distinction between subjective and objective into question; and replacing the totalization of images with “an outside which is inserted between them.”28 Even for Deleuze such insights long precede the discoveries of cinema.

As a statement of the principles of politics, Rancière’s “dissensus”—a gap in the sensible itself—might be seen to be part of a long tradition of radical thinking that locates the purest spirit of ideology-critique in the practices of artists and writers—or in the paradoxes (purposeless purposiveness) at the heart of the aesthetic. Lukács may be at the origin of this tradition. Étienne Balibar traces the theme of the “subject of history” to Lukács’s theory of the proletariat as the site at which the contradictions of capitalism become visible, the site, therefore, at which objectification may, and must, be transformed into political subjectivity.29 This theme—of a dissensus at the heart of the political subject—is crucial to the Frankfurt School tradition, as well as to the work of French thinkers as different as Alain Badiou, Jean-François Lyotard, and Deleuze. Its culmination, perhaps, is the position articulated by Deleuze and Félix Guattari in What is Philosophy? as paraphrased by Rancière: “art is politics.” “The success of a revolution,” write Deleuze and Guattari, “resides only in itself, precisely in the vibrations, embraces and openings it gives to men and women at the moment of its making and that composes in itself a monument that is always in the process of becoming. . .”30 As in Lukács’s notion of “imputed” class consciousness, the people in Deleuze and Guattari are always “missing”—always “to come.” This notion seems inseparable from Rancière’s insistence that political dissensus implies the lack of a “proper” place for the operation of politics, and the absence of any “natural” political subjects.

And yet, Rancière’s own thought differs from this, at least nominally. For within Rancière’s notion of dissensus as a “gap in the sensible itself” there is a further distinction between political dissensus and aesthetic dissensus. In an essay called “Literary Misunderstanding” Rancière characterizes these as two “divergent” paths, separated by the political operation of “subjectivation.” While literature “dissolves the subjects of utterance in the fabric of the percepts and affects of anonymous life,” the task of politics is to identify the anonymous—what Rancière calls “those without part”—as a “collective, an us.”31 This distinction between literature and politics brings us up against a number of difficulties, since for Rancière himself the mode of political “subjectivation” can never be consolidated—in the form of, for example, a subject position defined by its defense of particular interests. For Rancière, the subjects of a political demonstration are by definition “always precarious”; a political difference “is always on the shore of its own disappearance.”32 Given the forcefulness of these statements, together with the claim that politics “has no proper place nor any natural subjects,” it is difficult, and perhaps impossible, to distinguish between the political task of “subjectivation” and the literary operation of “desubjectivation.” What seems to differentiate art (literary dissensus) from politics, in Rancière’s conception, is simply that aesthetic dissensus enjoys the frame of a formal designator (the words “art” and “aesthetics”) to repair the dissensus. It is as an effect of these designations, then, rather than anything inherent in the activities thereby designated, that dissensus is turned into a mode of consensus. If aesthetics is a “regime”—born, as Rancière says elsewhere, “as the refusal of its name”—what is there to differentiate the refusal of the name that takes place as an axiom of aesthetics from the refusal of the name that takes place as an axiom of “politics”—other than this event of renaming?33 As modes of dissensus, the only difference between aesthetics and politics seems to be the name “aesthetics,” or the designation “regime,” both of which we owe to Rancière himself. Rancière’s distinction between aesthetics and politics would seem to be a version of the naming procedure that Rancière himself has called the operation of the police, and which he opposes to the political.34

And yet, if naming is the mark of the regime, who is to say when the name is “final” and when it is provisional, or—to use a term that I will return to—penultimate? Who is to say when Rancière himself is writing in the mode of the free indirect—that is to say, in the mode of represented speech and thought—and when he is not? When Rancière re-names the “aesthetic regime,” is this, too, not merely a penultimate word, a term used by Rancière not in his own name, with a definitive subjective commitment, but in the name of a thought, a regime, from which Rancière retreats at the very moment he speaks it? Is it possible that for Rancière, too, the “regime” of art retains an unnamed and unnamable element in which the work of political dissensus continues unrepaired? Rancière’s description elsewhere of the “absolute manner of seeing things” or the “absolutization of style” in Gustave Flaubert suggests as much. In this context, he says, “absolute” means not bound, but freed:

What have things been freed from? . . . They have been freed from the modes of linkage proper to characters and their actions that defined the genres of representation and determined the “styles” appropriate to them. … They are freed from the forms of presentation of phenomena and linkage between phenomena that define the world of representation. They have been freed from the nature that founds that world: from its mode of presenting individuals and linkages between them, from its modes of causality and inference: in short, from its entire regime of signification.35

And yet . . . There is always an “And yet”; for the mode of the “free indirect” is a mode of the impossibility of finalization, the mode of the word delivered with “a sideward glance,” the “word with a loophole” as Bakhtin calls it.36 It is by such means that Dostoevsky’s heroes retain “the possibility for altering the ultimate, final meaning of [their] own words” in the words themselves—and it is by such means that Dostoevsky retreats to an infinite distance from his heroes—or (which amounts to the same thing) approaches them to the point of an absolute proximity. “Judged by its meaning alone,” Bakhtin continues, “the word with a loophole should be an ultimate word and does present itself as such, but in fact it is only the penultimate word and places after itself only a conditional, not a final, period.”

The great fascination of Rancière’s work lies in the degree to which his own writing inhabits the free indirect mode, so flawlessly and consistently that it is almost impossible to attribute to it, or derive from it, a final position or point of view. Rancière too, then, precisely in his refusal of the name, may be said to represent another stage in the Lukácsian trajectory traced by Balibar, a late stage in which dissensus is inhabited by the intellectual not only in his or her discourse, his or her syntax, but in his or her writerly subjectivity. The unity of voice that every critical commentator or theorist aspires to is for Rancière no longer possible nor desirable.

In “Discourse in the Novel,” Bakhtin produces the following extraordinary passage:

Even when heteroglossia remains outside the novel, when the novelist comes forward with his own unitary and fully affirming language (without any distancing, refraction or qualifications) he knows that such language is not self-evident and is not in itself incontestable, that it is uttered in a heteroglot environment, that such a language must be championed, purified, defended, motivated. In a novel even such unitary and direct language is polemical and apologetic, that is, it interrelates dialogically with heteroglossia. It is precisely this that defines the utterly distinctive orientation of discourse in the novel—an orientation that is contested, contestable and contesting—for this discourse cannot forget or ignore, either through naiveté or by design, the heteroglossia that surrounds it.37

Here, the conditions of the free indirect seem to be extended to all novelistic discourse, indeed to every novel, no matter the frame or the intonation given to the utterances that appear within it. The boldness of the claim lies in its insistence that the novelist “knows”—even when everything in the work contradicts this—that manifest declarations of the novel are never unaffected by a “dialogizing background.” But what about when the novelist doesn’t “know” this? Where is such “knowledge” located then? In the text, in discourse itself, perhaps? And why stop at the novel? For if heteroglossia can affect the discourse of the novel even from outside the work, why not also the discourse of the historian, the sociologist and the philosopher?

When, say, the political theorist Adriana Cavarero coins the term “horrorism” in order to reconceptualize acts of violence without the agency that is implied in readymade concepts such as terrorism, or when the anthropologist Talal Asad critically dismantles the points of view from which Western liberal analysts have imported prior assumptions about motivation and causality into their readings of suicide bombings, we seem to be in a world in which free indirect—the free indirect subjective—is a plausible mode in which to undertake social and political analysis, one with its own distinct claims to responsibility.38 Free indirect, that is to say, is no longer a mode of represented speech and thought but an analytical register. No longer, then, a discourse that is inhabited subjectively, but a concept that is brandished objectively, and imbued with principles, conditions and limits, which is to say, a zone of application—an object. Is this what is meant by the elaboration of “free indirect” as a political concept? How can the penultimacy—to coin a term—of free indirect discourse be installed as a method and still retain its provisional quality? Doesn’t penultimacy, too, attain finality once it becomes a premise of the analyst? And are not all such gestures toward penultimacy rendered redundant or self-defeating by the de facto penultimacy of language?


In an essay written in 1986, J. M. Coetzee reflects on the challenges to the writer of fiction represented by the existence of the torture chamber, a place accessible (like the bedchamber) “to no one save the participants.” The torture chamber is thus emblematic of the limits of the novelist’s imagination; as such, it is at the “origin” of his or her calling, the compulsion to write. “In place of the scene he is forbidden to see,” says Coetzee, the writer “creates . . . a representation of that scene.”39 The predicament of the writer is that the limits of writing are thereby determined by the operations of a police logic (to use Rancière’s term)—the logic of a certain unrepresentability, a regime of visibility and invisibility. The problem, says Coetzee, is “not to allow [one]self to be impaled on the dilemma proposed by the state, namely, either to ignore its obscenities or else to produce representations of them. The true challenge is: how not to play the game by the rules of the state, how to establish one’s own authority, how to imagine torture and death on one’s own terms.”

Coetzee finds an answer to this dilemma—a novelistic answer—in the character of Rosa Burger, in Nadine Gordimer’s 1979 political novel Burger’s Daughter. Coming across a man flogging a donkey, Rosa perceives “the infliction of pain broken away from the will that creates it; broken loose, a force existing of itself, ravishment without the ravisher, torture without the torturer, rampage, pure cruelty gone beyond the control of the humans who have spent thousands of years devising it. . .”40 The passage is written not in free indirect discourse but, we might say, in the mode of the “free indirect subjective”; it describes a moment of perception without a transcendence, as Sartre might have put it, or “narration without perspective,” to use Bakhtin’s terms. Gordimer’s rationale for adopting this mode is clear to Coetzee. Does this man, he asks, “—‘black, poor, brutalized’—know how to live other than by brutality, doing unto others as has been done unto him?” Coetzee’s explanation for Gordimer’s use of the free indirect, in other words, is historical. Rosa looks forward, he says, to a time

when humanity will be restored across the face of society, and therefore when all human acts, including the flogging of an animal, will be returned to the ambit of moral judgment. In such a society it will once again be meaningful for the gaze of the author, the gaze of authority and authoritative judgment, to be turned upon scenes of torture. When the choice is no longer limited to either looking on in horrified fascination as the blows fall or turning one’s eyes away, then the novel can once again take as its province the whole of life, and even the torture chamber can be accorded a place in the design.41

The “time” that Coetzee is evoking is a postcolonial time, a time of liberation, when a quality of “humanity” will be restored to the relations of seeing and being seen; when visibility will be distributed equitably; when—to quote one of the most famous sentences from the work of Frantz Fanon—the black man will finally have “ontological resistance” in the eyes of the white man.42 In the meantime—and Coetzee is writing in 1986, a full eight years before the first free elections in South Africa—all discourse, all perception, all action (including nonaction) is implicated in regimes of inequality and injustice. Whatever Rosa does when she encounters the scene of brutality—whether she “bring[s] her authority to bear” on the man, by having him arrested, or drives past out of fear that she be thought “one of those whites who can care more for animals than people”—she is ensnared in positionalities that are unavoidably implicated in the apparatus of power. The web of implication extends beyond the diegesis to the figure of Gordimer, and to Coetzee himself. It is impossible to disentangle the writer’s gaze from the distributions of visibility that enable the writer to look in the first place, or the regimes of perception that attempt to repair the inequality with aesthetic effects—for example, the “false portentousness” or “questionable dark lyricism” that Coetzee finds in other treatments of the torture theme, such as Alex La Guma’s In the Fog of the Seasons’ End or Gillo Pontecorvo’s film The Battle of Algiers.43

Coetzee, like Rancière, refuses to offer prescriptions. Responding in an interview to a question about this passage in his essay—and implicitly about its failure to imagine a political subject into being—Coetzee differentiates between two kinds of duty: “To me, duty can be . . . an obligation imposed on the writer by society, by the soul of society, by society in its hopes and dreams; or it can be something constitutional to the writer, what one might loosely call conscience but what I would tentatively prefer to call an imperative, a transcendental imperative.” Having made this distinction, Coetzee adds: “I would not want to favor the first definition unhesitatingly over the second.”44

At which point one wants to add a sentence or two from the greatest theorist of the free indirect, Valentin Vološinov, the thinker for whom language and thus thought, beyond the understanding of the author, the linguist, or the recipient, even when presented as an authoritative speech act, and even when undertaken internally, in what Jean-Paul Sartre called “the intimacy of the For-Itself,”45 takes place irreducibly in the quasi-direct mode:

Tolstoy’s remarks about there being different kinds of thinking—“for oneself” and “for the public”—merely juxtapose two different conceptions of “public.” Tolstoy’s “for oneself” actually signifies only another social conception of addressee peculiar to himself. There is no such thing as thinking outside orientation toward possible expression and, hence, outside the social orientation of that expression and of the thinking involved.46

It is a point whose radicalism is completely grasped in Gordimer’s novel—which is to say, it is grasped by a character in Gordimer’s novel. Early in Burger’s Daughter Rosa reflects: “One is never talking to oneself, always one is addressed to someone. Suddenly, without knowing the reason, at different stages in one’s life, one is addressing this person or that all the time, even dreams are performed before an audience . . . It’s well known that people who commit suicide, the most solitary of all acts, are addressing someone.”47 Indeed, Rosa herself at this point is mentally addressing a third party, an absent lover named Conrad, a bourgeois whose views on the matter of address are directly opposed to Rosa’s.48

And yet . . . The capacity of the novel to comprehend and articulate such sentiments is a mark not of its ideological power but of its ideological limitation. As Bakhtin puts it, when discourse becomes an object of representation in the novel it is always as “ideologemes”: units of ideological significance. An aesthete can write an aestheticist novel only by inserting his or her aestheticism into the mind or mouth of a speaking person who “happens to be an ideologue for aestheticism.”49 Can we imagine a novel that gives effective ideological support to a regime of any kind, fascist or communist, democratic or autocratic, aestheticist or realist, monologic or dialogic? For Bakhtin it is inconceivable, for such authoritative elements inevitably “fall out of the artistic context” of the work and remain “dead quotation[s].”50 The environment of heteroglossia in which Rosa expresses the heteroglot quality of all thought and speech (“One is never talking to oneself, always one is addressed to someone. . .”) removes all privilege and authority from her expression, even as the expressed sentiment endorses that removal. This truth about language survives in Burger’s Daughter despite rather than because of its appearance in Rosa’s own discourse.

The other name for “heteroglot” is “free indirect,” a phrase that gives expression to the decentering and deauthorizing of free indirect discourse, and thus to the logic of novelistic thought itself. Between the material forms that heteroglossia can take within the novel—such as a proliferation of speaking characters—and the “dialogizing background” of the novel’s discursive environment there is a radical division. In the novel, whatever the author’s intentions, no political or aesthetic doctrine is able to appear overtly and retain its authority. That the irreducible distance between the thought of the work and the sentiments expressed in it is also conceivable as their infinite proximity is due to the defining political quality of the novel, which individual novels fall short of only to the degree that they have not realized their ideological potential: the obliteration of perspective.


Timothy Bewes is professor of English at Brown University, and the author of Cynicism and Postmodernity (Verso, 1997), Reification, or The Anxiety of Late Capitalism (Verso, 2002), and The Event of Postcolonial Shame (Princeton University Press, 2011). He is an Associate Editor of the journal Novel: A Forum on Fiction.


Published on May 24, 2017

1. Gérard Genette, Narrative Discourse: An Essay in Method, trans. Jane Lewin (Ithaca: Cornell University Press, 1980), 189.

2. Erich Auerbach, writing in an earlier moment about Woolf’s 1927 novel To the Lighthouse, observed that focalization is determined neither syntactically nor by any formal element, but rather by “intonation and context.” Erich Auerbach, Mimesis: The Representation of Reality in Western Literature, trans. Willard R. Trask (Princeton, NJ: Princeton University Press, 2013 [orig. 1946]), 535. For an alternative reading of the focalization of the first sentence of Mrs Dalloway see Alex Woloch, “Minor Characters,” Franco Moretti, ed., The Novel Volume 2: Forms and Themes (Princeton, New Jersey: Princeton University Press, 2006), 317-18. For Woloch the role of the novel’s opening line is to postpone the direct appearance of Clarissa Dalloway with the momentary perspective of Lucy, a minor character.

3. Franco Moretti, Graphs, Maps, Trees: Abstract Models for a Literary History (London: Verso, 2005), 82. The first description of free indirect style as a “peculiar mixture of direct and indirect discourse” is that of Adolf Tobler in 1887 (quoted by Moretti, ibid., 81).

4. D. A. Miller, Jane Austen, or The Secret of Style (Princeton, New Jersey: Princeton University Press, 2003), 58, 60, 59.

5. In a still more recent essay on Alfred Hitchcock’s cameo appearances in his own films, Miller formalized his method as “Too Close Reading,” a mode in which “objective” analyses all too easily “mutate” into “subjective” ones, and in which a critical essay can take on the quality of a “fantastic tale.” See D. A. Miller, “Hitchcock’s Hidden Pictures,” Critical Inquiry, Vol. 37, No. 1 (Autumn 2010), 126-127. In the light of this proposition Frances Ferguson re-reads Miller’s book on Jane Austen as a prior example of this “too-close” procedure. For Ferguson, Miller’s reading takes place on so personal a basis as to effect a departure from the “strongly generalizing tendencies in the criticism of literature” (Frances Ferguson, “Now It’s Personal: D. A. Miller and Too-Close Reading,” Critical Inquiry, Vol. 41, No. 3, Spring 2015, 524-25). Thus, for Ferguson, Miller’s “inordinat[e] close[ness]” to Hitchcock (as Miller himself refers to it, 127), as well as to Austen, may no longer have any spatiality to it at all. This would make Miller a prior explorer of the possibilities to be fathomed in this essay.

6. Jacques Rancière, The Politics of Aesthetics, trans. Gabriel Rockhill (London: Continuum, 2004), 13.

7. Jacques Rancière, “Are Some Things Unrepresentable?” in The Future of the Image, trans. Gregory Elliott (London: Verso, 2007), 118.

8. Jacques Rancière, The Politics of Literature, trans. Julie Rose (Cambridge: Polity, 2011), 9.

9. Rancière, The Politics of Literature, 44.

10. See for example Engels’s April 1888 letter to Margaret Harkness,

11. Mikhail Bakhtin, “Epic and Novel,” The Dialogic Imagination: Four Essays, trans. Caryl Emerson and Michael Holquist (Austin: University of Texas Press, 1981), 31.

12. Mikhail Bakhtin, Problems of Dostoevsky’s Poetics, trans. Caryl Emerson (Minneapolis: University of Minnesota Press, 1984), 237.

13. Bakhtin, Problems of Dostoevsky’s Poetics, 225.

14.James Kelman, How Late it Was, How Late (New York: Norton, 2005), 80, 105, 30, 51, 75.

15. Mikhail Bakhtin, “The Problem of the Text in Linguistics, Philology, and the Human Sciences: An Experiment in Philosophical Analysis,” Speech Genres and Other Late Essays, trans. Vern W. McGee (Austin: University of Texas Press, 1986), 119.

16. Bakhtin, “The Problem of the Text,” 121-22.

17. See V. N. Vološinov, Marxism and the Philosophy of Language, trans. Ladislav Matejka and I. R. Titunik (Cambridge, Massachusetts: Harvard University Press, 1986), 65-82.

18. Bakhtin, “The Problem of the Text,” 119, 121.

19. Bakhtin, Problems of Dostoevsky’s Poetics, 85.

20. Bakhtin, Problems of Dostoevsky’s Poetics, 79, 6.

21. Vološinov, Marxism and the Philosophy of Language, 100, 118, 67, 90, 68.

22. Vološinov, Marxism and the Philosophy of Language, 46, 120-21.

23. Jacques Rancière, Mute Speech: Literature, Critical Theory, and Politics, trans. James Swenson (New York: Columbia University Press, 2011), 115-16.

24. Rancière, Mute Speech, 116; Rancière, Politics of Literature, 10.

25. Jacques Rancière, “Ten Theses on Politics,” Dissensus: On Politics and Aesthetics, trans. Steven Corcoran (London: Continuum, 2010), 38.

26. V. N. Vološinov, Marxism and the Philosophy of Language, 146. “Subordinating conjunctions” is a term used by Ann Banfield in her study Unspeakable Sentences: Narration and Representation in the Language of Fiction (Boston: Routledge & Kegan Paul, 1982), to denote the words que, in French, and that or whether, in English, which establish positionality and hierarchy in the reporting of a thought or an utterance (71). As Banfield points out, free indirect style dispenses with—indeed, does not “tolerate”—such subordinating conjunctions. Their absence introduces an ambiguity to the status and the provenance of reported sentiments. Thus, in Nadine Gordimer’s novel Burger’s Daughter (which will be discussed later in the essay), as the protagonist Rosa narrates an old conversation with her lover Conrad about his past, we read: “He was vague about what he had done and how he lived” (New York: Viking, 1979, 20). Is Rosa reporting her own thoughts as she listened to Conrad speak, or her own thoughts in the presesnt, as she reports the conversation to us? “Two utterances,” says Vološinov, are brought together “in all their ideational fullness” (146). The rupturing of the dike between direct and indirect discourse can take a narrative as well as a grammatical form—as when, a few pages later in Burger’s Daughter, in an apparently direct address to the reader, Rosa informs us that the third-person description of her own profile in the preceding chapter was “concocted” after the fact with the use of “a hand-held mirror directed towards another mirror” (14). By such conceits Gordimer establishes the free indirect principle that operates at every moment of this text, even when the point of view is provisionally that of a conventional third person narrator.

27. Pier Paolo Pasolini, “The ‘Cinema of Poetry’,” Heretical Empiricism, trans. Ben Lawton and Louise K. Barnett (Washington: New Academia, 2005), 176-77.

28. Gilles Deleuze, Cinema 2: The Time-Image, trans. Hugh Tomlinson and Robert Galeta (London: Athlone, 1989), 148-9, 187-8.

29. Eva L. Corredor, “Interview with Étienne Balibar,” Lukács After Communism: Interviews with Contemporary Intellectuals (Durham: Duke University Press, 1997), 113-115; Georg Lukács, History and Class Consciousness: Studies in Marxist Dialectics, trans. Rodney Livingstone (London: Merlin, 1971), 159-72.

30. Gilles Deleuze and Félix Guattari, What is Philosophy? trans. Hugh Tomlinson and Graham Burchell (New York: Columbia University Press, 1994), 176-7; see Rancière, Dissensus, 172, 170.

31. Jacques Rancière, The Politics of Literature, 43, 36.

32. Jacques Rancière, “Ten Theses on Politics,” 39.

33. Jacques Rancière, “What Aesthetics Can Mean,” trans. B. Holmes, in Peter Osborne (ed.), From an Aesthetic Point of View: Philosophy, Art and the Senses (London: Serpent’s Tail, 2000), 18, 19.

34. Jacques Rancière, Disagreement: Politics and Philosophy, trans. Julie Rose (Minneapolis: University of Minnesota Press, 1999), 28-31.

35. Rancière, Mute Speech, 116.

36. Bakhtin, Problems of Dostoevsky’s Poetics, 232, 233.

37. Mikhail Bakhtin, “Discourse in the Novel,” The Dialogic Imagination: Four Essays, trans. Caryl Emerson and Michael Holquist (Austin: University of Texas Press, 1981), 332.

38. Adriana Cavarero, Horrorism: Naming Contemporary Violence, trans. William McCuaig (New York: Columbia University Press, 2009); Talal Asad, On Suicide Bombing (New York: Columbia University Press, 2007).

39. J. M. Coetzee, “Into the Dark Chamber,” David Attwell (ed.), Doubling the Point: Essays and Interviews (Harvard, MA: Harvard University Press, 1992), 364.

40. Coetzee, “Into the Dark Chamber,” 367; Nadine Gordimer, Burger’s Daughter, 208.

41. J. M. Coetzee, “Into the Dark Chamber,” 368.

42. Frantz Fanon, Black Skin, White Masks, trans. Richard Philcox (New York: Grove, 2008), 90.

43. J. M. Coetzee, “Into the Dark Chamber,” 366.

44. J. M. Coetzee, “Interview,” David Attwell (ed.), Doubling the Point, 340.

45. Jean-Paul Sartre, Being and Nothingness: An Essay on Phenomenological Ontology, trans. Hazel E. Barnes (London: Routledge, 1969), 222.

46. V. N. Vološinov, Marxism and the Philosophy of Language, 89-90.

47. Nadine Gordimer, Burger’s Daughter, 16.

48. “The will is my own,” Conrad insists to Rosa. “The emotion’s my own. The right to be inconsolable. When I feel, there’s no ‘we’, only ‘I’” (52).

49. Mikhail Bakhtin, “Discourse in the Novel,” 333.

50. Bakhtin, “Discourse in the Novel,” 344.