publisher colophon
  • A Critical Review of Derek Matravers’s Fiction and Narrative

Derek Matravers’s latest bookFiction and Narrative1—is a bracing review of many of the leading topics in the philosophical discussion of the intersection of—as his title indicates—fiction and narrative. A major aim of the book is to dethrone the prevailing view that the notion of the imagination plays a central role in the definition of fiction versus nonfiction. In addition, Matravers argues that the distinction we should care about in this vicinity is between representation and confrontation. Matravers takes up many other issues in this book, which is alive with argument. But these are the two matters upon which I focus in this review because they appear to be the most important.

Matravers dialectically paves the way for his own proposal regarding the representation/confrontation couplet by criticizing several prominent attempts to defend the fiction/nonfiction couplet in terms of imagination versus belief. These include the work of Kendall Walton, Gregory Currie, and, to a lesser extent, the position coauthored by Aaron Meskin and Jonathan Weinberg.

Interrogating Walton’s position is Matravers’s first order of business. As is well known, on Walton’s view, a fiction is a prop in a game of make-believe. Children play games of make-believe. For example, they might [End Page 569] play a game of pirates. In that game they have various props; straight tree branches, shorn of their leaves, play the role of swords, which the children use to duel one another. The game presupposes certain rules, such as that if you are touched with a sword/branch you are wounded. Walton proposes that we understand fiction to be the model of children’s games of make-believe.

For instance, when we read Gulliver’s Travels, we take it as a prop in our game of make-believe. Specifically: we make believe that we are reading the journal of one Lemuel Gulliver. Just as the branch is a sword in the game of pirates, so my copy is a copy of Gulliver’s journal, which I make believe conveys to me Gulliver’s assertions about his travels. Similarly, with respect to Dostoyevsky’s Notes from the Underground, we make believe that we are being addressed by the Underground Man. Unlike the children’s game of make-believe, these novels have been formally authorized or prescribed by an official author: Swift on the one hand, and Dostoyevsky on the other. These novels are mandated games of make-believe, whereas we could make up our own game of make-believe were we to stipulate that every pine tree we pass is a giant ogre to be rushed past and eluded.

Walton’s most famous example of his theory is not a literary fiction but an invented film—called The Green Slime, but with no connection to the actual film of that name. Walton asks us to envision a viewer, Charles, who is using the film as a prop in his game of make-believe. When the eponymous Green Slime lurches toward the camera, Charles makes believe that it is pursuing him, just as in the children’s game of pirates, a player without a sword (without a tree branch) is likely to either flee from or surrender to another child armed to the hilt with swords/ branches. Just as the swordless child will cower before the weaponized pirate/child, so Charles cringes before the make-believe onslaught of the Green Slime. Of course, the defenseless child is not really afraid of the make-believe pirate/child; nor is Charles actually frightened by the Green Slime. His emotion, like his situation, is make-believe. It is a make-believe emotion.

Matravers points out several telling disanalogies between children’s games of make-believe and our commerce with fictions. The most important, I think, concerns the de se aspect of Walton’s account. Walton’s version of make-believe is imagining de se—imagining about oneself. That is, we play the game from the “inside”; we ourselves as ourselves engage the prop. For example, Charles makes believe that the Green Slime has his number—that it is coming for him. This matches [End Page 570] the insider/player perspective of children’s engagement with games of make-believe. They’re in the game. However, this hardly seems a feature of consuming most, or even a great many, fictions. With most fictions, I have not been mandated to make believe that I have come upon some journal that purportedly records the experiences of its fictional author, such as in books like To Kill a Mockingbird.

It may be disputed whether we must be playing a game of make-believe that includes explicit fictional narrators. But even were that the case, there are so many fictions without mandated fictional narrators that the notion that we are making believe de se in response to them appears unlikely. Make-believe engagement, in other words, is not a necessary condition of our interaction with fictions. Moreover, if make-believe engagement is the case, then it does not seem that children’s (or even adults’) games of make-believe can provide us with a comprehensive model for the consumption of fiction. Though the child’s make-believe engagement with gobs of mud transforms them into pies, no necessary parallel de se engagement exists when it comes to novels. Thus, make-believe engagement cannot account for all fiction.

Others have questioned the de se component of Walton’s concept of imagination, and the example of Charles and the Green Slime has also seemed strained to some commentators. But Matravers’s criticism is nevertheless original in revealing that the problem lies at the very origin of Walton’s account of the initial analogy with children’s games.

After challenging Walton’s theory of imagination in his account of fiction, Matravers turns to Gregory Currie. Matravers correctly points out that simulation—imagining oneself in the circumstances of fictional characters—cannot frame the boundary between fiction and nonfiction, since we may simulate living people when we read the newspaper. This is true. However, simulation is not the only sort of imagination that Currie has maintained we mobilize when engaging fictions. It is what Currie thinks of as secondary imagining. Currie also cites primary imagining, which I conjecture is the kind of imagining Currie could claim is relevant in our negotiations with fictions.

Matravers is aware of Currie’s distinction between primary and secondary imagining, but says he finds it difficult to understand what Currie has in mind by “primary imagining” (Matravers, p. 30). This surprised me since I thought Currie’s primary imagining was the allegedly consensus view that Matravers dubs N1, namely, “What it is to be fictional that p (F(p)) is for there to be a prescription that we imagine that p” (Matravers, p. 21). Meskin and Weinberg’s theory, minus the boxology, [End Page 571] likewise identifies the imagination as crucial to accounting for fictional engagement.

What is common to many of the approaches to fiction that Matravers opposes, it seems to me, is their Gricean heritage. These theories seem to be roughly modeled on Paul Grice’s theory of meaning. This type of theory is what might be called an intention/audience-response model of communication. That is, for Grice, meaning—speaking very broadly—is the result of the complex intention of a speaker which intention is recognized by her audience and accepted for the reason that it is recognized as intended (see below).

Applied to the question of fiction, one version of a Gricean theory of fiction looks like this:

A structure of sense-bearing signs (e.g., a novel) X by sender S is fictional if and only if S presents X to an audience A with the intention that:

  1. (1). A recognizes that X is intended by S to mean P (a certain propositional content);

  2. (2). A recognizes that S intends A to imagine that P;

  3. (3). A imagines that P; and

  4. (4). that (2) is the reason for (3).

Or, to simplify matters, the sender S intends (and, in doing so, mandates or prescribes) that the audience, A, imagines that P. That is, the sender signals a fictive intention (imagine that P) which mandates the audience, A, adopts the fictive stance toward P (that it imagines that P).

In the preceding formula, condition (1) establishes that the audience understands the meaning of X (e.g., the sentences in the novel For Whom the Bell Tolls) by virtue of its uptake of the sender’s complex intention. In short, condition (1) is just a reprise of the Gricean theory of meaning. The definition of fiction really begins in earnest with condition (2).

Condition (2) requires that X be presented by the sender S with the intention that the audience imagine that P. But what does imagine mean in this context? The notion of the imagination can mean so many different things. For example, there is the visual imagination whereby we engage in mental visualizations. But that is not the concept of the imagination being invoked above. Rather it is what we can call the suppositional imagination. This is the use of “imagination” in phrases like “suppose x.” It contrasts with Walton’s notion of make-believe, since it has no necessary de se component as well as with Currie’s version of simulation. [End Page 572]

In order to get a handle on this notion of imagination, contrast it with the notion of belief. To believe X is to hold it in the mind as asserted. To believe that the earth is round is to hold in mind as asserted the propositional content “that the earth is round.” To suppositionally imagine, on the other hand, is to entertain a certain propositional content as unasserted—to hold it before the mind as unasserted. That is, to suppositionally imagine that “the French man-of-war bore down upon Captain Horatio Hornblower with guns blazing” is to entertain the propositional content “that the French man-of-war . . . ” as unasserted. It is not to take the propositional content as asserted; it is not to believe it. The prescription to believe, instead, would be a mark of nonfiction.

Reduced to a slogan, this account of fiction amounts to the mandate “imagine that P” in contrast to nonfiction, which at least involves the mandate “believe that P.” Moreover, the audience accepts the mandate to suppositionally imagine P (to adopt the fictive stance) because it recognizes that is what the sender intends (the fictive intention). Of course, in the case of nonfiction, the mandate to believe P (the mandate to adopt the nonfictive or assertoric stance) does not guarantee that the assertions presented are true but only that they are intended to be so regarded and thereby to be assessed by the relevant standards of truth.

In addition, this sort of approach appears to have a trim way of answering the question of what is “true in the fiction.” Namely, “what is true in the fiction” is whatever the author/sender mandates the audience to suppositionally imagine. Thus, it is true in the fiction The Hound of the Baskervilles that the “monster” is actually a big dog because Sir Arthur Conan Doyle intends us to suppositionally imagine that it is a big dog.

One way of looking at this approach is to note that it relies upon the mandate to adopt a certain propositional attitude—imagining—as key to differentiating fiction from nonfiction. Matravers questions the viability of this strategy on the grounds that fictions may involve believing as well as imagining, notably where we are dealing with truths in fictions.

As Matravers argues, truths in fiction present imagination theorists with a number of problems. The first problem is that there are explicit truths in fiction—truths that the authors of said fictions intend the audience to believe. When Victor Hugo says that Notre Dame is in Paris, he means you to believe it. He may intend you to imagine that Quasimodo lives there. But he certainly intends the reader to believe that Notre Dame is in Paris, just as he intends the reader to believe the theory he expounds about the history of symbols. Likewise, Stephen Crane intends you to believe that the North fought the South in the American [End Page 573] Civil War in the Red Badge of Courage, and Boris Pasternak intends you to believe that the Bolsheviks waged war against the White Russians in Dr. Zhivago. But how are imagination-accounts going to deal with this undeniable fact about fictions if fictions are to be categorized in terms of mandating the propositional attitude of suppositional imagining (which is characterized in contrast to that of belief)?

Furthermore, not only do explicit truths exist in fictions that authors mandate and readers believe; an indeterminately large number of implicit truths also lie therein that authors mandate readers to believe: implicit truths that must be believed if the fiction is to remain intelligible. In order to see this, recall that no fiction can proceed without presupposing that the readers will fill in many details that remain unstated in the work. Conan Doyle never explicitly records that Sherlock Holmes has five toes on each foot and that there is a left foot and a right foot rather than two left feet. The reader just supposes that.

In other words, the reader fills in or completes the fictional world. And her filling in of the fictional world in the right way is crucial to rendering the story intelligible. For example, when the protagonist halts as the villain points a revolver at him at point-blank range, the reader supplies the protagonist’s motive: the gun will kill him at that distance, the protagonist knows this, and he does not want to risk being killed. If the reader did not presuppose something like that, the situation would not make much sense.

But how does the reader fill in the fiction in the right way? She has a number of heuristics at her disposal. The first is the realistic heuristic; this tells her to fill in the story in terms of the kinds of details that obtain in the actual world. Of course, the realistic heuristic is not always pertinent; sometimes fictional worlds depart from the actual world; zombies rise from the dead. In cases like these, readers fall back on the conventions and presuppositions of the relevant genres. That is, they rely on what we can call the genre heuristic. And when this fails to render the story intelligible, we have other heuristics. For example, we may fall back on beliefs about the actual world shared by authors and audiences at the time that the fiction was composed. Call this the historicist heuristic. Which of the heuristics readers opt for depends upon which one or ones seem, upon reflection, to render the fiction maximally intelligible.

But what does this have to do with the objection to the imagination-account of fiction? Simply this: in light of the realistic heuristic, virtually any conceivable fiction contains truths that the author intends us to believe. The imagination-account is inadequate, since for any fiction [End Page 574] we are mandated to mobilize the propositional attitude of belief with respect to an indeterminately large number of details. So fiction cannot be defined essentially in terms of the mandate (the fictive intention) to adopt the fictive stance. With any fiction, we are most likely not only intended to imagine A, B, and C but also to believe (x, y, z . . . +n). That is, with any engagement with fiction it is not only that we will find it shot through with things we believe (especially given our reliance on the realistic heuristic) but which we recognize we are intended to believe.

That there are both explicit and implicit propositional contents in fictions that readers are mandated to believe is irrefutable. Matravers is absolutely correct about that. But does this mandate undermine the imagination approach to fiction?

At least two lines of response, it seems to me, are open to the imagination approach.

The first is to argue that whether or not we believe the explicit and/or implicit truths in the fiction is irrelevant. What we have been mandated to do is imagine them. We are mandated to imagine that Moscow is in Russia in War and Peace. That is the propositional attitude intended to govern our engagement with the novel. That we also believe that Moscow is in Russia is beside the point. For the purpose of consuming the novel, suppositional imagining is as far as we need to go. Supporting this perhaps is the common occurrence that when we come upon some claim in a novel of which we are uncertain, we are likely not to embrace it, since we know that much of the detail in a novel is invented.

Moreover, proponents of the imagination approach may also argue that it is not inconsistent for us to suppositionally imagine something we believe to be true. For the sake of argument, I may introduce a premise that I believe to be true with the locution “suppose x” as in “suppose the second law of thermodynamics.” Likewise for the purpose of fictioneering, the author may ask us merely to suppose something that in other contexts we would believe. But in this context, belief is not what is intended. Nor does this undercut the notion that there is a useful contrast between suppositionally imagining and believing, since what the imagination theorist is talking about here is which propositional attitude is intended to govern the encounter with fiction.

But this maneuver has obvious liabilities. For surely we are intended in some cases to believe rather than to simply suppositionally imagine certain authorial assertions. The chapter on whales in Moby Dick only makes sense if we believe that whales are as awesome as Melville argues in order to justify a story about whaling. If we knew that whales were no larger than minnows, Melville’s rhetoric wouldn’t work. [End Page 575]

Another response to what we might call the problem of belief in fiction might be to maintain that suppositional imagining is the default propositional attitude when it comes to fiction. That is, with fictions we are mandated to suppositionally imagine, unless reasons indicate that belief is what is intended. For example, in the Melville case, were we being asked to just imagine whales are awesome, that imagining would undercut his argument. Suppositionally imagining that the moon is made of green cheese goes nowhere toward proving that it is.

So, on this version of the imagination-account of fiction, fiction is defined in terms of an authorial intention to imagine P, unless we have reason to suspect that belief is intended. The overall, defining propositional attitude with fiction is suppositional imagining, although episodes of believing may be nested within the overarching complex authorial intention to suppositionally imagine. Here, of course, the question arises about when readers have grounds to think that believing is mandated. The answer to that question must be decided on a case-by-case basis in terms of what suppositions make sense of the text—what seems the most plausible way of rendering the text to be maximally intelligible. This, needless to say, is not an alternative to recognizing the author’s intention. Rather it is an albeit defeasible way of ascertaining the author’s intention.

Be that as it may, having, to his own satisfaction, defeated competing theories, Matravers presumes that the debate is open to the fielding of his own view according to which the fiction/nonfiction couplet should be abandoned in favor of a putatively superior distinction between what he calls confrontations versus representations. This distinction, moreover, he hypotheses is what theorists like Walton were really getting at all along. Representations, according to Matravers, are symbolic arrays, like photographs from yesteryear, that depict events and states of affairs outside of our possible sphere of action because they are removed from us in time, as in the case of histories, and/or remote from us in space, as in the case of a video on the nightly news of a demonstration in some far corner of the world. A crucial feature of representations, for Matravers, is that with representations, we cannot act to change the course of events depicted.

On the other hand, we can act in response to confrontations. A grandfather says to his family, “There are wolves at the mouth of our cave.” This declaration can lead to various responses. We may load our rifles, switch off their safety locks, and point them in the direction of the opening of the cave, while moving the children to the rear where [End Page 576] they will be out of harm’s way. That is, in the case of confrontation, we can respond to the recounted situation as it marks a face-to-face encounter in our egocentric space, thereby affording the opportunity to act in response to it.

The confrontation-versus-representation distinction, of course, evaporates the distinction between fiction and nonfiction in terms of the opposition of the propositional attitudes of imagination versus beliefs, since both motion pictures like From Russia with Love and camera footage like the Zapruder tape of the Kennedy assassination block responsive action in the egocentric space of the viewer. That is, representations, as Matravers understands them, are not necessarily marked by a unique propositional attitude. The distinction we really care about is supposedly between representations and confrontations.

The first thing to note about Matravers’s remapping of the conceptual field is that it does not seem to square with our ordinary language concept of representation. Dolls are representations in everyday parlance, but little girls can surely respond actively to them in their own, egocentric space. For example, they can change the doll’s diaper when it gets wet. Little boys can pet their stuffed puppies. And so on. Consequently, the incapacity to afford the possibility of interacting face-to-face in our egocentric space with a representation in our usual sense is not a necessary condition for that concept. At best, Matravers’s category is revisionist.

Traffic signs are also representations, but we can interact with them face-to-face in our egocentric space. We stop when we see the word “stop” painted on them. Nor need they literally rely on language. We stop at hexagonally shaped signs, not to mention red lights. So, once again, Matravers’s view does not track what we mean when we call something a representation.

Indeed, Matravers’s distinction between representations and confrontations seems quite arbitrary. Imagine a clerk in a liquor store is looking at a surveillance camera mounted over the doorway. He sees three men armed with assault rifles headed for the entrance of the store. That appears to count as a representation, since he is not having a head-on encounter with the content of the video screen, even though he might respond to it by pushing the alarm button under the counter that will send the police to the rescue. On the other hand, if his eyes had been focused just six inches lower at the sliding glass doors at the front of the store and had he hit the alarm as a result, that action would count as a confrontation. However, distinguishing between these two cases appears perfectly gratuitous, thereby raising questions about the sufficiency of the distinction across the board. [End Page 577]

Matravers does not deny that a distinction between fiction and nonfiction can be drawn, but such a distinction instead might rest on whether a work abides by the fidelity constraint (where the author includes only events she believes occurred, as in the case of nonfiction) or does not respect the constraint as in the case of nonfiction. Nevertheless, Matravers thinks representation is superior to confrontation.


Perhaps because of the way in which intuition pumps involving behavioral criteria play a role in elucidating fiction. Viewing Godzilla, we do not run from the theater screaming for our lives, nor do we alert the archbishop when reading The Exorcist. Commentators often cite these behavioral phenomena in discussing the nature of fiction. Specifically, they are often invoked as counterexamples to illusion theories of fiction. But, as Matravers points out, neither do we flee from the screen when we see simultaneously broadcast newsreel footage of tanks moving toward the camera in an actual wartime situation. Thus a difference in propositional attitudes does not putatively account for our behavior in these situations, since our behavior is the same in both cases. Something else is in play here, namely, that these cases are both cases of representation rather than confrontation. Nevertheless, even if Matravers is correct in undermining the frequent use of behavioral criteria in defeating illusion theories, the usefulness across the board of the confrontation/ representation contrast as a productive successor concept to the nonfiction/fiction couplet (as explicated in terms of different propositional attitudes) remains controversial given the anomalies cited above.

Noël Carroll
City University of New York


1. Derek Matravers, Fiction and Narrative (Oxford: Oxford University Press, 2014); hereafter abbreviated Matravers. [End Page 578]

Additional Information

Print ISSN
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.