WhyWeReadFiction

by Theory of Mind

  Let me return very briefly to the arguments of Part II to clarify how what I said there about fiction and our metarepresentational ability differs from what I am saying here. There, I considered the possibility that certain fictional stories (especially those featuring unreliable narrators) play in a particularly focused way with our ability to monitor our sources of information. They portray protagonists who fail, on some level, to keep track of themselves as sources of their representations of their own and other people's minds, and, by doing so, they force the reader into a situation in which she herself becomes unsure of the relative truth-value of any representation contained in such a narrative. Detective stories, I propose in this chapter, play a slightly different game with our metarepresentational ability. Rather than encouraging us to believe what a given protagonist (e.g., Lovelace or Humbert) is saying, only then to slap us with a revelation that we should not have trusted him in the first place, the detective stories want us to disbelieve, from the very beginning and for as long as possible, the words of pretty much every personage we encounter. The two types of narratives thus build on the same cognitive capacity for storing information under advisement, but they approach it from different angles.

  One may argue, then, that detective stories literally exist for assiduously cultivating what Dr. Sheppard would consider a "rather . . . suspicious attitude" in the reader. In this respect, whodunits can be enjoyable and even addictive in the same way as weightlifting can be enjoyable and addictive: the more you train a certain muscle, the more you feel that muscle and the more you want to train that muscle. Note that I am using the far-from-perfect bodybuilding analogy on purpose to stress that just as not everybody is an avid bodybuilder—though everybody has a body and is in principle able to lift weights to train isolated muscles—so also not everybody is an avid detective-novel reader or is even remotely interested in detective narratives. Those of us who do not work out with weights still get enough indirect exercise from our everyday activities to keep our muscles from atrophying, and, similarly, those of us who do not read detective stories (or even much of any fiction) still get plenty of relevant interaction with our environment to keep our metarepresentational capacity "in

  2: Reading a Detective Story

  shape." The assumption that reading detective stories works out our metarepresentational capacity thus allows us to account both for the enjoyment that we derive from such stories and for the fact that such enjoyment is not universal.

  Furthermore, even if weightlifting makes one generally stronger, and detective-novel-reading makes one a veritable expert in the genre, both experiences still remain in many ways decoupled from reality. Just as overdeveloping one's triceps, biceps, and trapezoids generally does not give the bodybuilder any particular advantage in her everyday activities2—it certainly does not make one more adept at handling such crucial items as a pen, a laptop, a phone, and a fork—so keeping on a steady diet of detective stories does not make one a particularly discerning social player. It does not help me see through somebody's lies and it does not help me to know which "clues" to pay attention to in order to get to the truth of a given matter. In fact, applying what I have "learned" from a murder mystery to my everyday life could make me a social misfit: there is an important difference between being able, in principle, to revise one's views based on new evidence and going around deliberately suspecting everybody of being not what they seem, "just in case." In this respect, detective narratives may be said to parasitize on our metarepresentational ability: they stimulate it without providing the kind of "educational" benefit that we still implicitly look for in what we read. Delight they do, but instruct they don't, or at least not in the traditional sense of the word instruction.3

  The detective narrative's emphasis on exploring the furthest limits of our metarepresentational ability is the reason I prefer to focus on the novel and not on the classical form of the genre, the short story. Literary critic Jacques Barzun has suggested that the short story remains the "true medium of detection," for turning an elegantly economic piece into a "tangled skein of 150,000 words" accomplishes little else than adding the "artificial bustle and bulge" of false leads. Note, though, that put in "cognitive" terms the difference between the short story and the novel acquires a new meaning. Unlike its shorter counterpart, the detective novel veritably luxuriates in mind-reading; it adds more minds for the reader to consider and more metarepresentational framing to keep track of (or, as Jack Womack puts it on a different occasion, "the difference between the stories and the novels is the difference between coffee and methedrine"4). Of course, one of the founding fathers of the genre, Edgar Allen Poe, was already quite aware that his short stories were all about mind-reading, for as the narrator of "The Purloined Letter" famously discovers, figuring out the crime requires the "identification of the reasoner's intellect with that of his opponent" (13). Generally, however, the format of the short story limits the number of minds that could be read in-depth and titillatingly misread.

  However heavy-handed it may be, the parallel between detective fictions and weightlifting works on yet another level: in a culture that does not have a concept of weight-training facilities, or that considers muscular bodies ugly, or that frowns upon women exercising in such an "unfeminine" fashion, or that thinks that there is something unbearably ridiculous about setting aside significant amounts of time and money for lugging around pieces of iron, weightlifting of the kind currently widespread in this country would not exist. By the same token, there is nothing historically inevitable about the emergence, wide cultural acceptance, and long-term prospects of the detective genre, however apt this genre happens to be in stroking our metarepresentational ability.

  This emphasis on historicizing is crucial for the cognitive-evolutionary approach to literature championed by this study, and one of its broader ramifications applies not just to the detective genre. As we learn more and more about our metarepresentational ability, this knowledge may allow us to account, at least on some level, for certain fascinating regularities that we encounter in already existing cultural representations, such as literary texts, but it will never predict what cultural representations we are bound to have or cannot have in the future. Those are grounded in future history and as such are unpredictable even if they build on the same cognitive predispositions that have been with us for hundreds of thousands years.

  Thinking of the detective narrative as engaging in a particularly focused way our metarepresentational ability and yet being anything but historically inevitable puts on a stronger footing our project of historicizing the "rise of the detective story" phenomenon. Briefly, critics have offered explanations for the emergence and cultural entrenchment of the genre that range from sociopolitical (e.g., Howard Haycraft's hypothesis of the relationship between the detective genre and democracy) and scientific (e.g., Ronald R. Thomas's correlation of the rise of the detective story with the development of forensic technology), to ideological (e.g., Routley's argument about the relationship between the detective story and the English puritan tradition) and aesthetic (Joyce Charney's view of the detective novel as a latter-day response to the same set of aesthetic needs that used to be addressed by the English novel of manners). The endeavor to historicize the nineteenth- and twentieth-century detective story is often complicated, however, by the acknowledgment that we can find

  2: Reading a Detective Story

  "proto-detective" narratives in much earlier epochs, from Daniel's interrogating of the elders in the biblical story of Susanna in the garden to Sophocles' Oedipus and Voltaire's Zadig.5 Such acknowledgments seem to undercut, at least on some level, our attempts to situate the detective story in the nineteenth-century or the twentieth-century historical milieu and to explain its popularity by specific sociocultural developments of the moment. For if there is a detective story already present in the Bible, how can we speak about its "emergence" in, say, the 1840s, with the stories of Poe?

  The cognitive framework lets us address thi
s issue directly. It suggests that if (some form of) the metarepresentational ability has been with us since the dawn of the human species, then people have always had the potential for being interested in the stories that engage this ability. Consequently, by completely vindicating our suspicions that we have "always" had some sort of detective narratives lurking in our cultural history, the cognitive framework allows us to move on, so to speak, and to focus on the sociohistorical and aesthetic factors that might have contributed to the appearance, in the nineteenth century, of the detective story as a culturally recognizable, new, and special literary genre.

  Furthermore, our perspective on the permutations of this genre from the nineteenth century until today may, too, change once we posit as the key underlying characteristic of the detective story its tendency to engage in a focused way our evolved cognitive ability to store information under advisement. That is, we can begin to see the recent history of the detective narrative as a cultural chronicle of writers' experimentation with our metarepresentational ability and our Theory of Mind, pushed to their limits in several different directions. In the process of such experimentation, writers learn to negotiate and redirect cognitive challenges that may have first appeared insurmountable for their readers.

  A detective story seems to be particularly fit for such an analysis because the genre is relatively young, and we have access to the feedback received by the experimenting authors. That is, we know what initially caused an uproar in the audience but gradually became widely accepted and what, on the other hand, continues to constitute a problem even as generations of authors have tried their hand at circumventing it. The larger point that underlies such an investigation and that carries over to our thinking of other genres is that literary history as a whole could be better understood if we considered our cognitive predispositions as an important factor structuring the individual author's attempts to break the mold of what constitutes an acceptable and desirable literary endeavor of their own day.6

  In what follows, then, I consider four features of the detective story and, in some cases, their respective changes over time. I suggest that these features acquire a new psychological and cultural significance when approached from a cognitive perspective. The first subsection of my argument, "One Liar Is Expensive, Several Liars Are Insupportable," examines the care with which any writer of fiction treats the destabilizing presence of a lying character, the proliferation of potential liars being, of course, a trademark of the detective story. The second, "There Are No Material Clues Independent from Mind-Reading," emphasizes the detective story's ultimate goal of reconstructing the state of multiple minds populating the scene of the crime. The third, "Mind-Reading Is an Equal Opportunity Endeavor," addresses the genre's practice of strategic obfuscation of selected minds. The fourth, "Alone Again, Naturally," offers a cognitive reading of the old rule according to which, in an effective whodunit, the detective should be either celibate or married.

  3

  METAREPRESENTATIONALITY AND SOME RECURRENT

  PATTERNS OF THE DETECTIVE STORY

  wo points of clarification are in order. First, in the rest of this Part III,

  I use the term metarepresentation interchangeably with the term metarepresentationally framed information, meaning, in both cases, "information (or representation) stored under advisement." For example, when in one of my case studies, Maurice Leblanc's "The Red Silk Scarf," the police inspector concludes upon observing the behavior of two suspicious men in the street that they must be "plotting something," I call his interpretation a metarepresentation because it is "good for now," that is, it provides a temporarily useful explanation of the states of mind behind the suspicious behavior, but it can be adjusted, confirmed, or discarded any moment once more information comes in. In other words, I take it as a given (even though I do not say it again and again in every such case) that this explanation is stored with some sort of metarepresentational "tag," such as "the inspector thinks" or "we think," and that it is the implicit functional presence of such tags that makes it possible for us and for the

  3: Metarepresentationality and the Detective Story

  inspector to revise our interpretations as we go along.

  Second, I use here more insistently than in the previous sections such expressions as a "strong" and a "weak" metarepresentational framing to indicate that there are different degrees of advisement under which we "store" representations. For example, if I say to you that the rest of this section is divided into four parts, you have no particular reason to distrust me, and so you store this information with a "weak" metarepresentational tag, "Zunshine says that. ... " If, however, you are reading a detective story, you are encouraged by the laws of the genre to store nearly every attribution of the mental state behind each character's behavior with a very "strong" metarepresentational tag. If, for example, a potential suspect, "Flora," says that she left her room on the night of the murder because she wanted to get some water, the "Flora says" part of the representation—that is, its source tag—ensures that we still take her explanation into account, but we are strongly prepared to find that it is not true.

  The concept of variously weighted metarepresentational framings provides us with a useful framework for comparing detective novels with other works of fiction that have at different times been productively likened to them, such as Austen's Emma. Austen's novel has been described as "the most fiendishly difficult of detective stories,"1 and, indeed, its ending requires from us the type of cognitive work associated with the endings of detective novels. In a typical detective narrative, once the murderer is found out and his/her motivation is explained, we have to think back and revise our earlier interpretations of the events of the story, an important metarepresentational readjustment. Similarly, in Emma, once we are told the truth about Frank Churchill and Jane Fairfax, we have to reflect back onto the entire novel and modify our earlier interpretations of certain "clues," such as the timing of Frank's first arrival at Hartfield, the gift of the piano, Jane's insistence on getting her mail herself, and so on. Note, however, that when we read Emma the first time, we store the interpretations of these "clues," mostly provided by Emma, with relatively weak metarepresentational framing because, although ready to readjust them to some degree, we do not expect that they will have to be revamped so drastically. By contrast, the "real" detective novel alerts its readers early on to the fact that every bit of interpretation provided by characters ought to be distrusted until the end—an example of a very strong metarepresentational framing. Of course, within the same detective story, we can store information provided by different suspects, or even the same suspect on different occasions, under very different degrees of advisement; and, moreover, as we go on reading, we constantly modulate the relative strength of metarepresentational framing used to process the characters' presumed or claimed mental states. Still, the whodunit is associated with a much stronger internal metarepresentational framing than, say, a comedy of manners, such as Emma.2

  (a) One Liar Is Expensive, Several Liars Are Insupportable

  The reader of the detective story is supposed to "suspect everybody" (Paretsky, Bitter Medicine, 48). This constant readiness to keep under strong advisement any current explanation of any character's mental state comes at a price. To understand why it is so, we can turn again to the argument of the first section, in which I showed that Mrs. Dallotuay at times pushes our ability to process embedded intentionalities beyond our cognitive zone of comfort (i.e., beyond the fourth level). I think that it is significant that on such occasions Woolf does not try to make us guess at her characters' states of mind. Instead, she tells us quite explicitly what they are thinking, feeling, or desiring. In the scene at Lady Bruton's that I discussed earlier, we are told what Lady Bruton feels as she watches Hugh; we are told what Hugh thinks as he unscrews the cap of his pen and begins to write; and we are told what Richard thinks as he watches Hugh and observes Lady Bruton's reaction to Hugh's implicit assertions. The scene is challenging bec
ause the reader has to process a string of five- and six-order intentionalities. But at least Woolf does not require us to store the information about Lady Bruton's. and Hugh's states of mind under advisement by having implied, for example, that Lady Bruton and Hugh just pretend to be thinking about the letter to the editor and are really concerned about something else, and so Richard's complex reconstruction of their states of mind could be all wrong, and we have to wait for another ten or ninety pages to find out what Lady Bruton and Hugh were really thinking about. That is, within the world of the novel, we are allowed to consider the thoughts of the characters at this particular junction not as tentative guesses to be verified later but as architectural truths that can circulate freely among and affect indiscriminately our cognitive databases concerned with the lives and feelings of Mrs. Dalloway s characters.

  Now imagine a scene in a novel that embeds five or six levels of intentionality, as in, "A says that B thinks that C wants D to consider E's idea that F believes that X." This is already fairly difficult to follow for any

  3: Metarepresentationality and the Detective Story

  reader. Let us, however, complicate it even further and suggest that this is a scene from a detective novel, whose credo is to "suspect everybody." What it would mean is that not only do we have to process five or six embedded levels of intentionality, but we also have to consider on top of it that either A, or B, or C, or D, or E, or F; or both A and B; or C, D, and F; or all six of them are lying. I am not saying that is impossible to write such a scene (in fact, it may have been written at some point), but I strongly suspect that at least in the context of the literary history of our present moment, readers may find it rather incomprehensible. An author could play with multiplying the levels of embedded intentionality, as Woolf did, or an author could deliberately mislead us about the thoughts, desires, and intentions of her characters, as Sayers says all detective story writers should do; but it may take a presently unforeseen form of literary experimentation to usher in a work or a series of works of fiction that could successfully do both. At this point in our literary history, an effective whodunit can offer us red herrings again and again, but it tends to stay around or below the fourth level of embedded intentionality, and more reliably so than a non—detective story.