How to Do Things with Novels

m

Twentieth-Century Literature

Reviews How to Do Things with Novels Why We Read Fiction: Theory o f Mind and the Novel by Lisa Z unshine C olum bus: O h io State University Press, 2006.198 pages

Ellen Spolsky

T h e title o f Lisa Z unshine’s new book reminds m e o f j. L. Austin’s title How to Do Things with Words. Like Austin, Zunshine promises a straightfor ward approach, as if to say: “R elax and listen to some straight talk about a topic others have overcomplicated.” Zunshine implies that h er jo b is to discuss the simple truth that real readers naturally like good stories. It’s hard to argue w ith her as I w rite this review in the week after the final H arry Potter b o ok was released and sold 2,652,656 copies in the first 24 hours (Guardian Unlimited). Z unshine thus seems to promise her readers— prim arily students o f literature and secondarily experienced academic readers and the general reading public— that they will no t be asked to struggle w ith unfamiliar scientific term s or radical ideas bu t will be rewarded w ith an interesting answer to the old question o f w hy we like to read fiction. She delivers w hat she promises and, like Austin, opens the way for continued explora tion o f the topic. A lthough her analyses o f texts em erge ou t o f her ow n close encounter w ith contem porary cognitive science and evolutionary psychology, her discussion o f empirical data from these fields is usually described as suggestive. Claims for the existence o f m ind as a neurological capacity, or even as a brain module, like vision o r hearing, are offered in W hy We Read Fiction but are subordinated to examples o f how useful it is. T h e proofs o f the approach will com e no t from outside o f the literary experience b u t from the readers’ ow n sense that it does indeed describe the way understanding in life and in literature feels. T he T heory o f M ind itself, or ToM , the theoretical scaffolding o f this study, offers a way o f talk ing about the ability to m ind read— that is, to infer the internal states o f m ind o f others by observing their behavior. Considering readers’ interac-

Twentieth-Century Literature 53.3

Fall 2007

394

Published by Duke University Press


Review tions w ith literary texts from this new perspective no t only offers a new vocabulary for som ething w e’ve always done— search out and explore the beliefs and motives o f characters— but also allows some new speculation about w hy people everywhere like reading or hearing stories, and always have. A lthough I can only hint at it here, I believe this new approach to reading literary texts adds significandy to w hat w e have already learned about language use from Austin. Like Austin’s book, which is a record o f spoken lectures, Why We Read Fiction is a w ritten version o f talks and classroom presentations in w hich the speaking voice is rarely muffled. Z unshine’s voice is that o f a generous and self-effacing teacher w ho summarizes the scene being analyzed for you, introduces new term s w ith hom ely examples, and belitdes herself rather than condescend.The b o o k ’s first sentence is: “Let m e begin w ith a seemingly nonsensical question,” and its second-to-last makes a claim she labels “modest,” although it is hardly that. As w ith Austin’s book, the disparaging w ink at academic theorizing suggested by the title is disin genuous. T he questions about language use and about the hum an love o f stories may be simple, b u t answering them , for both authors, involves m aking significant theoretical proposals about the structure o f the context o f language use and about the best way to account for the hum an ability to distinguish the m ore from the less relevant contextual inform ation and to understand o u r ow n position in a web o f interpersonal relationships. T he modesty o f the b o o k ’s rhetoric and claims, along w ith its u n pretentious choice o f texts familiar in the A m erican undergraduate cur riculum , are ju st w hat is needed for the subject, since even though claims m ade by cognitive scientists are currently popular in the media, so far they offer mostly suggestive rather than conclusive evidence about the brain function o f readers. Cognitive literary criticism is m aking waves not because it offers hard science to a mushy field o r because it supersedes older form s o f academic criticism or proves them useless o r mistaken. O n the contrary, Z unshine’s readers will understand and agree w ith her hypotheses only if they can, as they read w ith her, becom e aware o f their ow n processes o f understanding and interpretation— processes that are already in place but are normally subconscious. Zunshine states her claims as descriptions o f w hat “w e” think, how w e react, how we understand as we encounter the words, thoughts, and actions o f the characters in a novel. H er claim that readers’ theories o f m ind drive o r guide their understand ing o f literary texts gradually brings us to the recognition that Zunshine

395 Published by Duke University Press


Ellen Spolsky is n o t only describing w hat we do w h en we understand other people (insofar as we do), but that our understanding o f fictional characters fol lows from that real-life ability. “We,” then, are n o t only students o f literature b u t also all those w ho have ever listened to or told stories, since all humans, the argum ent goes, have an evolved m ind-reading capacity. T h e theory that we can all infer w hat others think or believe on the basis o f w hat we hear from them and observe about th em is part o f an account o f the universal hum an ability to live in groups— that is, to manage a cooperative social life. T he assump tion is that hum ans w ouldn’t have survived as a species if we couldn’t som ehow gauge w hat others were thinking w ithout their telling us, or w hat they were intending to do before they did it. Those o f our ancestors w ho did survive were those w ho had developed the ability to avoid the hostility o f others by predicting it (and running), in good time. Zunshine thus introduces to literary discussion the claim that all new b o rn hum ans need to understand their hum an condition o f intersubjec tivity, and the sooner the better. T hey are evolved to use their experience to track eye movements, facial expressions, bodily posture, and gestures for evidence o f others’ unseen thoughts, feelings, and m ost im portant, inten tions toward themselves. T he basis for our understanding stories, and also for o ur attraction to them , is that we all do com e to understand the paral lels betw een o ur ow n intentionality and that o f others, norm ally by the age o f three o r four. Because each o f us from infancy has been studying faces, o ur “default interpretation o f behavior is that it reflects a person’s state o f m ind” (4). Every successful encounter w ith another person, in life or in a story, “reinforces our tendency to make that kind o f interpretation first.” O u r ability to read no t only the faces and minds o f others but also pictures and even cartoon sketches, and o u r ability to enliven characters described by words on a page, suggests that we are always ready and able to create characters and to infer their internal states from linguistic and visual clues, almost no m atter how bare they be.Thus, w hen reading nov els, w e magnify even the slightest clues to help us understand the “minds” o f th e characters nam ed. N o t only are we evolved to be successful at this kind o f inference but also we can no m ore tu rn off o u r ability to read minds than we can hear our native language as merely noise. Authors de pend o n this ability and exploit it. “Fiction engages, teases, and pushes to its tentative limits our m ind-reading capacities” (4). Starting w ith a set o f



Review familiar texts (Mrs. Dalloway, Clarissa, and Lolita, am ong others) and add ing a discussion o f detective fiction (making entirely clear how little her theories depend on “high” art), Zunshine integrates the new perspective in ways that extend and deepen our understanding o f some im portant recent concerns in literary theory and criticism. Several o f the issues that have provoked debate am ong literary critics and theorists in recent years, specifically the phenom enology o f readers’ response, the identification of unreliable narrators, and the structure and meaning o f genre classification, are here opened up in new ways. In the last quarter o f the last century, reader-response theorists sought to describe the dynamic o f the reading process from the point o f view of readers whose understanding o f a text is built out o f personal experience yet is constrained by the semiotic codes and values o f their cultures and communities, w hich give structure to their experience. In Wolfgang Iser’s view o f reading, for example, the m eaning o f a w ork is a jo in t creation o f the author and the reader— som ething that materializes in midair, so to speak, betw een the text o f the book and the reader’s m eaning-m aking activity. M eaning-m aking, in Iser’s view, is som ething learned w ithin a culture and thus is different for different people, although, as Stanley Fish demonstrates, its rules m ight be so uniform w ithin a reading com m unity as to be virtually invisible. Some theorists argued that the meaning readers themselves create m ight be even m ore im portant and m ore interesting than the m eaning o f “the poem itself” or the author’s intended meaning. Some thought it m ight actually be the only m eaning available anyway. Considering the m aking o f m eaning from the perspective o f T heory o f M ind, readers use their ability to read minds in order to follow literary characters, and to follow their reading o f other characters’ minds, into several dimensions. Assuming that authors manipulate our m ind read ing by controlling and structuring the way they feed us inform ation, we can watch characters catch signals about w hat other characters m ight be thinking o f them , or o f others, in reasonably com plex patterns o f interac tive guessing. Except for the new vocabulary, this is som ething we always did. B ut in Z unshine’s opening example we see a new question on the horizon. She asks us how we know that Peter Walsh’s trembling (when he visits Clarissa Dalloway on the m orning o f her party, not having seen her for many years) “is to be accounted for by his excitem ent at seeing his old love again . . . and n o t by his progressing Parkinson’s disease” (3). Asking n o t about a reader’s response bu t about how the reader w ould account



Ellen Spolsky for that understanding, Zunshine’s question is almost but not quite the same as earlier theoretical inquiry. R eader-response theorists argued that we know w hat the words m ean because we know the social and literary codes that govern social and literary interactions. “K now ing” the codes, in this view, means know ing the default values. If you’re reading about a social encounter in a novel and no t about a case history in a medical text, then assume that trem bling in the context o f the relationship betw een a m an and a w om an signifies intense feeling. Experienced readers know w hich codes to apply and how to choose w hich o f the possible expla nations is likely to be the intended one. B ut Zunshine turns out to be getting at som ething different: she asks how it is that everyone, no m atter w hat their social context, assumes automatically that they can understand other people by observing their bodies. She is not asking for the evidence for our interpretation bu t rather w hy w e take that evidence to be evi dence. She inserts an earlier level o f questioning. W hat is it, she asks, that allows us to assume that we can understand internal meaning— intentions and beliefs— on the basis o f external evidence such as how people look and behave and w hat they say. In her discussion o f Mrs. Dalloway, then, and Clarissa, am ong others, Z unshine moves beyond the generalization that “every single act o f w riting and reading fiction” depends on and exploits o u r theory o f mind, and explores how different w riters “push to their limits certain aspects o f the general, constant, ongoing experimentation w ith the hum an m ind that constitutes the process o f reading and w riting fiction” (73; Z unshine’s italics).The investigation very soon helps us see how powerful is our ability to keep track o f “who thought, wanted, and felt w hat and when” (5). N o t only can we infer w hat someone is thinking, we can infer w hat they are thinking about other’s actions and thoughts. W h en Mrs. Dalloway’s husband R ichard observes their friend H ugh take out an expensive fountain pen to w rite a letter he has com posed at the request o f Lady B ruton to the Times, Virginia W oolf relates the small incident so that (here is Z unshine’s partial summary): “R ichard suspects that Lady B ruton indeed believes that because, as H ugh says, the makers o f the pen think that it will never wear out, the editor o f the Times will respect and publish the ideas recorded by this pen” (27). Grammatical analysis o f this sentence w ould clarify the levels o f em bedding it involves. Zunshine cites some recent experim ental w ork in w hich the limits o f our ability to correcdy untangle such em bedding were tested. T he experi ments suggest that after four levels o f recursive em bedding, m any people



Review lose track o f the intentionality as described. Looking at these exchanges from this point o f view allows Zunshine to distinguish am ong different styles o f narration and offer an explanation o f why W oolf’s novel “places extraordinarily high demands on our m ind-reading ability.”W hile we can all manage “A wants B to believe that C thinks that X,’’W oolf asks more o f us. For example, you’ll probably find the following nonliterary example impenetrable: “W hen you told m e you didn’t know w hat came o f M aria’s having telephoned Joel, were you assuming I’d recognize that you suspect the call was to ask him to encourage Alan to w ithdraw his resignation?” Thus it’s n o t surprising that inexperienced readers are afraid o f Virginia Woolf. Zunshine reports, however, that w hen she explains to her students that “W oolf tends to play this particular kind o f cognitive ‘m ind gam e’ w ith her readers, it significantly eases their anxiety . . . and helps them to start enjoying her style” (35). As the com phcations w ithin narratives multiply, so do the opportu nities for misunderstanding. It is virtually conventional in nineteenthcentury English novels for the plot to be shaped by a m ajor character’s misreading o f another character’s m ind early on, and, as happens in many o f Jane A usten’s novels, the rest o f the plot brings the protagonists around to better understanding and thus to the happy ending. (Zunshine dis cusses Pride and Prejudice and Emma.) Some characters, o f course, d o n ’t m isunderstand, and readers may learn to identify a reliable m ind reader in the story and depend on his or her interpretations. T he w riter, as we saw ofV irginia Woolf, depends on but also may challenge readers’ abili ties to keep track o f w ho thinks w hat about w hom w hen. Sometimes the reader can see that a character has m isunderstood, but sometimes the reader is misled along w ith the character. To describe these com plex possibilities, Z unshine introduces a concept o f m etarepresentation to describe the hum an m ind’s talent for form ing and using representations about representations. T he m etarepresentations tag the representations to w hich they are attached w ith source inform ation and w ith qualifying evaluative and contextual meaning. T hey link representations w ith other inform ation that may be useful w hen it is retrieved. T hey footnote the representation, as it were, storing the inform ation by w hich one can later ju dge the reliability o f the representation by recording w ho said it and in w hat circumstances. In fiction the tagging system works the same way, but it’s the author, n o t the world, w ho controls the am ount o f inform a tion available and the tim ing o f its release. A quick look at a ballad in the



Ellen Spolsky repertoire o f Joan Baez will show just how im portant m etarepresentation is in determ ining action and how the author a n d /o r genre control it. In a tale o f thw arted love, the shiny object o f the title, The Silver Dagger, seems at first to be the m ajor threat to the lovers’ union, but in four short stanzas the speaker makes clear that her m other’s strongest w eapon is her words, and it is they that do the damage. T he m other’s lesson is about the need to understand metarepresentation, and is based on h er ow n experience o f being lied to. D o n ’t sing love songs, you’ll wake my m other She’s sleeping here right by my side A nd in h er right hand a silver dagger, She says that I can’t be your bride. All m en are false, says my m other, T hey’ll tell you wicked, lovin’ lies. T h e very next evening, they’ll court another, Leave you alone to pine and sigh. M y daddy is a handsome devil H e ’s got a chain five miles long, A nd o n every link a heart does dangle O f another m aid h e ’s loved and wronged. G o court another tender maiden, A nd hope that she will be your wife, For I’ve been warned, and I’ve decided To sleep alone all o f my life. T h e traditional density o f the ballad form lets us see quickly how dependent the interrelationship o f the four personae is on m ind reading. T h e first-person speaker tells us not only that “All m en are false” but also how she knows: “says m y m other.” Trusting h er source, she tells the m an to “ Go court another tender maiden,” m aking explicit her trust in her m o th er’s words: “ For I’ve been warned, and I’ve decided.” “All m en are false” is the focal representation— the crucial content on w hich the speaker’s four actions tu rn (her opening com m and to som eone not to sing, her decision, refusal, and recom m endation to him). She has tagged the content w ith a label (my m other told me) that will rem ind her o f



Review its source. She uses it in com bination w ith h er ow n knowledge o f her fathers behavior to decide to trust its reliability.The last sentence, further, tags her ow n representation (“I’ve decided”) for her audience, assuring us that it’s her ow n decision— no t her m other’s silver dagger— that backs her decision. N o te that the audience o f the song can make an additional m etarep resentation about the content based on its ballad genre. T he last stanza in this ballad provides inform ation that suggests a last-m inute réévaluation o f the moral, although it is one the audience makes, no t the speaker. T h e picture we develop o f the protagonist in the first three stanzas leads us to believe we are dealing w ith the breakup o f a love affair based on reasonable advice from a m other to a daughter. B ut after hearing in the last line that the speaker rejects no t only the m an to w hom she speaks but all m en, we may decide that the m other’s advice unfairly hobbles her daughter’s future by leading her to decide to “sleep alone all o f [her] life.” As Zunshine describes it, it is no t only the speaker w ho has the ability to tag inform ation. We the audience also hold inform ation “under advise m ent,” allowing us to change o u r m ind about the value or truth o f the w hole narrative w hen new inform ation emerges. As we have seen in the short ballad narrative, and as we read a novel, know ing the origin o f any particular scrap o f representation is crucial. In fact, whatever we know about the context o f the scraps o f inform ation that faute de m ieux guide our lives helps us reuse it felicitously, to use Jo h n Austin’s term for the successful perform ance o f a speech act. And as Zunshine demonstrates, “our tendency to keep track o f sources o f our representations— to metarepresent them — is a particular cognitive endow m en t closely related to our m ind-reading ability” (47). Examples o f our pow er to assess representations for their reliability by m aking inferences about their source occupy several interesting sections o f W hy We Read Fic tion. Since 1961, w hen Wayne B ooth called attention to the phenom enon o f unreliable narrators, critics have broken dow n the category further and w ith further subtlety.1 Given how difficult it can be for readers o r for a character in a novel to follow all the layers o f reliability at w hich inform a tion comes to them and to tag them properly, each w ith contextual com plications, it w ould seem that far from being the default assumption, the reliability o f the inform ation available to readers and to other characters in the book is n o t occasionally but typically doubtful. T he possibilities o f em bedded misunderstandings are dizzying and, depending on the genre,



Ellen Spolsky lead to tragedy or to farce. Zunshine describes one such error in Clarissa, a novel that articulates, w ith a hitherto-unprecedented intensity and detail, the them e o f the correlation betw een arduous m ind-reading and tragic m isunderstanding.. . . M ind-reading is a crucial aspect o f o ur everyday existence, but a character too occupied w ith figur ing out other people’s states o f m ind, and worse, flaunting his ability to “see through” other people, runs a grave m etarepresentational danger: he can easily lose track o f him self as the source o f his representations o f the other persons m ental world. (89) Zunshine also cites an episode from the T V series Friends as an ex ample o f how the same kinds o f complications can lead to farce. T he close dependence o n genre o f too m uch unreliability is an im portant point, and one easy to miss because o f the overall and correct claim Zunshine makes forToM , namely that readers approach the characters in books and understand the understanding o f authors and o f characters in books by means o f the same skills they depend o n for their everyday com m unica tive (interpersonal) com petence. At the same time, novels are no t life, and crucially, although ToM works well enough often enough for it to be described as a felicitous evolutionary advance, novels and other fictional narratives are precisely n o t interested in the way things norm ally work. T h e m ind-reading successes o f everyday life fall into the background o f novels, while the failures are news or, as M ary Louise Pratt describes them, “tellable” (136). T h e boundary genre o r test case in w hich the jo b o f sorting useful knowledge from disinform ation is, as Z unshine recognizes, the detective story, to w hich the last third o f her book is dedicated. H er interesting exam ination reveals, for example, how the genre is constrained by the high demands m ade on the readers’ m ind-reading abilities o f having to assume (as the detective must) that everyone is lying. It is her discussion o f the way m ind reading interacts w ith the distinctive demands o f differ ent genres that leads Zunshine to the concluding discussion o f w hether different genres m atch different ways o f know ing (different ways o f m ind reading). This question seems to suggest that if they w ork in different ways, they do different things. And if they do things at all, are they doing things that make a difference? Are they teaching us o r training us in any way? To use our innate intersubjective skills m ore efficiently, perhaps,



Review o r m ore ethically? D o novels, by exercising our m ind-reading abilities, train us to live w ith others? This is a conclusion Zunshine hesitates to embrace. If we understand literary texts because we understand other people, and whatever is consciously designed or contrived by an author is su perim posed upon or grounded in the real world o f hum an brains and know ing, then, minimally, reading novels should afford the same kind o f practice in m aking appropriate guesses about the unseeable minds o f oth ers as real interpersonal actions do. Except that the minds o f characters aren’t minds: they’re fictional creations. Is it possible, then, that the fictionality o f a text cancels or interferes w ith or brackets this experience— keeps it from teaching us— in the same way that know ing w e’re watching a play keeps us from ju m ping up to save D esdem ona before O thello strangles her?2 D oes the genre determ ination (play o r novel, for example) give a reader permission to bracket the intended o r unintended ethical teach ing o f a w ork o f art? It certainly seems impossible to dem onstrate that readers o f Anna Karenina or Madame Bovary learn not to be unfaithful to their husbands. Does fiction do m ore than “engage and titillate our m etarepresentational capacity” (79)? Zunshine seems careful never to overreach herself, having learnt an im portant lesson from earlier literary scholars w ho have been criticized for the counterintuitiveness o f their claims about w hat cognitive science can tell us about literature. She never allows herself to claim m ore than that novels provide pleasure, (certainly a safe claim, because it’s w hat most people w ould say if asked w hy they read novels), but she does define pleasure in a new way. She sees the pleasure afforded by fictional narratives as grounded in o u r awareness o f the successful testing o f o u r m ind reading adaptations, in the respite that such a testing offers us from o ur everyday m ind-reading uncertainties, o r in some com bination o f the two . . . [with the proviso that] the joys o f reading fictional minds are subject to some o f the same instabilities that render our real-life m ind-reading both exciting and exasperating. (20)

O n the very last page o f the book Z unshine m entions several ways in w hich our exercise ofT oM in reading fiction may reach further than the above description allows. M uch w ork remains to be done to specify the nature o f the pleasure o f reading, and I’d guess that m ore w ork on



Ellen Spolsky the heuristic pow er o f specific genres w ould be relevant here. Zunshine has used the term s fiction, stories, and novel interchangeably throughout h er text, but I w ould think that clearer distinctions am ong the terms w ould b rin g us closer to answering questions about the relationship betw een fictional teaching and delighting. Novels certainly make the claim o f exhibiting truths; even if the narratives aren’t themselves true, and even w hen they are outrageously fictional (thinking o f the H arry Potter books again), they certainly can claim to be “true to the truths o f the hum an heart,” in N athaniel H aw thorne’s words. We will gain, I sus pect, by w orking out the kinds o f claims that different genres and even different subgenres conventionally make on readers. D etective stories, for example, expect readers to use their everyday logic and knowledge o f the real w orld to com pete w ith the detective in solving a mystery, but it goes w ith o u t saying that the crim e itself, the plot, is contrived. Fiction is prob ably n o t a useful genre term because fact and fiction are so thoroughly m ixed in texts and in life. I w onder if Zunshine w ould agree that we read novels and exercise o u r m ental endow m ents precisely because the genre gives us permission to do so w ith o u t the consequences that such “w ork” w ould have in the real world. T he genre o f novel (not o f story, or o f fiction) may be plea surable for the very reason that we can, say, think about adultery w ithout the consequences o f indulging in it. A nd as long as you cover your book in a plain brow n wrapper, there w ouldn’t even be such consequences as m ight arise if others know w hat you’re reading. This is still in the realm o f speculation, b u t I’d say that the distinctions betw een fiction, narrative, story, and novel are going to have to been taken m ore seriously before a fuller answer to w hy we read novels can be given. Z unshine’s b ook is not only a good beginning; it will also serve, as Austin’s How to Do Things with Words did, as a stimulant to im portant fur ther work. We are already seeing, in the year since the book appeared (it is already in its third printing), m ore detailed w ork on the operation ofToM in reading novels and on its neurological underpinnings. B ut I am particu larly interested in seeing the developm ent o f this w ork in organizing and providing structure to the wastebasket we now call context. ToM should be able to contribute to a theoretically satisfying answer to the question o f why author’s intentions have always seemed naturally to be o f prim ary im portance and yet were so easily dethroned w hen challenged by French and Am erican deconstructionist theory and criticism. T he brain question



Review is: just how do we decide (and the decision has to be instantaneous) what parts o f the context o f an exchange should be attended to? Zunshine’s b o o k sets us on an interesting path toward understanding how we know w hat is interesting.

Notes 1. By the 2004 publication of James Phelan’s Living to Tell About It, there were at least six distinguishable kinds of unreliability, but the discussion is far from over. 2. See John Tooby and Leda Cosmides and my answer to them in SubStance. See also Suzanne Keen on what novels can do.

Works cited Austin, John L. How to Do Things with Words. Oxford: Oxford UP, 1962. Booth, Wayne C. The Rhetoric of Fiction. Chicago: U of Chicago P, 1961. Fish, Stanley. Is There a Text in This Class? The Authority of Interpretive Communi ties. Cambridge: Harvard UP, 1980. Guardian Unlimited 3 Sept. 2007. . Iser, Wolfgang. The Act of Reading: A Theory ofAesthetic Response. Baltimore: Johns Hopkins UP, 1978. Keen, Suzanne. Empathy and the Novel. Oxford: Oxford UP, 2007. Phelan, James. Living to Tell About It. Ithaca: Cornell UP, 2004. Pratt, Mary Louise. Toward a Speech Act Theory of Literary Discourse. Blooming ton: Indiana UP, 1977. Spolsky, Ellen. “Why and How to Take the Fruit and Leave the Chaff.” SubStance 30.1-2 (2001): 178-98. Tooby, John, and Leda Cosmides. “Does Beauty Build Adapted Minds? Toward an Evolutionary Theory of Aesthetics, Fiction, and the Arts.” SubStance 30.1-2 (2001): 6-27.
