NEUROSCIENCE AND NARRATIVE theory have much to learn from each other. There is an explanatory gap between the neural correlates of consciousness and the lived experience of telling and following stories, but as I have tried to show, this gap can be a resource rather than an obstacle. The challenge, no doubt easier said than done, is to understand and take advantage of how the differences that divide these two fields provide complementary perspectives that are irreducible but (at least potentially) mutually illuminating. The problem is not how to reconcile the sciences and the humanities in a grand epistemological synthesis (E. O. Wilson’s [1998] dream of consilience) or how to resolve the conflict between the so-called two cultures. What is required, rather, is to figure out how to conduct productive exchanges across this divide from positions of disciplinary strength, where each side recognizes what the other can do that its theories and methods cannot. Identifying places where such exchanges might occur and suggesting what they might look like has been one of this book’s primary aims.
What might narrative theory learn from such an exchange? For one thing, as I have argued, the findings of contemporary neuroscience about cognition, language, and emotion make some literary and narratological theories less plausible than others (especially those still haunted by the legacy of structuralist narratology). I have also tried to show that some neuroscientific research (on the disjunctive temporality of brain functioning, for example, or on the action-perception circuit) can shed light on much-discussed topics in narrative theory (how stories organize time and how plots imitate actions).
The benefits are not all one way, however. As I have argued, neuroscientific investigations of simulation in cognition, metaphor, and social interactions need to be informed by literary and narrative theories that have analyzed with great precision and detail the paradoxical workings of as-relations and the as-if in fictional stagings of experience. These theories, drawn from the works of Nietzsche, Iser, and Ricoeur (among others), show why the as-structure of simulation is more variable and open ended than causal or mechanical models imply, which explains in turn why the grounding of metaphors in embodied experience is differentially graded on a spectrum from the concrete to the abstract. Hence the biocultural variability of embodied metaphors for experiences like anger or pain that are both universal across our species and culturally and historically differentiated. Similarly, as phenomenological theories of reading suggest, the doubling of self and other in our experience of the intentionality embedded in a text casts doubt on the unidirectional assumptions of simulation models that underlie simplistic claims about the ability of literary works to improve theory of mind and empathy. Understanding the relation between stories and the brain requires competence in narrative theory as well as in neuroscience. Neither can do this work alone.
If one of the long-standing goals of narratology has been to explain the triangular relationship between stories, language, and the mind—to understand how our ability to tell and follow stories correlates with basic linguistic and cognitive processes characteristic of our species—then it is incumbent on narrative theory to keep up with the science. It may have once been the case that so called first-generation cognitive science, based on an artificial intelligence model of the brain, lined up well with the assumptions of structuralist narratology. The logical, algorithmic relations between anatomically based modules of a computer-like brain might then correspond to the rules of operation of universal grammar or to the orderly structures of language (langue) underlying speech (parole), and it would have made sense for narratology to propose cognitive schemes, frames, or scripts analogous to these modules and preference rules defining the formulas governing their operation.
As I explain in chapter 1, however, this model of the brain has fallen out of favor as the science has evolved. Extensive and ever-growing evidence provided by new scanning technologies has revealed that a modular, computer-like conception of the brain is too rigid and orderly to account for the dynamic, recursive interactions across the cortex and between brain, body, and the world through which cognitive patterns form, dissolve, and re-form. Language is a whole-brain phenomenon (Huth et al. 2016), and different parts of the cortex can be repurposed to support various and variable cognitive functions (Anderson 2010), as when the brain learns to read by recycling parts of the visual cortex that had evolved for invariant object recognition (Dehaene 2009). Our capacity for language is based on recurring, interactive experiences, not innate grammar-like structures (Nadeau 2012), and emotions are similarly an anatomically variable function of personal and cultural experiences, not universal structures with defined “neural signatures” (Lindquist et al. 2012, Barrett 2017). A bushy conception of the brain as a dynamic assembly of interacting, mutually constituting elements has displaced the logical, orderly modular theory, and cognitive narratology needs to revise its research program accordingly.
Science is a historical enterprise, not simply a collection of facts, and its theories change as new evidence casts doubt on previously accepted models. Fields in the humanities like narratology that make assumptions about cognition and the body need to continually review and evaluate those assumptions in light of developments in the scientific research. Neuroscience will not settle all of our controversies, however. There is no danger that the scourge of scientific positivism will close off inquiry and debate. To begin with, falsification is asymmetrical. Science can invalidate some views that are inconsistent with well-documented findings, but it can’t necessarily tell us what is right. Some theories turn out to be wrong, but much room remains for productive disagreement about how to proceed. Although theories of language and emotion based on a modular theory of the brain don’t stand up well against growing bodies of evidence, many questions remain about how best to understand these aspects of our cognitive lives, and conflicting interpretive approaches are still viable. Just as there are multiple ways in which our evolutionarily developed cognitive equipment can be used to create different narrative and cultural worlds, so different hermeneutic programs for making sense of this pattern-forming activity can be productively pursued—but there are constraints on what works and what doesn’t (see Armstrong 1990).
A dynamic, recursive model of the bushy brain is consistent with a variety of different ways of analyzing the interaction of brain, body, and world in the exchange of stories. Drawing on the work of Ricoeur and Iser, I have preferred a phenomenological model of the circuit of figuration that crosses back and forth between different aspects of experience as we exchange stories—how the configuration of action in narration and emplotment (mimesis2) draws on prefigurative personal and cultural experience (mimesis1) and can in turn transform the figurative cognitive processes of recipients (mimesis3) (see chapter 1). But as I have also suggested, other pragmatically-oriented approaches in second-generation cognitive narratology offer promising insights, consistent with contemporary neuroscience, into the dynamic interactions through which stories are constructed and exchanged. For example, Terence Cave (2016) combines the theories of implication from relevance theory with Gibson’s concept of affordances to propose a cognitive model of thinking with literature that emphasizes the improvisatory possibilities made available by narrative forms. Elaine Auyoung (2018) applies the concept of fluency from communication theory to clarify how mimetic fiction can come to seem real. Karin Kukkonen (2016) uses Bayesian models of predictive processing to explain how the probability design in a narrative plays with the expectations that guide our cognitive activity in reading and in life. James Phelan’s (2017) rhetorical theory has much to say about the ways in which authors seek to influence their audiences by manipulating various aspects of narrative communication. This is not an exhaustive list but, I think, an indication of some of the many lines of inquiry that make this new period in the history of narrative theory exciting and promising.1
Some terms and concepts from classical structural narratology still provide useful tools for analyzing the circuit of figuration through which stories are produced and exchanged. For example, I have repeatedly invoked the foundational narratological distinction between story and discourse (Chatman 1978) in order to explore the interaction between the different temporal processes and modes of action that characterize the telling (the act of narration), the told (the organization of events), and their reception (the activities of narrative comprehension). Similarly, Gérard Genette’s (1980) magisterial explanation of narrative anachronies lays bare disjunctions in the temporal organization of stories that are interestingly correlated to the peculiarities of our temporally decentered brains. These terms and concepts often need correction or supplementation, however, because a grammatically-oriented model lacks the ability to clarify and explain the epistemological and cognitive workings of narrative—for example, as I show in chapter 4, a linguistic model is inadequate to explain how the much-discussed ambiguities of free indirect discourse play with the experiential paradoxes of intersubjectivity or how stories from different historical periods configure the as-structures of cognition to stage what it is like to be conscious through different versions of the as-if. The lexicon of classical narratology, although sometimes dauntingly complex and obscure, can provide invaluable instruments for making sense of the processes through which narratives build and break patterns. What is wrong is to reify or ossify these terms by ontologizing them or freezing them into a taxonomy.
The drive to classify and taxonomize, a pervasive characteristic of narratology that goes back to Aristotle, is both a good and bad thing. Categorization is a basic cognitive activity at work across the different sensory processes of seeing-as through which our embodied brains play with patterns to configure our worlds. From opponency in vision to the exploratory activity of different sensory and perceptual processes, our responsiveness to differences is basic equipment for living. The error is to reify these categorizing and differentiating processes into cognitive modules or formal structures that provide more logic and order than is evident in the bushy brain’s ever-shifting balancing act between building and breaking patterns. Modes of categorization do crucial cognitive work, and the terms and concepts of classical narratology can provide tools for understanding these processes, but a classificatory scheme cannot do justice to the to-and-fro between the competing imperatives of order and flexibility to which narratives contribute by playing with concord and discord.
Stories offer a potentially productive resource for exploring many of the issues that have recently emerged as central to the neuroscience of the bushy brain—for example, the processes of neuronal assembly, oscillatory binding, and brain-to-brain coupling that neuroscientists have come to recognize as crucial to the workings of the dynamic system of intracortical and brain-body-world interactions that constitute the so-called connectome (see Raichle 2011, Dehaene 2014). Because of its reliance on an outmoded conception of brain structure and cortical modularity, formalist cognitive narratology is not well equipped to provide the theoretical guidance this research needs, but one of my hopes is that neuroscientists might find useful suggestions about possible research questions and areas to explore in the correlations I lay out between narrative theories of time, action, and intersubjectivity and the corresponding cognitive science. The instruments available to neuroscience are not always adequate to these challenges. For example, as I have explained, the temporality and sociality of narrative interactions are especially recalcitrant to measurement by current scanning technologies. But often these limitations point to vital areas of brain functioning that neuroscientists increasingly recognize the need to account for—how timing is crucial to the oscillatory processes of neuronal synchronization, for example, and how coupled brains in collaborative interactions behave differently from brains in isolation. What stories do that neuroscience has a hard time measuring is often exactly what neuroscience needs to understand.
It is also important to keep in mind what literature can reveal about cognition that necessarily eludes the measuring instruments of science—the qualia of what it is like to be conscious, for example, that fictional narratives can stage in different ways beyond what the quantifications of empirical methods can register. It is essential as well that cognitive scientists not ignore or run roughshod over these limitations by oversimplifying matters whose complexities narrative theory can and should instruct them about—for example, how fictional stagings of other worlds double self and other in ways that are not reducible to unilinear simulations and that consequently may not necessarily produce prosocial effects. (For the same reasons, predictions about the necessarily antisocial effects of violent films or video games are also overly simplistic and deterministic.) One of the hallmarks of science is its inventiveness in designing experiments that circumvent the limitations of its measuring instruments, but narrative theory can help differentiate between illuminating creativity and blunt-edged obfuscation (as in experiments that invoke vague notions of literary fiction to test equally vague responses to theory of mind measures, neither of which do justice to the stylistic complications of narratives or to the variability of the reading process).
Some practitioners of cognitive literary studies have recently called for an empirical turn to test theoretical claims about reading and narrative against experimental evidence (for the best work of this kind, see Scientific Study of Literature, founded in 2011 as the journal of the International Society for the Empirical Study of Literature). This is in many respects a welcome development. Ascertaining what the facts are is not a simple and straightforward matter, however. The empirical turn sometimes seems motivated by an aversion to theoretical speculation, but experimental work can’t do without theory. Experiments test hypotheses, and these hypotheses need to be based on good science and also on good theories about narrative and reading. Rather than regarding experimental work as an alternative to literary and phenomenological models of the reading process, empirical studies should look to narrative theory and phenomenology as a source of hypotheses to test and as frameworks for making sense of their findings. I have cited some important work in this field, by researchers like Richard Gerrig (2010, 2012) and David Miall (2011), because their experiments shed interesting light on important scientific and theoretical questions about memory-based processing and the role of emotions in guiding expectations. Empirical work is a process not of collecting facts but of designing informative experiments, and good experimental design requires an understanding of the issues deserving exploration, based on the best-regarded science and the relevant literary and narrative theory, and an appreciation of the limits as well as the powers of various methods for gathering evidence. Testing hypotheses is a theoretical enterprise, not an alternative to theory.
Experimental work on narrative and reading also needs to keep in mind how readers can make sense of stories differently not only because of their different levels of competence but also because of their different interpretive frameworks. Skepticism about a so-called WEIRD subject pool (Henrich, Heine, and Norenzayan 2010)—whether undergraduate students from Western, educated, industrial, rich, and developed countries are a neutral, unbiased source of evidence about cognitive processes—is also relevant to neuroscientific and psychological experiments about stories and reading. As biocultural hybrids, members of any subject pool will have some universal and some contingent, historically and culturally variable characteristics. In one sense the ability to tell and follow stories is natural to our species—even small children can do it! But it is also a cultural acquisition, something we learn to do, and something we can learn to do better (or so I as a teacher of literature believe).
Any description of how we tell and follow stories is likely to have a prescriptive dimension. For example, Iser’s phenomenological description of the reading process also offers guidance about how to better understand stories (by attending more thoughtfully to how we fill gaps and indeterminacies, how we build and break illusions, how we project expectations that are in turn surprised, or how we double our consciousness with the consciousness held ready by the text). One of the ways in which we learn to read stories is by following the guidance of others who we think have something to teach us, and narrative theories at their best may help us become better readers of narrative. What counts as better is not measurable on a linear and univocal scale, however, that leads teleologically from bad to better to best. There are degrees of competence in reading, but at some point equally competent readers may begin to differentiate themselves by adopting opposing, sometimes mutually exclusive, hermeneutic allegiances and acquiring different reading practices, which in turn may reflect the contingencies of their cultural, historical situations.
One of the paradoxes of our biocultural hybridity as readers is that our universally shared cognitive equipment can be deployed in a variety of not always reconcilable ways, guided by different presuppositions and interests, so that equally valid interpretations of a text can disagree fundamentally about its meaning (see Armstrong 1990). Empirical studies of literature must remember that opposed reading practices may divide even the most competent readers because they belong to different interpretive communities, with different values and aims. These differences are not something that an empirical approach can or should hope to overcome by digging down to the simple facts of the matter. They are, rather, fundamental aspects of reading and interpretation that a scientific study of literature needs to understand—and that theoretical accounts of hermeneutic pluralism and its epistemological limits can help to explain.
Empirically tested scientific theory will not by itself tell you how to read, although knowing something about the cognitive processes that narratives draw on and set in motion may heighten our appreciation of the ways in which stories play with pattern. This is one of the limits to what science can tell the humanities and a reason not to fear its determinism. Interpretation is a circular process of projecting hypotheses about how parts fit together into wholes and adjusting these hypotheses as anomalies and inconsistencies emerge (see Armstrong 2013, 54–90). These guesses and the revisions we must make to them as we read and interpret are an open-ended, to-and-fro, interactive process that cannot be reduced to deterministic procedures or mechanical formulas. Our hermeneutic guesswork can be guided, however, by various models that offer suggestions about hypotheses to test, and that is a reason why narrative theories of whatever stripe (classical or postclassical, cognitive or linguistic, guided by formal interests or informed by concerns about gender, sexuality, or politics) can be sources of productive, interesting interpretations and, by inspiring creative guesses, can make us more insightful readers. For example, Genette’s elaborate categorization of narrative temporality, voice, and focalization by itself won’t make us more competent interpreters of stories, but his terms and concepts have unquestionably been a rich source of suggestions for how to understand narratives of all kinds.
Genette’s work is also a good example of another question that scientific approaches to narrative need to deal with: what should be the database? Can an adequately comprehensive theory of narrative be constructed on the basis of a single work, even one as complex and ambitious as Proust’s In Search of Lost Time? No theorist of narrative knows all of the world’s literatures, from all periods and cultures. There have been, of course, many admirable, courageous attempts at encyclopedic coverage (for a recent example see Pettersson 2018), and one reason why Scholes and Kellogg’s Nature of Narrative (2006 [1966]) continues to illuminate is the rare and productive partnership of this modernist and medievalist with complementary expertise. The examples I have offered in this book reflect my own limitations from a lifetime of work teaching and writing about a particular body of narratives (including a studies in fiction course, with examples from Balzac to Pynchon, that Bob Kellogg first suggested I teach when he was my chair at the University of Virginia, and that evolved into a seminar on realism and modernism when I was Bob Scholes’ colleague at Brown). The narratives I know best—especially the works of literary impressionists James, Conrad, and Ford and modernists Joyce, Woolf, and Faulkner—are by no means representative of all stories, but as I have tried to show, their self-consciousness about the epistemological processes set in motion by representation and narration makes them particularly useful cases for foregrounding the cognitive workings of narrative figuration of interest to me. Any examples offered by a narrative theory will reflect the cultural, historical contingencies of their time and place—their situation in the cognitive archive—and this is a necessary consequence of our biocultural hybridity. But this hybridity should also make it possible to extrapolate from the particular to the general and to identify species-wide processes and properties deployed by any given example from this archive. That is my hope, in any case, and it is a wager that all students of narrative must make—and that an empirical, scientific approach cannot avoid.
Science also cannot dictate or determine the moral and pedagogical implications of narrative. The play of doubling in the exchange of stories and the variability of the as and the as-if in narrative figurations of experience resist reduction to mechanical, univocal formulations of the social powers of narrative, whether Steven Pinker’s overly optimistic claims for the moral education of epistolary novels or Mark Bracher’s well-intentioned but authoritarian, monolithic pedagogy of reverse engineering through habitual reinforcement of preferred cognitive categories. Stories can have noxious as well as beneficial moral consequences, fueling conflict and violence or promoting empathy and compassion. Narratives can reinforce boundaries between us and them, or they can challenge prevailing conceptions of justice and injustice and encourage democratic conversation about reconfiguring our sense of responsibility toward others. This book tries to explain this contradictory state of affairs in scientific as well as humanistic terms. The three-fold exchange of figurative patterns (the pre-, con-, and refiguration cycle) as we tell and follow stories is grounded in various neurobiological processes and the interactions they enable and constrain between our brains, bodies, and worlds. This cycle of figuration is open ended and unpredictable, however, which is a good thing for many reasons having to do not only with our neurobiological health (preventing cognitive patterns from getting locked into rigid habits) but also with the creative potentialities of our existential and cultural lives. Neuroscientifically informed analyses of the figurative processes involved in narrative can help explain why this is so, but they do not let us off the hook from making difficult moral and social choices.
Neither neuroscience nor narrative theory can claim to have the last word about stories. Narratives are crucial to the life of our species, and it is unlikely that such a complex phenomenon would yield its secrets to any one approach. Active conversation between neuroscience and narrative theory across the explanatory gap about how we tell and follow stories is yet another chapter in a long history of reflection and debate about this aspect of our cognitive lives. Turning to science for answers to questions long debated in narrative theory will not end these arguments. Plenty will remain to disagree about, but those disputes are likely to be more rigorous and productive if neuroscience is part of the conversation.