Journal ISSN: 2371-4549
An Introduction to a new Cluster on Genre
by the NovelTM Research Group
Genre is a word whose time has come — and gone — and might now, perhaps, be coming back again. Debates about particular literary kinds have been common in literary criticism since Aristotle's Poetics, but they acquired a new intensity and reflexivity in the third quarter of the twentieth century, as structuralists and post-structuralists struggled to redefine the concept of genre itself. From Northrop Frye's Anatomy of Criticism (1957) through Jacques Derrida's "Law of Genre" (1980), genre theory gave scholars a way to connect literary works to durable cultural patterns — or challenge the possibility of that connection.1
As Robyn Warhol points out in her introduction to The Work of Genre (2011), the concept of genre has spent much of the last thirty years in eclipse. This is not to say that readers — or publishers — ever stopped dividing literary works into epics, lyrics, mysteries, and young adult dystopias. Readers can hardly open a novel without thinking about genre, consciously or unconsciously. As John Frow explains, "generic structure both enables and restricts meaning, and is a basic condition for meaning to take place."2 But genre has become the kind of critical term that scholars use casually, and with an implicit shrug. Several observers have traced this shift to the ascendance of historicism in literary study. It began to seem "downright wrong," Warhol suggests, "to detach form, structure, and theme from a text's specific historical moment, or to sketch out taxonomies of literary types in the abstract."3
One can of course reconcile genre theory with historicism by treating genre as a historical construct. Genres can be viewed, not as natural literary kinds, but as generalizations about the organization of literary production or reception inferred from evidence in a particular period. Today, some version of this theory is often taken as an implicit starting-point for conversation. But it is a theory so flexible that it provides little guidance on many questions, not least the question of how genres resonate across historical epochs, of why, as Rita Felski puts it, "past texts still matter and how they speak to us now."4 As John Rieder has remarked, theories that identify genre with historical practices of categorization can lead to "a kind of tautology, an assertion that the genre is whatever the various discursive agents involved in its production, distribution, and reception say it is."5 That premise will not necessarily tell literary historians how to proceed when different observers, or periods, disagree. Nor can it adequately account for the flash of recognition often experienced by both lay and scholarly readers when encountering works written centuries ago.
For instance, if genres emerge from a period's own practices of categorization, why are so many books written about nineteenth century "science fiction"? We inherit the term from Hugo Gernsback in 1929; a strictly historical theory of genre might not use it to describe literary movements before that date. Should historians let Mary Shelley and Jules Verne go their separate ways, as nineteenth-century readers mostly did? Or can Gernsback's term be allowed to retrospectively define a previously nameless set of similarities that binds the pulps of the 1920s to earlier traditions of lost-world romance, dream vision, and utopia? Contemporary genre theory doesn't provide an answer, or even a clear criterion that we could use to find one. So critics confronted with a dilemma like this often find it simpler just to keep writing about nineteenth-century "science fiction," while adding air quotes to the phrase. Genres are acknowledged to be "fluid and tenuous constructions"6 at the same time as they remain rhetorically indispensable; the word genre itself sometimes seems to have acquired air quotes and a shrug.
And yet, there has been a resurgence of critical writing about genre in the last ten years. The reasons may have as much to do with popular culture as they do with new critical theories or methods. The things that used to be called "genre fiction" are newly central to cultural production. Warring comic-book franchises have acquired enormous market share, while cable channels are subdividing into long tails capable of targeting every possible variant of post-apocalyptic teen romance and self-conscious vampire comedy. It is perhaps not surprising that science fiction and fantasy have become as important to genre theory as "the epic" used to be, or that the concept of the "microgenre" is one of the salient critical innovations of the twenty-teens.7
Computational methods of analysis have arrived in the humanities just in time to take advantage of this proliferation of categories. When Netflix created 77,000 new microgenres —including "Emotional Fight-the-System Documentaries" and "Foreign Satanic Stories from the 1980s" — it was almost inevitable that Alexis Madrigal would try to reverse-engineer the process that constructed them.8 Computational methods of clustering (and data provided by Spotify) have similarly encouraged Glenn McDonald to visualize musical microgenres in Every Noise at Once.
Emerging work in the field of computational literary criticism has similarly begun to tackle the question of genre and the social organization of texts. Researchers have shown that computers can reproduce and predict human judgments about categories of fiction9 as well as model market-based decision-making.10 But we are still a long way from having more generalizable things to say about the novel as the so-called "genre of genres," as one of the more complex textual forms that can be stitched together from other forms to such a degree that it has "the appearance of being merely a secondary syncretic unification of other seemingly primary verbal genres."11
This question of the novel's coherence, of the ways in which its grouping organizes its larger social significance as a cultural artifact, was the animating idea behind the first year of research of the multi-year project on Text Mining the Novel. The essays that resulted take many different approaches to the problem. Some of them use models of genre in a predictive way, to explore the boundaries of genres or compare them to each other. Others extract explanatory insights from computational models, in order to better understand the role of epistemology or gender in fiction — or the nature of fiction itself. Still others attempt topic modeling or unsupervised clustering, in order to detect literary genres, or genre-like patterns, that may have gone unrecognized in critical debate.
In each case, common assumptions about the history of the novel are put under pressure through the insights offered by a computational understanding of the problem. Whether it is the question of the transhistorical coherence of genres like science fiction or detective fiction (Underwood); the surprising stylistic unity of Dead White Men that could be considered a genre in its own right (Wilkens); the eighteenth-century prehistory of the novel as captured through a variety of market-driven labels such as the tale, romance, or history (Algee-Hewitt et al); the detection of the philosophically oriented novel that queries the novel's epistemological foundations (Erlin); the relationship between gender, character and genre in the classical realist novel of the nineteenth century (Jockers/Kirilloff); the grounding of the novel's fictionality in a phenomenological investment in human encounter and doubt (Piper); or finally, an interrogation of the ways in which gender and genre theory intersect (Mandell), each of the pieces in this cluster offer new arguments about the novel's history and the place of categorization in that history.
It would be a mistake to assume that these authors have resorted to numbers in an attempt to make the concept of genre more "objective." Numbers have many uses: they can help a scholar compare large groups of examples, for instance, or describe a continuous gradient. But they aren't inherently objective, any more than words are. On the contrary, numbers will be used here to construct models, confessedly imperfect representations of cultural practices. As Andrew Piper has mentioned in his introduction to the first issue of Cultural Analytics, this modeling process is a way to acknowledge assumptions and reflect self-consciously on "the contingencies of knowledge."12 Indeed, because computational techniques frequently generate continuous probability distributions or make use of multiple clustering algorithms, they can be particularly effective in illuminating the fluidity of genre, thus supporting an understanding of genre in terms of family resemblances rather than fixed taxonomies.13
But if objectivity is the wrong word, it may still be fair to say that these essays are, collectively, searching for a way to put literary conversation about genre on a firmer footing than it presently occupies. The timeless taxonomy of modes and genres that Northrop Frye once hoped to sketch is not likely to make a comeback. Genres are clearly historical constructions. But our current repertoire of critical strategies doesn't always help us untangle their histories. Instead, the premise that a genre is whatever historical actors say it is tends to dissolve the history of genre into a maze of tautologically valid (but conflicting) assertions made by "writers, producers, distributors, marketers, readers, fans, critics and other discursive agents."14
Although genres are perspectival categories, we can aim for something better than a mere transcription of participants' opinions. Even -- and perhaps especially! -- historicist critics need to be skeptical about the testimony of contemporary observers. Writers deeply immersed in a particular system of production may pass silently over boundaries that seem self-evident. Conversely, an observer with a personal stake in recent debates may exaggerate the importance of relatively slight formal differences. To think skeptically about historical testimony, even scholars who understand genres as historical institutions will need another point of comparison.
Text analysis may not provide a complete picture of the history of genre, but it does give us a second reference point — a touchstone that we can use to unpack historical assertions about genre, or compare them to each other. This kind of evidence could make conversation about genre more interesting than it has been for the last thirty years, simply by giving historicism something to push against. In fact, it's possible to combine historical testimony and text in a single model. Instead of relying strictly on formal templates, or strictly on the conflicting claims of historical actors, we can work in a more dynamic way — putting textual and historical evidence in dialogue with each other, and allowing new insights to emerge from the tension between them. That is what we hope to have fostered here: a mode of inquiry about genre that revives substantive debate, and avoids falling back on a nominalistic shrug.
- Northrop Frye, Anatomy of Criticism (Princeton: Princeton University Press, 1957); Jacques Derrida and Avital Ronell, "The Law of Genre," Critical Inquiry 7, no. 1 (1980): 55-81. [↩]
- John Frow, Genre, 2nd. Ed. (London: Routledge, 2015), 10. [↩]
- Robyn Warhol, "Genre Regenerated," In The Work of Genre: Selected Essays from the English Institute, ed. Robyn Warhol (Cambridge, MA: English Institute in Collaboration with ACLS, 2011). [↩]
- Rita Felski, "Context Stinks," New Literary History 42, no. 4 (2011): 577. [↩]
- John Rieder, "On Defining SF, or Not: Genre Theory, SF, and History," Science Fiction Studies 37, no. 2 (2010): 191. [↩]
- Sherryl Vint and Mark Bould, "There Is No Such Thing As Science Fiction," In Reading Science Fiction, ed. James Gunn, Marleen Barr, and Matthew Candelaria (New York: Palgrave, 2009), 48. [↩]
- For reflection on the current spurt of generic hybridity, see Gary Wolfe, Evaporating Genres: Essays on Fantastic Literature (Middletown: Wesleyan University Press, 2011). Molly O'Donnell and Anne H. Stevens are currently organizing a Big Book of Microgenres. [↩]
- Alexis C. Madrigal, "How Netflix Reverse-Engineered Hollywood," The Atlantic, January 2, 2014. [↩]
- Sarah Allison, Ryan Heuser, Matthew Jockers, Franco Moretti, Michael Witmore, "Quantitative Formalism: An Experiment," Stanford Literary Lab, https://litlab.stanford.edu/LiteraryLabPamphlet1.pdf; Matthew L. Jockers, Macroanalysis: Digital Methods and Literary History (Urbana: University of Illinois Press, 2013), 63-104; Mariona Coll Ardanuy and Caroline Sporleder, "Structure-based Clustering of Novels," In Proceedings of the 3rd Workshop on Computational Linguistics for Literature, CLfL (EACL 2014): 31-39; Paul Vierthaler, "Fiction and History: Polarity and Stylistic Gradience in Late Imperial Chinese Literature," CA: Journal of Cultural Analytics (2016): "Fiction and History: Polarity and Stylistic Gradience in Late Imperial Chinese Literature". [↩]
- Vikas Ganjigunte Ashok, Song Feng, and Yejin Choi, "Success with Style: Using Writing Style to Predict the Success of Novels," In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing (EMNLP 2013): 1753-1764; Andrew Piper and Eva Portelance, "How Cultural Capital Works: Prizewinning Novels, Bestsellers, and the Time of Reading," Post45 (2016) ; Jody Archer and Matthew Jockers, The Bestseller Code: Anatomy of the Blockbuster Novel (New York: St. Martin's, 2016). [↩]
- M. M. Bakhtin, The Dialogic Imagination, trans. Caryl Emerson and Michael Holquist, (Austin: University of Texas Press, 1981), 321. [↩]
- Andrew Piper, "There Will Be Numbers," Journal of Cultural Analytics, May 23, 2016. [↩]
- On genres as family resemblances, see Alastair Fowler, Kinds of Literature: An Introduction to the Theory of Genres and Modes (Cambridge, MA: Harvard University Press, 1982). [↩]
- Vint and Bould, "No Such Thing As Science Fiction," 48. [↩]