In May I was invited to write an essay on the short story. I said yes and then asked myself what I know about the genre. I made my most serious inquiries as an undergrad, when I was an aspiring short story writer myself. I read almost all of Hemingway, Kafka, and Flannery O’Connor. I read Chekhov and Barthelme and Coover. I read Rick Moody and George Saunders and David Foster Wallace. “Octet” seemed like the pinnacle of something. Like every other serious writer I knew I had my head blown off by Jesus’ Son. A good list with a predictable selection bias. Later I read James Baldwin and Samuel Delany and Junot Díaz and Mary Gaitskill and Lorrie Moore and Alice Munro and Ottessa Moshfegh and Leslie Marmon Silko and Isaac Bashevis Singer and John Edgar Wideman.
My knowledge felt piecemeal. I didn’t know what it added up to. I didn’t know how these writers fit into traditions of the short story. I didn’t know who they thought of as their interlocutors, who they published with, who I was missing. So I thought I might gather some data.
I used Claude Code to build a spreadsheet with every story (with its date, its author, their gender, their MFA if they have one) in Granta, Kenyon Review, Missouri Review, Ploughshares. That worked. I gathered more. I got The Drift and Ecotone and Hobart and n+1 and Wigleaf. I got Argosy and Cosmopolitan and The Saturday Evening Post. I now have 191 magazines, with more than 57,000 writers and more than 264,000 stories.
What use is this? I can do a lot with it. I can run social network analysis to see which writers tend to cluster among the same magazines, giving me an overview of literary communities at various scales. I can see which MFA programs have special relationships with which journals. I can see the gender breakdown over time of the contributors to each journal. Just gathering and getting an overview of the data is illuminating: I had no idea that the early twentieth century pulps were so vast. There is a pulp writer in this admittedly very incomplete dataset who published six hundred and twelve stories, another who published seven hundred and fifty-five, another—god bless H. Bedford-Jones, King of the Pulps—who published eight hundred.
I can also build (heuristic! hyperfallible!) power rankings to see who my education foregrounded and who I’ve overlooked. I build these rankings by putting the story data in conversation with anthologization: Best American Short Stories, O. Henry, Pushcart; and with credentials data I have: who attended MacDowell, won a Guggenheim, a MacArthur, an National Endowment for the Arts grant, a Nobel, a PEN/Malamud, a Rona Jaffe, won or was a finalist for an National Book Award, an National Book Critics Circle Award, a Pulitzer, a Story Prize. I cobbled together a formula that grants fractions of points for each story, a point for each venue, two points for each anthology appearance, one point for NEA or MacDowell, three for a Nobel, and two for the other credentials.
Some of this data comes from the amazing Post45 Data Collective (okay I cofounded it, used to edit it, and sit on the editorial board), which peer reviews and publishes literary and cultural datasets open access: the NEA data comes from there, published by Alexander Manshel, with his team. Major prize data, including the NBA and Pulitzer, came from Claire Grossman, Juliana Spahr, and Stephanie Young’s dataset. I also use Jordan Pruett’s NYT Hardcover Bestseller Fiction dataset. Nora Shaalan helped with the New Yorker data.
I share several variations on the rankings below. There’s the full rankings, which include stories dating as far back as 1841 and as recent as this year. To get a clearer sense of the contemporary short story, I also am sharing a version with only writers who published their first story (in this dataset, not necessarily in their life) after 1980 and another for those who published their first after 2000. I’m cutting these off at the top 1000 for now.
Two things. (1) It’s clear that Joyce Carol Oates is the GOAT. She blows the rest out of the water, even Munro and Updike. I did not know her lore. (2) HUGE CAVEAT!! Before you get mad, remember: this dataset is partial; it is missing a lot; it is missing almost every publication outside the United States—I’m trained as an Americanist and, by training, profoundly provincial; it is missing many years from O. Henry and Pushcart, which I plan to get; it is missing big chunks of some magazines, such as Epoch, Prairie Schooner, and TriQuarterly; it is missing magazines I couldn’t get because of archival challenges and some of these are very important, like Fence, Grand Street, Iowa Review, Michigan Quarterly, and, maybe most of all, Tin House, not to mention the very partial representation of the pre–WWII scene, here. So writers who appeared often in Tin House, say, like Charles Baxter, Stuart Dybek, Deborah Eisenberg, and Joy Williams, should be ranked even higher than they already are. It is, again, a fallible heuristic. But it also has surfaced writers I had, indeed, overlooked. It turns out that Peter Orner and George Singleton and Laura van den Berg are among the most recognized short story talents of our time. I did not know.
Finally: this is a work-in-progress and I am all ears for feedback. I’m still early in this work. I believe it’s valuable to share my work as I go, to make what I’m doing an occasion for public conversation. If you have magazines, anthologies, or credentials you think I should get, let me know. Even better: if you have data you want to share, let me know. If you find mistakes, let me know, if you have tips you feel like sharing about how to make this better, let me know. If you think my ranking algorithm ranks some people too high or some people too low, let me know. I plan on eventually making all the data public, probably next year. I’d like this to be widely useful and to be as robust as it can be. Maybe later I’ll post some other analyses. But for now…
Just one other comment: most highly prolific authors (e.g., E. Phillips Oppenheim, Mary Roberts Rinehart) rely on formulae — plot patterns, stereotypes, devices — that make much of their output predictable, if comfortable. But Joyce Carol Oates is exceptional in being both incredibly prolific and continuously experimental. Is there such thing as a typical Oates story or novel? Godspeed to the academic or biographer who undertakes a comprehensive study of her work — a project of several decades, at least.
Of course this project as it currently stands is canonically fundamentally circular. With no direct evaluation of literary quality. I’d suggest that this project would do well to reach as far outside canonical circles as possible. That means, among other things, getting the stories from American socialist and other left counterestablishment magazines and anthologies going back through time, and into the Harlem Renaissance periodicals and anthologies, and the Black periodicals going back into the 19th century. And beyond that there are all kinds of interesting and non-canonical self-published work and anthologies, DIY and otherwise.
Of course in doing this you would still not be doing qualitative literary analysis but quantitative and sociological work. And if you don’t get out of canonical circles, it will remain ideologically and otherwise very limited compared to what it would be otherwise. Getting out of canonical circles should grant you greater possibility to make more and greater unexpected discoveries, possibly even those that are genuinely breakthrough and challenging to what is known or believed to be known, something beyond canonical and commercial accounting.
It’s one thing to take all such material and analyze canonical and commercial American literature of the short story form, which has been done all but to death in various ways. It’s quite another project entirely to gather and analyze American literature in the short story form writ large.
Your current task may be less daunting than it appears because a statistical sampling of canonical literature is likely to reveal the essence of it, due to the kind of cloning and censoring features and effects of commerce and the canon. Your research effort might be better spent expanding into non-canonical, non-cloned material, which may be more likely to contain the unexpected, the high-variance, the different, the information-rich, more normatively telling or imperative cases. The canonical portion, being repetitive, is likely to have diminishing returns past a sampling threshold.
Go to the next level and the project could become very daunting, far moreso, I think, but also far more revealing. So much of the canon and commerce consists of cloning, essentially. Get out of the world of the clones and a whole other universe opens up, but whether that can be quantitatively accounted for rather than qualitatively analyzed, I’m skeptical. Still, I think that’s also where the sociology and ideology becomes more complex and interesting.
And of course so much of this is an accounting of publishers and editors, their control, whereas the writers are often the mere aftereffects of publishing and editing ideology, decisions, and control. The writers are the shadow cast by commercial and canonical publishing and editing, the establishment. Editors, publishers, and writers who exist outside of the dominant structures have the potential to be the most illuminating, complicating, and challenging figures. And writers who exist outside of all editing and publishing control have the potential to be more illuminating, complicating, and challenging still.
But if you’re intent on mapping the entire network of influence and gatekeeping, within establishment circles, and tracking any changes over time, then you probably need all the canonical and commercial data rather than just a sampling. Still if you don’t get outside canonical and commercial realms, and thoroughly so, then you’re looking at a very distorted view of American literature, no matter how exhaustively perused.