Reader Selection and Market Acceleration: Are We Living in a Backward World?

Given the insane length of this post, I would recommend downloading the PDF version. Besides, it’s easier to read the footnotes that way. Some of which are pretty fun, I think.

Much in the same way it’s impossible for me to choose a single part of Franco Moretti’s Distant Reading that I like the best, I can’t quite settle on what it is about Moretti and his approach to literary studies that gets me jazzed. He’s iconoclastic and disruptive (his initial paper on “distant reading” was intended to jar the comparative literature department at Columbia out of its rut and resulted in a number of aggressive critiques), he’s obsessed with data and quantitative analyses that are less about close reading and interpretation and more about asking larger questions that can be backed by data (which has a sort of kinship with baseball sabermetrics), and his writing is almost anti-academic in the way it conveys a sense of wonder and exploration. I’ve never met the man, but his writing is stimulating, fun, and varied.

Hitting a number of different topics, the ten essays in this collection touch upon, in no particular order: data analysis of the length, type, and structure of titles for novels in the eighteenth century; the spread of Hollywood movies; evolutionary theory as applied to the literary marketplace; a geography of literary history; the vast structural differences in the European and Chinese novelistic form, which developed in parallel; and how the tracing of a single textual element (clues) through the stories of Sir Arthur Conan Doyle and his rivals can explain why Sherlock Holmes survived and 99.5% of the other books published at that time have been eliminated in the “slaughterhouse of literature.”

For me, personally, I haven’t had this much fun reading a critical work of literary scholarship in ages. Maybe ever. And to go back to that sabermetric thing, the way Moretti’s mind works throughout this collection reminds me of the great baseball stats books—the Bill James Historical Baseball Abstract, obviously, and Baseball Between the Numbers—tend to function: they ask large questions rooted in comparison and evaluation, then turn to the analysis of large data sets (or large swaths of history) to deduce facts and frameworks about individual works or groups of works. Admittedly, I don’t read a ton of contemporary literary criticism, but this seems a long way away from the traditional analysis of meaning found in most monographs and is based in the practice of “close reading.” (Selected at random: Melville’s Vision of America, Moby-Dick and Calvinism, Twentieth-Century Interpretations of Moby-Dick, Moby-Dick a Hindu Avatar, or, simply, Meaning of Moby-Dick.) In the words of Moretti, “We do not need more interpretations [. . .] not because they have nothing to say, but because, by and large, they have already said what they had to. A lot of good work has been done on the relation between meaning and meaning; far too little on meaning and forces.”

If I had to point to one bit of Moretti’s book that captured my critical imagination though, it has to be the part about reader selection and market magnification. This comes up in two separate essays (“The Slaughterhouse of Literature” and “The End of the Beginning”), both related to an experimental data-centric analysis of Sherlock Holmes stories and other detective stories of the time.

In one of Moretti’s graduate seminars, he and his class chose a single “unit of analysis”—in this case, clues—to focus on while reading mystery stories, and saw how those played out in all the texts at hand. So, rather than doing a “close reading” of a ton of different mystery stories of the late 1800s, they read these pieces searching just for the presence or absence of clues. (Like looking through baseball stats for players with high strike-out to home-run ratios, and ignoring the rest of the data.) This lead to a tree-like diagram in which some stories showed the presence of clues, whereas others of the time did not. The stories with clues were then subdivided into ones with clues that were “necessary,” which were further subdivided into stories with “visible” and then “decodable” clues. Of all the twenty-one stories they looked at (twelve of which were Holmes stories), only four ended up with clues that were “necessary,” “visible,” and “decodable”—and all four were written by Doyle. (Worth noting that eight of his twelve stories didn’t meet this categorization. Moretti posits that this might be due to Doyle’s focus not on the mechanism that made his best stories—like “The Red-Headed League”—work so well, but on his desire to build up Holmes as an almost mythological detective. If all the clues were decodable and visible to the reader, just how good of a detective is this Sherlock?)

That’s interesting in itself, and Moretti develops this idea as a possible explanation for why Holmes stories survived instead of any of the alternatives. Even if readers—or Doyle himself—couldn’t explain why they were drawn to these sorts of stories, this crucial formal unit (“clues”) worked in their readerly brains and lead them to prefer these sorts of stories. Market forces lead to more Holmes-esque writing, Doyle’s books stay in print, and decades later, academics cotton on and Sherlock Holmes stories are ratified as the paragon of a certain type of detective literature.

All that’s fascinating and makes me want to draw all sorts of trees and charts in my world lit class, but here are few paragraphs from Distant Reading that I want to mention before trying to develop my own argument. The first is a quote Moretti uses from an article by the economists Arthur De Vany and W. David Walls on an economic model for the film industry:

Film audiences make hits or flops . . . not by revealing preferences they already have, but by discovering what they like. When they see a movie they like, they make a discovery and they tell their friends about it; reviewers do this too. This information is transmitted to other consumers and demand develops dynamically over time as the audience sequentially discovers and reveals its demand . . . A hit is generated by an information cascade . . . A flop is an information bandwagon too; in this case the cascade kills the film.

In other words, word-of-mouth generates hits, creates buzz or an “information cascade” that develops almost exponentially. This likely sounds familiar to anyone involved in book publishing. What I’m most interested in here—and want to apply to the current book market—are two processes that Moretti teases out: acceleration and selection. Moretti here:

As more readers select Conan Doyle over L. T. Meade and Grant Allen, more readers are likely to select Conan Doyle again in the future, until he ends up occupying 80, 90, 99.9 per cent of the market for nineteenth-century detective fiction.

And, Moretti again:

A few sellers for the entire market; just like Holmes for the mystery niche. But it’s important to disentangle the two discrete processes that converge onto this single outcome: the process that centres on readers, and on their selection of Doyle’s formal solution over those of his rivals; and the other one, in which the market amplifies that initial selection over and over again. Readers and markets, in other words, are both causal agents, but in different ways: in the sense that readers select, and then markets magnify.

This is where I want to turn from Moretti to the current literary marketplace, especially as it relates to translations and the creation of buzz and acceleration. My working observation (I hesitate to use the word hypothesis given the paucity of data that I have at hand) is that the acceleration of two related phenomenon—media and the ability to comment on media—has resulted in a situation in which the processes of “reader selection” and “market magnification” have become divorced from one another and are now essentially reversed.

In a million ways, this is an impossible topic to talk about and analyze. The word-of-mouth buzz that leads to market magnification is generated from a number of different sources (booksellers, professors, friends, celebrities, tastemakers) and is notoriously hard to pin down and run experiments on. So, to make this as concrete as possible, I’m going to focus on my favorite hobby-horse of hate: anticipatory lists.

I have a corollary idea that (hopefully) justifies yet another post about lists.¹ Over the past decade, the number of works of international fiction and poetry that you could consume as a reader—or, more pertinent to this post, review as a critic—have expanded greatly. In 2008, there were 361 different titles; in 2016, 602. That’s only 241, but how many critics read more than 100 books last year? Fifty? One hundred?

And that 100 books is taking into account all the possible books, of which, works of original English-language fiction and poetry take precedent to a degree that so far exceeds the number of works in translation published on an annual basis that a critic reading a work in translation is almost the same as random.² Which ones do they choose to read? Why?

Historically, advance review trade magazines—Publishers Weekly, Kirkus, and Library Journal, among others—provided a sort of roadmap of what to pay attention to and what could be ignored. An average issue provides an overview of 60-70 titles, a handful of which are “starred,” a similar handful of which are works in translation.³

It would be interesting to track this: How many “major” reviews do “starred” titles get versus those that are simply reviewed in a trade magazine, versus those unreviewed? How do sales compare across these categories? These are ideas I’m totally marking for future posts . . .

But more to the point, I’m not sure if trade magazines are the major gatekeepers critics pay attention to anymore.

This is the most easily criticized statement in my post so far, so have at it, knowing that I’ve anticipated all of your immediate objections and still decided to put this out there.⁴ The majority of book criticism being written today is operating in one of two spheres: the personal-professional blog (everything from Flavorwire to Quarterly Conversation to Words Without Borders to Asymptote to Tony’s Reading List to Complete Review to Arabic Literature (in English)) or the professional-legacy arena (which includes online journals like Salon or Slate which pay their writers, to Wall Street Journal, Washington Post, L.A. Times, NY Times, The Nation, and other entities that are more obviously corporate and concerned with revenue flows, be it from subscriptions or paid advertisements). Both of these spheres live or die by the page view/like/retweet. Thanks to this moment of quantifiable late-capitalism, basically everyone has to justify themselves by generating clicks. Granted, the local blogger who “does it for fun” can engage with books (or, more likely, social politics) in a way that pays no mind to the number of visitors their piece draws in, but: that doesn’t mean anything, since criticism with no stakes is hardly criticism; and also, everyone writes out of self-interest and a desire for “enough” people to read and respect them.

The point being: Everything is based in shares/retweets/pageviews. Quantifiable results. This is an age of measurables. So how should a critic meter out the 100 books they read in a given year?

There are no stats on this (yet), so let me speculate irresponsibly: In that position, I would read 50 books by pre-ordained “literary” writers (let’s pretend for a hot second that I’m a critic of literature generally), so, in 2017, I’d read the new Roxane Gay, Paul Auster, Rachel Cusk, Ali Smith, George Saunders, J. M. Coetzee, Viet Thanh Nguyen, etc., etc., etc., books, along with the two dozen titles that are hyped at BookExpo America, at the American Library Association Summer Conference, at ABA’s Winter Institute, and so on and forth.⁵ Let’s pretend that takes up seventy of the hundred titles I’ll read in any given year. Twenty of the remaining are books that catch my eye for one reason or another—I met the author, the publicist convinced me to take a chance, it’s a book I think I can make my brand on. The last ten—translations.

That was all an epically long route to trying to demonstrate how “lists” function. There are a dozen websites (re: tastemakers) out there who do “best of XXXX month” lists capable of guiding my reading to the ten most viable works of translated literature.⁶ None of the sites I have in mind (such as Flavorwire, Buzzfeed, The Millions, and Vol. 1 Brooklyn, the latter two of which are discussed below) are nearly influential enough⁷ to causally determine which books get covered, but the slightest bit of research points to a significant overlap between books included on these lists and books that receive reviews in more mainstream, traditional (re: more widely read) outlets.

In March 2016, Flavorwire and the BBC highlighted seventeen different books to read that month. (Two titles appeared on both lists: What Is Not Yours Is Not Yours and Margaret the First.) Of these books, fifteen were reviewed by the New York Times, and the two that weren’t (Heads: A Biography of Psychedelic America and Blackass were reviewed in the San Francisco Chronicle and NPR respectively. It’s not like they went unnoticed.) Out of the 10,000 or so trade titles published that month (using Bowker data for fiction, science, biographies, sociology, poetry, history, philosophy, etc.), these two sites chose seventeen books, and had an almost one to one correspondence with which titles our country’s most influential newspaper chose to review. It would be more interesting—and illuminating—to look at sales data for these various titles and see how that correlates with being included on lists and/or reviewed in the New York Times, but I don’t personally have access to that data.

In the end, a list of books to look forward to is no more than a list of books to look forward to, but there is an inherent value to being included on a list. If nothing else, these list-making sites precipitate—and participate in the creation of—the “information cascade” that overwash successful titles.

Let’s look at a couple of examples of how these lists treat the books they choose to highlight. First up is Vol. 1 Brooklyn’s recent post on books to read in January 2017:

Han Kang’s first novel to appear in English, The Vegetarian, was one of the most jarring works of fiction we’ve read in a while. Human Acts takes a broader view of humanity, focusing on a host of reactions to the death of a young man in a political action in South Korea. We’re looking forward to experiencing her prose in a new context with this novel.

Two things about this write-up: it posits a pre-selection criteria on Han Kang’s previously published book (“one of the most jarring works of fiction we’ve read in a while) while openly admitting that the list-maker has yet to read the book being put forth as one to read this month (“we’re looking forward to experiencing her prose in a new context with this novel”).

In case you think this is a one-off example of choosing to promote a book before reading it, here are some other lines from this list: “we’re eager to read this,” “we’ve been curious about what’ll come next,” “given the setting and thematic aspects of this book, it may also be a timely read,” “we’re eager to see a new side of his work show up here.”

(As a side-note, I did the exact same thing when I was writing previews for this website. I understand and acknowledge the challenges to featuring only books you’ve read on those sorts of post. This isn’t meant to criticize Vol. 1, but to look at how that particular type of post functions and what it means for book culture and book marketing as a whole.)

The write-ups in the Great First-Half 2017 Book Preview from The Millions are written in such a way to make it at least seem like the list-makers have read these books that they’re including. (Although I highly doubt they’ve read many/any of the titles coming out in June or beyond.) But there’s still a lingering question of why these particular books have been chosen. Have the editors of The Millions read even 5% of the works of fiction coming out between January and June of this year? 1%? .01%? There’s no question that these selections are primarily anticipatory, that they assume these are the books that culture and readers will embrace once they’re available, the books that will be reviewed and bought instead of the titles left off of the list. This happens by necessity, but also raises the question of what elements are at play to make the contributors anticipate these particular books instead of others. (The answer to this is probably pretty obvious, but still, are they looking forward to the “best” books, or the most hyped?)

Again, having a book appear on one of these lists doesn’t cause it to become popular or successful, but given the overlap between a) these lists, b) mainstream reviews, c) bookstore availability, and d) sales (all of which should be investigated more in a different post), it’s very likely that the particular books on these lists will be the ones dominating the cultural conversation. Everything converges to support a very small number of titles, pre-selected from a pool of titles far too large to thoroughly explore and evaluate.

I have suspicions as to why certain titles are pre-chosen to “make it,” along with some understanding of the power dynamics and book market logistics that help these titles become successful, but here’s what I’m more interested in: the way in which book coverage seems to have reversed the “selection” and “market magnification” processes of book promotion. Readers and critics don’t read hundreds of titles and uncover a new form or technique that makes a novel particularly successful and then promote it through the marketplace; instead, the books that will be promoted and disseminated widely are already selected, known to the culture almost as soon as they’re announced⁸, and only then do critics and readers find reasons why these books are so great. We start from a point of view that these are the “best” books, the ones “worth reading” and then work backward. No one has the time or patience to read all the books coming out in 2017 from contemporary Argentine writers, so we just assume the tastemakers have it right and that the one book included on these lists/reviewed by mainstream media/displayed on a bookstore table must be the best representative of that category of writing. And even if a reader disagrees, how many other Argentine books have they read this year to compare the pre-selected title against? Less than ten, I’m certain.

We live in an age in which it’s not just information that’s accelerated, but the markets themselves. The magnification aspect of the book market—in which sales increase exponentially—is set in place before readers ever even see a particular book. This keeps the messiness of reader responses as far from the market as possible, consolidating power by offering readers a choice that never really was a choice at all. Or at least is very limited and predetermined by the sphere of critics, tastemakers, and promoters, who used to react to the market, instead of driving it. The biggest players in this market gravitate towards this situation, since it’s much safer and more predictable that way. To know which books will do well enough, to cover the titles that people will definitely be talking about (thus perceived as “important”), to stock the titles that are most likely to sell—the closer these things are to certain, the more stable and profitable the industry is. Hits can come out of nowhere and far exceed sales expectations, but it’s best if that happen in a context in which you already control the baseline for as close to 100% of the products you’re putting into the market as possible.

¹ I have to point out that this post is not a rant about list as “dumb,” but an analysis of how lists function. Which elevates it. In my eyes, at least.

² This is a maybe bullshit claim, but here’s my breakdown. On average, over 50,000 works of fiction are published in the U.S.annually. Translations made up 1% of that in 2016. (There were 500 works of fiction published in translation in the U.S. last year.) If I randomly selected novels from my local bookstore—independent or Barnes & Noble—it’s likely that at least 2 in 100 would be translated. Given the plethora of “known” English writers with new books published in a given year—50? 100? 500?—more than 90% of the works read by an average critic will have been originally written in English, leaving what?, 10 books maximum from the rest of the world? Feels like chance to me.

³ I can actually calculate this, and will. Either for this piece, or for a future one, depending on how much time I have in the very near future.

⁴ Come at me, bro?

⁵ You’re not a real critic until you review the most reviewed authors.

⁶ “Viable” as in, I’ll review them and someone will pay me for those clicks.

⁷ Alexa site popularity rankings are a good enough means for demonstrating the relative influence of the websites I have in mind. First off, the list-making sites, the ones that I think help sort the great unread books and provide a sort of precipitory guidance: Flavorwire is the #15,576 most popular site in the world (not necessarily for their books coverage, but), The Millions is #175,585, Vol 1 Brooklyn #1,979,614, and Lit Hub #43,012. By contrast, the New Yorker is #1,413, New York Times #103 (again, not just books coverage), NPR #662, and Slate #920. None of this is surprising, but it does tie into my core idea: to get more eyeballs, small sites want to anticipate what the significantly more well-read sites will eventually cover. It’s infinitely better to be part of the information cascade that outside of it.

⁸ Garth Risk Hallberg’s City on Fire is a perfect example of this. The novel is decent enough, but before anyone had even read it, it was guaranteed coverage in every possible outlet simply given the size of its advance. From Ron Charles’s review in the Washington Post: “Having reportedly paid nearly $2 million for the manuscript, Knopf must be praying that City on Fire is worth its weight in Goldfinch. Such irrational exuberance can’t buy a spot on the bestseller list, but it can guarantee coverage. So prepare yourself for what passes for a book publicity juggernaut: Over the next few weeks, you’ll read about this novel everywhere, and you’ll hear the young author interviewed on NPR, and you’ll see pyramids of City on Fire at your local bookstore. And at some point, you’ll wonder, ‘Should I read this novel — or three others?’”