In an unfamiliar city, I rarely walk into an empty restaurant. How good could the food be if nobody eats there? At least a crowded restaurant’s one that many people have already chosen, and that ain’t nothing. The number of patrons has no direct relationship with the quality of the food, but anyone choosing what to do with their time, money, or attention needs some way to predict quality. Or if not quality in an objective sense, an indication that there might be something on the menu to their liking. Crowdedness is a signal–quite possibly a weak one and certainly an imperfect one, but one that helps us to manage uncertainty. If we’re hungry, we need some heuristic that helps us to actually sit down and eat.
Recently the British newspaper The Guardian released a ranked list of what its editors called the “100 best novels of all time.” How do we know they’re the best? Because a lot of people say they are. Such a standard is about as helpful as peering in the restaurant window. To be fair, the Guardian‘s voters aren’t random. They’re accomplished folks. There are worse ways to choose a novel to read this summer than to be guided by the collective expertise of 172 prominent authors, critics, journalists, and academics (some are many at once). If Ian McEwan, Maggie O’Farrell, Merve Emre, and Salman Rushdie all say James Joyce’s Ulysses is one of the greatest novels, well then surely it is. Look at them all over there in Joyce’s Restaurant, eating with relish the inner organs of beasts and fowls. I guess I’ll try the mutton kidneys.
Then again, what if, unlike Joyce’s Leopold Bloom, I prefer not to give my “palate a fine tang of faintly scented urine”? The Guardian‘s list, conversation provoking as it is, makes it challenging to know if I’m inadvertently following the urine-loving crowd. The point isn’t just de gustibus non est disputandum; rather, it’s that there’s a deeply social element to taste, taste-making, and predictions about quality that the Guardian‘s final ranked list, plated and served, happens to obscure.
That’s why I created this digital network to see what it’s actually made of. Even more interesting than which novels received more votes are the networks of taste and taste-making that bind certain books together. As sociologists have long observed, cultural value accrues over time in contests among ideological coalitions. Network visualization, as Manuel Lima puts it in his book Visual Complexity, is “the cartography of the indiscernible.” Networks can make visible the underlying clusters and communities through which literary hierarchies form.
Note, for example, the subterranean connection Chimamanda Ngozi Adichie’s Half of a Yellow Sun shares with Arundhati Roy’s The God of Small Things in the Guardian data. Among those who listed Half of a Yellow Sun (Yagnishsing Dawoor, Diana Evans, Charlotte Northedge, Katherine Viner, and Emma Loffhagen), all but one (Viner) listed The God of Small Things as well. The network of voters raising up contemporary postcolonial fiction reveals far more than a ranked list in which Roy’s novel is #32 and Adichie’s is #62. Similarly, nothing predicts whether you chose Toni Morrison’s Beloved (#2) better than whether you listed Octavia Butler’s Kindred (#71) or Elena Ferrante’s My Brilliant Friend (#51). Fiction’s sisterhood networks run deep.
But other social forces are at work as well. When I peer in the window at Joyce’s Restaurant (more simply, voters for Ulysses), I’m presumably seeing folks whose tastes include experimental modernism, Irish lit, adaptations of epic and so on. But I must assume that among those in there already, there’s some number of people -maybe it’s small, maybe it’s large –who, as I’d do in a foreign city, entered because others were already in there. This is what social scientists call “social proof” – the tendency to infer value from others’ choices. It creates strange “rich get richer” effects where that which gets chosen gets chosen even more, and pleasure, observed, becomes its own advertisement. This is why crowdedness alone is a complicated signal. It’s hard to know whether you’re following the quality (Proust’s: kick-ass madeleines!) or following the crowd.
Networks have the virtue of at least making this phenomenon visible. It shouldn’t be too surprising that older and perhaps already more venerable novels have an advantage. Middlemarch was the top vote-getter and thus takes up the most space in the visualization. But if George Eliot’s novel were an English midlands gastropub, you’d better believe some people are tucking in for no better reason than that it’s been open since 1872. Anything open that long must have something going for it.
The point isn’t to cast any shade on Middlemarch or Ulysses but instead to foreground the sociological point that even readers as accomplished and sensitive as Stephen King, Tessa Hadley, Colm Tóibín and the like are subject to all kinds of social pressures. Their tastes emerge in relation to pre-existing canons and conversations; their public declarations are subject to scrutiny and critique. Identity politics, geopolitics, and nationalism can all subtly or not so subtly influence accounts of “greatness.” It’s easier to pronounce something great if you believe others think it’s great too, and the power and status of those others you’ve got in mind make a big difference.
“Being cultivated,” as Pierre Bayard has written, “is a matter not of having read any book in particular, but of being able to find your bearings within books as a system, which requires you to know that they form a system and to be able to locate each element in relation to the others.” More so than a hierarchical list, The Guardian‘s data as a network reveals “books as a system.” Whereas certain books enjoy tight, devoted readerships, other frequently-listed novels operate as shared spaces attractive to various communities with diverse tastes. It turns out, for example, that numbers one (Middlemarch), two (Beloved), and three (Ulysses) on The Guardian‘s list each exists in complex relation with the others. All of those novels, we might say, have their own gravitational pull. Middlemarch received 56 votes, Beloved got 43, and Ulysses 36, and yet only 6 out 172 voters chose both Ulysses and Beloved. 30 picked Ulysses but not Beloved; 37 picked Beloved but not Ulysses. Lovers of lush, prophetic Morrison only rarely align with students of “Stately, plump Buck Mulligan.” Middlemarch, apparently, has tables for both crowds.
Overall, the network of books and voters gives us a richer appreciation of the kinds of hidden forces that generate popularity and prestige. Ideally, a map like mine leaves us with clearer signals about communities of taste by helping us name and filter key sources of noise like canonical gravity and rich-get-richer effects. There’s no pure literary value outside of such social forces, but at least we can try to access more of what matters when choosing a novel. I’m not really one for organ meat, but I am willing to follow Ian McEwan over to Joyce’s. Instead of the mutton kidneys, I think I’ll have what Molly Bloom’s having.
