The Travolta moment

There’s a moment halfway through Jonathan Franzen’s The Corrections when Enid Lambert, the matriarch of the novel’s dysfunctional Midwestern family, visits a doctor on a cruise ship. It’s an important scene—Enid leaves with a handful of antidepressants that will play a big role later in the story—and Franzen lavishes his usual amount of care on the sequence, which runs for a full nine pages. But here’s how he chooses to describe the doctor on his first appearance:

He had a large, somewhat coarse-skinned face like the face of the Italian-American actor people loved, the one who once starred as an angel and another time as a disco dancer.

I adore The Corrections, but this is an embarrassing sentence—one of the worst I’ve ever seen pass the pen of a major novelist. It’s particularly surprising coming from Franzen, who has thought as urgently and probingly as any writer alive about the problem of voice. But it’s also the kind of lapse that turns out to be unexpectedly instructive, precisely because it comes from an author who really ought to know better.

So why does this sentence grate so much? Let’s break down the reasons one at a time:

1. Franzen clearly wants to tell us that the doctor looks like John Travolta, but he’s too shy to come out and say so, so he uses more than twenty words to convey what could have easily been expressed in two.
2. In the process, he’s false to his character. No woman of Enid’s generation and background would have any trouble coming up with Travolta’s name, especially if she were familiar with his role in Michael, of all movies. It’s not like she’s trying to remember, say, Richard Jenkins.
3. Worst of all, it takes us out of the story. Instead of focusing on the moment—which happens to be a crucial turning point for Enid’s character—we’re distracted by Franzen’s failure of style.

And the punchline here is that a lesser novelist would simply have said that the doctor looked like Travolta and been done with it. Franzen, an agonizingly smart writer, senses how lazy this is, so he backs away, but not nearly far enough. And the result reads like nothing a recognizable human being would feel or say.

I got to thinking about this after reading John McPhee’s recent New Yorker piece about frames of reference. McPhee’s pet peeve is when authors describe a person’s appearance by leaning on a perceived resemblance to a famous face, as in this example from Ian Frazier: “She looks enough like the late Bea Arthur, the star of the nineteen-seventies sitcom Maude, that it would be negligent not to say so.” Clearly, if you don’t remember how Bea Arthur looks, this description isn’t very useful. And while any such discussion tends to turn into a personal referendum on which references are obvious and which aren’t—McPhee claims he doesn’t know who Gene Wilder is, for instance—his point is a valid one:

If you say someone looks like Tom Cruise—and you let it go at that—you are asking Tom Cruise to do your writing for you. Your description will fail when your reader doesn’t know who Tom Cruise is.

And references that seem obvious now may not feel that way in twenty years. McPhee concludes, reasonably, that if you’re going to compare a character to a celebrity, you need to pay back that borrowed vividness by amplifying it with a line of description of your own, as when Joel Achenbach follows up his reference to Gene Wilder by referring to the subject’s “manic energy.”

When we evaluate Franzen’s Travolta moment in this light, it starts to look even worse. It reminds me a little of the statistician Edward Tufte, who famously declared that graphical excellence gives the viewer the greatest number of ideas in the shortest time with the least ink in the smallest space. In his classic The Visual Display of Quantitative Information, he introduces the concept of the data-ink ratio, which consists of the amount of data ink divided by the total ink used to print a statistical graphic. (“Data ink” is the ink in a graph or chart that can’t be erased without a real loss of information.) Ideally, as large a proportion of the ink as possible should be devoted to the presentation of the data, rather than to redundant material. As an example of the ratio at its worst, Tufte reprints a graph from a textbook that erased all the data points while retaining the grid lines, noting drily: “The resulting figure achieves a graphical absolute zero, a null data-ink ratio.” And that’s what Franzen gives us here. In twenty words, he offers no information that the reader isn’t asked to supply on his or her own. To be fair, Franzen is usually better than this. But here, it’s like giving us a female character and saying that she looks like Adele Dazeem.