April 22, 2004

Attention Deficit

Mark's posting on Camille Paglia's charges of decline in attention is right on the mark -- this is just an antique jeremiad in new packaging. People have been saying the same thing for centuries, with no more justification than anecdotal observations. (As Montesquieu said, looking back on the long line of complaints about the state of culture: "If all of this were true, we would be bears today").

One person who has made an honest effort to quantify these effects is Todd Gitlin. In an article in The Nation a few years ago called "The Dumbdown," he reported a study he'd done that showed that that the length of the average sentence in novels on the New York Times bestseller list has decreased by more than 25 percent over the last sixty years, while the average number of punctuation marks per sentence has dropped by more than half.

But that method is subject to lots of confounds -- for one thing, the bestseller lists are computed very differently now. And I was curious enough about this to do my own little study, with Brett Kessler, which revealed a very different pattern.

Kessler and I did similar calculations, not for bestsellers but for articles from The New York Times and Science. (For the Times we took the lead sentences of the most prominent story for each of 40 consecutive daily issues starting in October 1 of every twentieth year, going back to 1856. For Science we used a slightly different but roughly equivalent method for selecting articles, beginning in 1896.)

We found that both sentence length and number of punctuation marks per sentence had indeed declined slightly over the period that Gitlin looked at -- in 1936 the average Times lead sentence was almost 38 words long; in 1996 it was less than 35 words long. The figures in Science were analogous -- between 1936 and 1996, sentence length dropped from 27.1 to 25.8, though the average number of punctuation marks per sentence actually increased slightly. In fact that was part of a century-long cycle:

 
NYT
Science
 
Words per sentence
Punctuation per sentence
Words per sentence
Punctuation per sentence
Year
Mean
StdDev
Mean
StdDev
Mean
StdDev
Mean
StdDev
1856
31.8
19.6
2.6
2.7
-
-
-
-
1876
29.9
15.7
1.7
1.8
-
-
-
-
1896
30.0
13.4
2.0
1.8
34.9
23.9
3.3
4.1
1916
38.1
20.8
2.1
2.6
36.6
20.4
2.1
2.4
1936
37.8
11.1
1.8
1.3
27.1
11.8
1.3
1.3
1956
21.0
6.3
0.7
0.9
30.2
13.6
1.9
1.8
1976
34.0
9.7
1.3
1.0
23.2
10.5
1.8
1.9
1996
34.7
11.8
1.9
1.1
25.8
11.2
1.8
2.1
MEAN
32.2
1.8
29.7
2.0

But what does all this mean? The differences are pretty small, and in any event it would be hard to argue that either publication has been dumbed down over the course of the past 70 years, or that it requires less attention to read Science now than it did then.

And even if you were determined to interpret declining sentence length as an indicator of declining attention capacity, you'd be led to a curious conclusion. In fact, the mean sentence length in the Times reached a low of 21.0 in 1956, but since then it has been climbing -- by 1996 it was up almost 75 percent from its low. And the average number of punctuation marks per sentence has almost doubled over that period. Similarly (if less dramatically) for Science, where there has been a 10 percent increase in mean sentence length over the past 20 years.

In short, sentences have gotten longer and more complex since Camille Paglia's youth. If you're looking for a decline in attention, you might start with hers.

Posted by Geoff Nunberg at April 22, 2004 02:49 PM