Self-reference and narcissism

It is fast becoming a commonplace of American criticism that frequent use of the first person betokens narcissism. Last month Stanley Fish reported that he had listened closely to President Obama and had detected a growing preference for big I’s over little we’s. “[T]he note of imperial possession, the accents and cadences of a man supremely aware of his authority and more than comfortable with its exercise,” have creeped into his speech, Fish concluded. But it was not only the Left devouring one of its own. Last week Wall Street Journal columnist Peggy Noonan—best known as a speechwriter for President Reagan—belittled Governor Sarah Palin for being “self-referential to the point of self-reverence.” In the July 3rd announcement of her resignation from Alaska’s governorship, Palin kept saying “I’m, I’m, I’m,” Noonan complained. Over at the Language Log, Mark Liberman submitted Noonan’s claim to careful scrutiny (h/t: Neil Verma).

Adopting “two crude measures of ego-involvement,” Liberman compared Palin’s announcement to three similar speeches—Richard M. Nixon’s concession in the 1962 California gubernatorial election, Lyndon B. Johnson’s announcement in 1968 that he would not seek reelection to the presidency, and President Nixon’s resignation in 1974—and found that, by these measures, “Palin is more ego-involved than LBJ, but less than Nixon.” She used the various forms of the first-person singular four percent of the time, while Nixon was at 6.1% and 4.6% and Johnson at just 2%. Liberman also calculated the ratio of the first-person plural to the singular, observing that a “higher ratio suggests less ego-involvement,” and found that Johnson had the highest ratio (1.37), but that Palin’s (0.81) was strikingly higher than Nixon’s (0.17, 0.48).

I want to take Liberman’s analysis one step further, not to defend Palin—frankly, she doesn’t need my help—but to show that the folk psychology about frequency of the first person is badly off the mark. In short, self-reference is not evidence of narcissism, because historically even Nixon’s rate of I-talk is within the range of normal.

Plagiarizing Liberman’s method, I examined three English-language classics of the eighteenth century that were written in the first person and three from the nineteenth. Here are the results. (Please forgive the lack of a table.)

Earl of Chesterfield, Letters to His Son (1746–71)
Words = 286,074
1st sing. = 8,636
% 1st sing.= 3.0%
1st pl. = 777
Pl./sing. ratio = 0.089

Laurence Sterne, Tristram Shandy (1759)
Words = 190,268
1st sing. = 6,641
% 1st sing. = 3.5%
1st pl. = 816
Pl./sing. ratio = 0.123

Benjamin Franklin, Autobiography (1771)
Words = 65,935
1st sing. = 2,963
% 1st sing. = 4.5%
1st pl. = 678
Pl./sing. ratio = 0.229

Charles Dickens, David Copperfield (1850)
Words = 362,889
1st sing. = 22,959
% 1st sing. = 6.3%
1st pl. = 2,701
Pl./sing. ratio = 0.118

Mark Twain, Adventures of Huckleberry Finn (1884)
Words = 116,519
1st sing. = 4,914
% 1st sing. = 4.2%
1st pl. = 1,062
Pl./sing. ratio = 0.216

The Personal Memoirs of U. S. Grant (1885–86)
Words = 241,878
1st sing. = 4,692
% 1st sing. = 1.9%
1st pl. = 1,758
Pl./sing. ratio = 0.375

What do these figures prove? Accusations of narcissism cannot be sustained by citing the frequency of self-reference, not even as contrasted with the use of the first-person plural. While Ben Franklin has the second-highest percentage of references to himself, he also manages the second-highest ratio of plural to singular uses. In Liberman’s language, he displays both a relatively high degree of ego-involvement and clear evidence of relatively less ego-involvement. Again, it is not surprising to find that Grant, a military man, has the highest ratio of plurals to singulars among the six authors. What is surprising is that Nixon achieved a higher ratio (0.48 versus Grant’s 0.38) when he told the country at last that he was giving up the presidency.

I am left with two hypotheses, neither of which the folk psychologists and critics of American political discourse have entertained. First, the frequency of the first person is pretty likely to be a product of culture and history. The eighteenth-century British writers use the first-person forms less than the other four on my list. And in the twenty-first century heavy use of the first person is an accepted norm. An accepted grammatical norm, I might add. All it may demonstrate is a preference for constructing sentences in a certain way—a relatively easy way.

Replying to criticisms of first-person narration, the novelist David Isaak points out that the first-person encourages a straightforward construction that can wear upon readers:

I’ve heard more than one person comment that first-person narratives tend to start too many consecutive sentences with “I,” giving the impression we are listening to a Mexican folk song (“Ai—Ai—Yi—Ai . . .”). Fine—but I’ve seen just as many third person manuscripts starting paragraph after paragraph with “He.” Is “hee-hee-hee” somehow better?Isaak doesn’t notice that he starts both of his sentences the same way. But this is not to fault him. Starting sentences with the I is the default construction in current English, especially in informal discourse when the speaker’s (or, as here, the writer’s) mind is not on the form of what is being said.

And thus the second hypothesis. Person reflects genre. Despite the fact that he is an eighteenth-century author like Sterne and Chesterfield, Franklin uses the first person more often because he is writing an autobiography, a literary kind that, except when it is an exercise in long-winded self-concealment, like The Education of Henry Adams, depends helplessly upon the first person. Similarly, to accuse David Copperfield of “ego-involvement”—he uses some form of the first person 6.3% of the time—does not seem quite right. David is as much a “camera” as Christopher in The Berlin Stories; he is at least as interested in the people in his life as in himself. Consider, for example, the passage in which David first studies Uriah Heep in Mr. Wickfield’s office:It so happened that this chair was opposite a narrow passage, which ended in the little circular room where I had seen Uriah Heep’s pale face looking out of the window. Uriah, having taken the pony to a neighbouring stable, was at work at a desk in this room, which had a brass frame on the top to hang paper upon, and on which the writing he was making a copy of was then hanging. Though his face was towards me, I thought, for some time, the writing being between us, that he could not see me; but looking that way more attentively, it made me uncomfortable to observe that, every now and then, his sleepless eyes would come below the writing, like two red suns, and stealthily stare at me for I dare say a whole minute at a time, during which his pen went, or pretended to go, as cleverly as ever. I made several attempts to get out of their way—such as standing on a chair to look at a map on the other side of the room, and poring over the columns of a Kentish newspaper—but they always attracted me back again; and whenever I looked towards those two red suns, I was sure to find them, either just rising or just setting.David refers to himself eleven times in this passage—exactly five percent of the words are first-person forms—while referring to Heep just seven times (ten, if the references to Heep’s eyes almost as impersonal objects are included). Yet his entire attention is on Heep, not himself. The narrative strategy is to register Heep’s effect, because that is how—at least for Dickens—a man is to be judged.

Unless first-person genres and their self-referential purposes are taken into account, complaints like Fish’s and Noonan’s about “self-reverence” and the “imperial possession” are empty moralizing.

Update, I: Here are the numbers for this Commonplace Blog.

Words = 179,427
1st sing. = 2,362
% 1st sing. = 1.3%
1st pl. = 453
Pl./sing. ratio = 0.192

After the basic components of a sentence, the most common words here have been not (1,284), novel or novels (724), literature or literary (706), book or books (531), all of the various variations on the word Jew (386), and then American (359). No idea what to make of all this.

Update, II: Three more sets of figures.

Wilkie Collins, The Moonstone (1869)
Words = 197,669
1st sing. = 10,827
% 1st sing. = 5.5%
1st pl. = 1,189
Pl./sing. ratio = 0.110

Booker T. Washington, Up from Slavery (1900)
Words = 74,130
1st sing. = 3,120
% 1st sing. = 4.2%
1st pl. = 718
Pl./sing. ratio = 0.230

Woodrow Wilson, Presidential Addresses (1913–18)
Words = 92,886
1st sing. = 1,294
% 1st sing. = 1.4%
1st pl. = 2,069
Pl./sing. ratio = 1.599

The Moonstone is the first example of an unreliable narrative given by Wikipedia. Its numbers are nearly the same as those of Dickens’s novel, published eighteen years earlier. The age? The genre? Washington is close enough to Franklin to suggest that something around 4% is the rate at which autobiographies drop into self-reference. Meanwhile, Wilson’s use of the first person almost exactly mirrors Grant’s. His ratio of plurals to singulars, though, is the highest I have found, identifying an integral element of his political rhetoric.