Obama's Imperial 'I': spreading the meme

« previous post | next post »

I guess it's now officially a Media Meme: Obama's "royal we has flowered into the naked 'I'". First Terence Jeffrey ("I, Barack Obama"), then George Will ("Have We Got a Deal for You"), now Stanley Fish ("Yes I can"):

By the time of the address to the Congress on Feb. 24, the royal we has flowered into the naked “I”: “As soon as I took office, I asked this Congress.” “I called for action.” “I pushed for quick action.” “I have told each of my cabinet.” “I’ve appointed a proven and aggressive inspector general.” “I refuse to let that happen.” “I will not spend a single penny.” “I reject the view that says our problems will simply take care of themselves.” “I held a fiscal summit where I pledged to cut the deficit in half by the end of my first term.” That last is particularly telling: it says, there’s going to be a second term, I’m already moving fast, and if you don’t want to be left in the dust, you’d better fall in line.

There’s no mistaking what’s going on in the speech delivered last week. No preliminary niceties; just a rehearsal of Obama’s actions and expectations. Eight “I”’s right off the bat: “Just over two months ago I spoke with you… and I laid out what needed to be done.” “From the beginning I made it clear that I would not put any more tax dollars on the line.” “I refused to let those companies become permanent wards of the state.” “I refused to kick the can down the road. But I also recognized the importance of a viable auto industry.” “I decided then…” (He is really the decider.)

The trouble with this idea, as often with the insights of the punditocracy, is that there's no evidence that it's true. Worse, evidence is easily available to disconfirm it.

In yesterday's post ("Fact-checking George F. Will"), I counted, and discovered that in that "speech delivered last week" — the one about the auto bailout — President Obama used forms of the first-person pronoun at a rate of 1.7% (i.e. 42 instances in 2423 words). Compared to Obama's press conference of 2/9/2009 — before the 2/24/2009 date that Prof. Fish identifies as the pronominal turning-point — this is a lower rate, not a higher one: the 2/29/2009 press conference exhibited a rate of 2.6% (205 of 7,775).

Furthermore, if we compare the first press conferences of the previous two presidents, we find higher rates yet: 3.9% for Bill Clinton (275 in 6,935) and 4.5% for George W. Bush (300 in 6,681).

You'd think that Prof. Fish would have done some counts of this sort before displaying his analysis in the pages of the New York Times, especially since he's explicit about the fact that he's inferring the president's attitudes by counting (and analyzing) pronouns rather than by using ESP.  Unfortunately, at some point in the past few decades, literary scholars seem to have abandoned the assumption that claims, even quantitative ones, ought to be testable.

But maybe I'm being unfair. Fish explicitly compares Obama against Obama's own earlier practice, and only implicitly against other recent presidents. (And speeches are a different genre from press conferences.) So I guess I should go back and look at some of Obama's pre-presidential transcripts as well.  Any bets on what I'll find?

Here goes. Fish mentions 8 of Obama's speeches. I've arranged them in chronological order below, and given each one a link to a transcript. (I've added one that Fish didn't mention: Obama's first presidential press conference, which I analyzed yesterday, because it provides a useful point of comparison to the first press conferences of previous presidents.) The rightmost column gives the percentage of words in the transcript that are first-person singular pronouns (I, me, my, myself, mine):

1 the presidential announcement 2/10/2007 2.1%
2 the Iowa victory address 1/3/2008 1.9%
3 the Philadelphia speech on race 3/18/2008 1.8%
4 the nomination acceptance 8/28/2008 2.5%
5 the Grant Park victory speech 11/4/2008 1.9%
6 the inaugural address 1/20/2009 0.2%
7 the first presidential press conference 2/9/2009 2.6%
8 the address to congress 2/24/2009 1.6%
9 the GM bankruptcy speech 6/1/2009 1.7%

(Recall again that in comparison, William J. Clinton's first two presidential press conferences had a rate of 3.9%, while George W. Bush's first two presidential press conferences had a rate of 4.5%.)

Prof. Fish tells us that there's a "pattern of pronouns" in Barack Obama's speeches, whereby an increase in the use of 'I' and 'me' and 'my' marks his transition from the "restraint and modesty" of a politician whose "self-assertion is immediately muted" to the "imperial possession" of a "guy [who]'s completely in charge, making decisions, giving instructions, deploying resources, assigning tasks — a combination point guard, quarterback and clean-up hitter." And according to Fish, the "flowering" of the "naked 'I'" is first evident in the 2/24/2009 address to Congress, and "the note of imperial possession" is fully on display in the 6/1/2009 GM bankruptcy speech. Maybe so — but in fact, the rate of use of first-person singular pronouns in those two speeches is lower than in *any* of the six earlier speeches that Fish cites — except for the inaugural address, where first-person singular pronouns were almost completely absent, presumably for stylistic reasons.

Maybe president Obama is now "completely in charge, making decisions, giving instructions, deploying resources, assigning tasks". That's more or less what the leader of the executive branch of government is supposed to do, after all. But his (or his speechwriters') rate of use of first-person singular pronouns hasn't increased — in fact, maybe it's gone down.

[In fairness to Fish, he doesn't make his case only or even mainly on a simple count. He suggests a distinction between the earlier speeches where "Both syntactically and substantively, the 'I' is subordinated to the projects to which it is dedicated", and the recent speeches, where the "royal we" emerges in the inaugural, and then "flowered into the naked 'I'" in the address to Congress.

On the other hand, he does wave some numbers around ("Eight “I”’s right off the bat … Accompanying the “I”’s are a bevy of “my”’s"), so it's worth establishing that simple pronoun counts undermine his argument rather than supporting it.

And after giving the texts a modest amount of scrutiny, I can't see any form of (inter-subjectively stable) linguistic analysis that would support Fish's notion that Obama's earlier uses of 'I' are "syntactically and substantively … subordinated to the projects to which it is dedicated". Perhaps some reader can figure out how to turn this idea of Fish's into a testable claim, in a way that would support his argument.

Until then, I'm going to assume that Stanley Fish has joined Terence Jeffrey and George F. Will in a classic display of confirmation bias: they feel that president Obama is arrogant and uppity, and so his uses of the first-person singular pronoun are striking and salient to them, in a way that previous presidents' pronouns were not. I freely admit, of course, that I have only a little more evidence for this opinion than they do for their views on Obama's egotism.  ]

[Update — a bit more on presidential first-person plural pronouns here, and a deeper analysis of Fish's "Royal we" view of Obama's inaugural here, and another pundit joins the pack here. ]


  1. Outis said,

    June 8, 2009 @ 8:35 am

    Even if Fish's facts weren't wrong, what's wrong with the president using the first person pronoun? He is the president, he is the policy and decision maker, he does represent the US. I'm sure if he kept saying "we", he would be attacked for shirking responsibility.

  2. Aaron Davies said,

    June 8, 2009 @ 8:44 am

    Doesn't Fish explicitly disbelieve in truth anyway? At least that gives him a bit of an excuse Will lacks…

    [(myl) Yes, more or less. ]

  3. Amy Reynaldo said,

    June 8, 2009 @ 8:45 am

    I wonder how much of the commentary hinges on white privilege. Perhaps Will, Fish, et al., simply don't notice when a white president tosses around the first person singular, but it jumps out at them in a POC (that's "president of color," of course). Perhaps they expect powerful white men to sound commanding and are still getting used to the experience of listening to a particularly powerful black man sounding—and being—commanding.

    [(myl) Indeed — I added a similar speculation (about confirmation bias) to the end of the post, as you were adding this comment. ]

  4. AA said,

    June 8, 2009 @ 8:59 am

    Shouldn't you count the percentage of I's to the number of subjects in a speech rather than to the total word count? I think I's are noticeable because each sentence no matter how long has one subject and if a large percent of these are first person singular it does give an impression of excessive self-centredness.

    [(myl) Be my guest — the links to the transcripts are in the posts above.

    There are lots of different ways to count such things — ratios to other pronouns, ratios to number of tensed clauses, etc. etc. If you tried enough of them, you might find one to suit. Or maybe not. My main point is that Prof. Fish didn't even try to make a testable claim. He just read over (or listened to, or remembered) some speeches, and gave us his reaction. But you can be sure other media voices will now take it as established that Obama, once in office, started over-using first-person singular pronouns. ]

  5. Chris said,

    June 8, 2009 @ 9:00 am

    If he used "we" in the inaugural address (has anyone checked?), I would interpret that as "we Americans": we live in troubled times, we have great challenges ahead of us, etc. That seemed to be the tone of the inaugural, from what I remember.

    Speaking on behalf of the nation does seem somehow related to the royal we, but it's not quite the same (even discounting the fact that Obama was elected to represent us) as the more famous uses like "We are not amused", which clearly does not signify that the entire English nation was not amused.

    I don't expect Obama to say, or even think, "L'etat, c'est moi!" – although some of his predecessors might have. A community organizer is in no danger of confusing the symbol with its referent.

  6. Chris said,

    June 8, 2009 @ 9:05 am

    Since Amy's comment appeared while I was writing mine, I'll add that speaking on behalf of the nation is also something likely to seem extra salient when done by someone who isn't the hearer's idea of a typical American (i.e. a white male).

    Anyone who thinks about it ought to realize that there's 300 million Americans and they're not all alike and therefore the idea of a "typical American" is asinine. But maybe I shouldn't blame Plato too much – the way essentialism keeps cropping up all over the place in human thought strongly suggests that it's a flaw in the human brain, and he's just the best-known victim who took it to the ridiculous extreme of exalting the abstraction over the reality.

    Or maybe Sapir-Whorf is right and it's the fault of a language that has words like "typical". What languages have no words for typical, and are people who speak them better at grasping the importance of variation?

  7. Dan Scherlis said,

    June 8, 2009 @ 9:08 am

    Strange. I once again fail to understand what Fish is driving at. It's almost as if his texts were without instrinsic meaning.

  8. Dan Lufkin said,

    June 8, 2009 @ 9:35 am

    Fish has gotten so cranky lately that I'm thinking of taking his emblem off the trunk of my car.

  9. kip said,

    June 8, 2009 @ 9:54 am

    I'm not sure that the ratio of first-person-singular pronouns vs. total word count is the best measure. Wouldn't the ratio of first-person-singular pronouns vs. first-person-plural pronouns be more effective in testing his claims?

    If Obama tends to be wordier than Bush and Clinton, it could be that both singular and plural first-person pronouns have a lower rate compared to overall word count, but the ratio of singular to plural is still higher.

    [(myl) As I said in response to another comment, there are lots to things to count, and lots of ways to model the resulting patterns of numbers. The problem is that pundits like Will and Fish aren't counting or checking anything, they're just giving their impressions — which are probably based on various sorts of illusions, especially group-think and confirmation bias.

    If we were seriously attempting an empirical investigation of political rhetoric, your suggestion would a plausible thing to try. But that's not what Prof. Fish is up to, as far as I can tell. ]

  10. Andy Hollandbeck said,

    June 8, 2009 @ 12:47 pm

    I just don't get it. Our president is taking personal responsibility for the actions of our government. Have we become so jaded by the constant doublespeak of politicians that when a powerful man takes responsibility, we see as somehow wrong?

    He didn't say "My cabinet was instructed…" or "A decision was made…" or "Not a single penny will be spent…" Could it be that conservatives are more anti-S&W than even Geoff Pullum, preferring weak, impersonal, and passive to strong, personal, and active?

    Or are they just preparing to re-run Bob Dole for the 2012 election?

    [(myl) I don't think that the issue is what "we" see. Rather, (some members of) the chattering classes need to associate a convenient stereotype with each public figure. In this case, the "Obambi" concept is not working out (remember that one?) and so OK, how about Julius Caesar? ]

  11. Ryan Denzer-King said,

    June 8, 2009 @ 1:24 pm

    Maybe I'm being overly naive or accomodating, but it seems to me that there must be something that caught Fish et al.'s eyes. Perhaps it'd be be worthwhile to compare frequences of 1st-person pronouns in Obama's main clauses vs. subordinate clauses, relative to Bush and Clinton, or sentence-initial pronouns.

    [(myl) Maybe. Or maybe, as Amy Reynaldo suggests above, it's the gap between normal presidential reality and what Will, Fish et al. expect from some who looks like Barack Obama.

    Anyhow, I don't think that it should be our job to find some measure that rescues pundits from the appearance of writing nonsense. ]

  12. Jim said,

    June 8, 2009 @ 1:43 pm

    "I just don't get it. Our president is taking personal responsibility for the actions of our government."

    And he is talking like a citzen, a single citizen, who happens to be in a certain position for a time.

    "Have we become so jaded by the constant doublespeak of politicians that when a powerful man takes responsibility, we see as somehow wrong?"


    "I just don't get it. "

    Yes, you do get it.

  13. J. W. Brewer said,

    June 8, 2009 @ 2:24 pm

    It's probably a common and bipartisan affliction in political rhetoric for the same type of behavior (syntactic or otherwise) that is considered evidence that my guy is decisive and responsible is also perfectly good evidence that your guy is arrogant, self-centered, drunk with power, etc. In modern American politics, those who are sufficiently self-effacing that they can't be credibly accused of arrogance, lust for power etc. tend not to end up getting very close to the Presidency.

    Will seemed focused on "I," whereas Fish expanded his focus to some uses of "my." There's a possibility (for those looking for alternative research projects that certainly neither pundit engaged in) that "I" in particular is sometimes found more grating than other first-person singular pronouns. I seem to recall (details very fuzzy) at some point getting usage advice from an English teacher that it was bad for too high a percentage of sentences in an essay to begin with "I" and then figuring out this objection could be worked around by recasting every second or third sentence to begin with wording like "It seems to me that" or "In my view . . ." It should be possible to quantify the extent to which particular politicians do or do not use such workarounds in lieu of what Fish calls the "naked 'I.'" Not that I'm volunteering to do that quantification or suggesting that Prof. Liberman ought to have done so.

  14. Mike Anderson said,

    June 8, 2009 @ 2:32 pm

    Dang! Just when I needed a good, short case study about research design for an observational study, along comes Liberman with a doozy! Thanks for a most edifying example.

  15. bulbul said,

    June 8, 2009 @ 4:01 pm

    but it seems to me that there must be something that caught Fish et al.'s eyes.
    I think the point of this whole affair is that there isn't and that these guys don't even bother to look.

  16. chad said,

    June 8, 2009 @ 4:13 pm

    Your attack on Dr. Fish is so narrow minded. You're trying really hard to empirical, but do you really think comparing 1 or 2 press conferences between three presidents is a large enough sample to represent some kind of significant scientific truth?

    [(myl) No, but it's counts from five more press conferences than George Will or Stanley Fish gave us. ]

    Your own bias towards empiricism (numbers=objective truth) is puzzling but unsurprising given the highly-academic, drier than a desert post you wrote. Not to mention the intellectual fallacy of a comparison that takes all of the president's words completely out of historical and cultural context. Dr. Fish's analysis may not be quite as empirical, but he is a lot closer to the truth than you think you are. Speaking of egos run wild…

    [(myl) As always, our marketing department stands ready to refund double your subscription price in case of less than complete satisfaction. But if you'd like something moister, we may be able to oblige. ]

  17. Amy Reynaldo said,

    June 8, 2009 @ 5:17 pm

    Chad, what's the historical and cultural context? "ZOMG! President Obama isn't a white man and yet he talks like he's the leader of the free world anyway"? I dunno—I hear Obama speaking and what I hear is smooth, sonorous gravitas, a command of complex issues. I don't hear overweening ego.

    It's all political sniping. I didn't like the content of Reagan or either Bush's speeches for the most part. I didn't like their style. Why? Probably because I disagreed with their politics. If Will and Fish were honest, they'd quit pretending that their problems with Obama relate to his rhetorical style and acknowledge that they're generally not going to like anything he says.

  18. EconTech » Links for 2009.06.08 said,

    June 8, 2009 @ 7:31 pm

    […] Obama's Imperial 'I': spreading the meme: Mark Liberman addthis_pub = 'econtech'; addthis_options = 'email, digg, delicious, google, newsvine, reddit, stumbleupon, technorati, twitter, more '; Posted by: computer.economist on 8 Jun 2009 at 17:30 -0500 Categories: Links […]

  19. The other Mark P said,

    June 9, 2009 @ 6:24 am

    Dr. Fish's analysis may not be quite as empirical, but he is a lot closer to the truth than you think you are.

    "Not quite as empirical" huh? By my reckoning, Fish's analysis is not even remotely empirical. He's made it up. It's BS.

    It is poor form to argue that one opinion is better than another on the basis that it is higher level thinking when the actual facts on the ground argue the reverse.

    We have seen decent factual evidence that Fish is wrong. If Fish's analysis is closer to the truth, show us how. Mere statement that he is correct is no more valid than Fish's original position is. Unless you believe the "say it three times and it becomes true" version of argument.

  20. Phil Sheehan said,

    June 9, 2009 @ 12:12 pm

    How much of the right-wing distress stems not so much from the word itself but from the way the President pronounces it? He pronounces it: the word "I," as if it were "eye" rather than "oi" or "ah" or any of the other variants. And he's a POC from the Midwest; how dare he speak as if he were an East Coast intellectual?

    "I" may occur more often in speeches by Willy and the Shrub, but neither of them pronounced it so clearly. Deferential? Perhaps, but I think corn-pone folksy is more likely.

    Sooner or later we're going to have to acknowledge what's happening. Obama is being trashed because he's a black man who had the nerve not only to seek the Presidency, but actually to win it. The Limbaughs and the Gingriches — and, alas, the Fishes — may not themselves be racists, but they know their audience.

  21. Idiot pundits « Later On said,

    June 9, 2009 @ 2:04 pm

    […] pundits are too quick to type: they get an idea, and right away they're pecking out a column. Here's the latest: I guess it's now officially a Media Meme: Obama's "royal we has flowered into the […]

  22. One more myth on the pile « Around Teh Table said,

    June 9, 2009 @ 6:02 pm

    […] Seems that in light of reality, not so much. […]

  23. Texas Aggie said,

    June 10, 2009 @ 12:43 am

    You have to take into account the source of these statements. George Will has proven himself to be a clueless dope time after time. Any analysis that he announces can be safely ignored (see his columns on global warming). Stanley Fish has a long history of not being able to analyze anything. Remember that he is one of the original deconstructionists who got blown away by some physicist publishing a work of complete and deliberate baloney in one of the deconstructionist journals. What ever happened to them, by the way, and why doesn't Fish go join them?

  24. Understanding Government » Blog Archive » Taking the “I” out of . . . Chief Executive? said,

    January 15, 2010 @ 9:56 am

    […] often little more than one big personal pronoun, to attack the president for this, especially when it turns out Obama doesn't use the word "I" more than other recent presidents.  But there sometimes does seem to be an odd focus on himself in the president's words […]

  25. Warum amerikanische Journalisten Obamas Pronomen zählen « USA Erklärt said,

    September 6, 2012 @ 9:30 am

    […] bezeichnet werden. Die "allgemeine" Variante des Vorwurfes stimmt demnach zumindest für Obama nicht: [B]ut his (or his speechwriters') rate of use of first-person singular pronouns […]

RSS feed for comments on this post