November 17, 2005

Has George W. Bush become more disfluent?

George W. Bush, in his 11/16/2005 press conference in Kyoto:

Obviously, the extent to which uh [0.295]
the Japanese government wants to give reconstruction money to Iraq is up to the Japanese government, and [pause 0.187]
to- to the- and I- as to the- [pause 0.205]
the- the uh deployment of troops, it's up to- [0.421] it's up to the government. [pause 1.237]
's what happens in democracies -- government makes decisions that uh [pause 0.598]
that uh that they're uh capable of living with, and that's [pause 1.966]
that's what we said, ((we)) said, do the best you can do; [pause 0.530]
make up your own mind, it's your decision, not mine.
(audio clip)

These are exemplary sentiments, but their expression is surprisingly chaotic, given that the question was a predictable one and the answer is a routine piece of diplomatic boilerplate.

Regular readers of this blog know that I think that George W. Bush has gotten a bad rap on the question of fluency and verbal facility. (You can find more discussion of this issue by searching Language Log for "Bushisms".) In reference to the 10/8/2004 presidential debate, for instance, I observed that both the moderator (Charlie Gibson) and John Kerry committed verbal flubs that were more striking than any of George W. Bush's miscues. One of Kerry's disfluencies:

And I'm gonna [pause 0.258]
put ((in)) place a better homeland security. [pause 0.239]
Effort. [pause 0.259]
Look at it. [pause 0.256]
95% of our containers coming into this country are not inspected today. [pause 0.769]
When you get on an airplane, your car- your [pause 0.421]
bag is- is- is [pause 0.184]
((b-)) x-rayed [pause 0.494]
but the cargo hold isn't x-rayed. Do you feel safer? [pause 0.686]

(audio clip)

As stress and distraction increase, people are more likely to break down verbally. Because we know this, we may pay more attention to disfluency when we suspect that the speaker is distracted and under stress. So my impression that George W. Bush has become more disfluent recently may be a fact about me and the conventional assessment of the political situation, rather than a fact about him. Still, I can add one small quantitative comparison to the subjective impressions. Compared to his 10/3/2005 nomination speech for Harriet Miers, his 10/31/2005 nomination of Samuel Alito had twice as many self-corrections in barely half the time. (More exactly: my transcription of Bush's speech nominating Harriet Miers shows two self-corrections in 1202 words over 547.7 seconds; my transcription of Bush's speech nominating Samuel Alito shows four self-corrections in 672 words over 314.8 seconds.)

I don't know how these self-correction rates compare to those typical of other public figures in similar situations (reading short speeches to a small live audience and a large broadcast audience). I do know that the duration of W's pauses in such speech reading, at least in recent times, is much longer than I would expect in presentations of this type (and longer than his pauses are in unscripted speech). But on this measure, at least, his Miers and Alito nomination speeches are entirely consistent, as this boxplot comparing the distribution of within-sentence and between-sentence pauses in the two speeches shows:

Within each "box", the horizontal line shows the median value, while the top and bottom of the box are the 75th and 25th percentile. The "whiskers" show the extreme values (as long as they are within 1.5 times the interquartile range of the box -- outliers beyond that range are shown as individual points).

As I pointed out in an earlier post, W's pauses in the Miers nomination are roughly twice the duration of those in Miers' remarks on the same occasion, and I'm confident that his current pausing in read speeches is in general much longer than the norm for American public figures. It's easy to speculate that this is a tactic he's adopted in defense against being criticized for making mistakes in speaking. It might be interesting to model the variation in pause durations in terms of local properties of the message, other than the obvious within-sentence vs. across-sentence distinction made above. In the case of the first presidential campaign debate back in 10/2004, I found that the duration of Bush's pauses was not significantly correlated with the duration of the immediately following spoken phrases (r=-0.05), but did correlate with the duration of the immediately preceding phrase (r=0.54).

With access to a complete set of recordings over time, it would be fairly easy to sample various relevant quantities and rates for different public figures in different sorts of material at different times and to explore various hypotheses about what was going on in their minds as well as in their rhetoric. I'm a bit surprised that more of this has not been done.

[For completeness, here's the whole Q&A where the Bush remarks came from, as transcribed on the web site. Obviously (and appropriately, in my opinion) most of the false starts and self-corrections have been edited out.

Q Concerning the dispatch of self-defense forces to Iraq, the 14th of next month is the time limit of the stationing. What kind of explanation did you make to the President about that? And how did President Bush evaluate that-- appreciate Japan's position on this? And what do you expect Japan to do further in Iraq on this issue?

PRIME MINISTER KOIZUMI: Concerning Japan's assistance toward Iraq, including the activities of the self-defense forces, we will want to see that the Iraqi people, themselves, bring democratic and stable nation by the power of the Iraqis, themselves. And they are making efforts toward that goal. Certainly there are political difficulties, but they are making progress.
So, against that background, as a responsible member of the international community, Japan should seriously consider what we could do to help the situation there. That has been our position, and there is no change in our basic stance.
What kind of assistance we are going to make in December? First, toward the reconstruction of Iraq, what we can do -- that, first, we have to think about, and then multilateral forces and other nations are involved in helping reconstruct Iraq. As a member of the international community, we have to join them. And further, on the basis of the importance of the U.S.-Japan alliance, we have to take all those things in a comprehensive manner, so that we seriously think what we could do to help the Iraq situation, and we make judgment on that basis.

PRESIDENT BUSH: Obviously, the extent to which the Japanese government wants to give reconstruction money to Iraq is up to the Japanese government. And as to the deployment of troops, that's up to the government. That's what happens in democracies -- governments make decisions that they're capable of living with. And that's -- that's what we said, said, do the best you can do; make up your own mind, it's your decision, not mine.


[Update: As an example to show that George W. Bush has often been extremely fluent and clear in presenting complex ideas extemporaneously, listen to this clip from one of his town meetings on Social Security reform (remember those?), held on June 2, 2005, in Hopkinsville KY. I can't provide any valid evidence that the pattern of the Kyoto clip has become stronger and more common, and the pattern of the Hopkinsville clip weaker and less common, but that's my subjective impression. ]

[Update #2: More on this topic here.]

Posted by Mark Liberman at November 17, 2005 08:45 AM