December 27, 2006

A participle too far?

At the fourth annual Language Log Christmas party, someone seems to have decanted an extra bottle or two of rum into the eggnog, and so we're still catching up on pre-Christmas news. One item that shouldn't pass unnoticed was the introduction of part-of-speech terminology into the national dialogue on Iraq.

You can read the whole story in Sheryl Gay Stolberg's NYT weblog post "A Choice of Tenses in War of Words" (12/18/2006). Tony Snow, the White House press secretary, was trying not to admit that Colin Powell's statement that we're losing contradicts President Bush's statement that we're winning. (It's hard out there for a press secretary.) Stollberg describes the outcome this way:

But as to what the president believes about the present tense question — winning or losing? – Mr. Snow stammered a bit, then suggested this was a matter better left to grammarians than press secretaries.

“It’s one of those things,’’ he said, “where you end up – it all ends up trying to – you’re trying to summarize a complex situation with a single word or gerund, or even a participle.’’

It's not surprising that Mr. Snow is a bit confused about the terminology here. The Cambridge Grammar of the English Language has a section (p. 82) with the heading "A distinction between gerund and present participle can't be sustained". Some highlights:

Historically the gerund and present participle of traditional grammar have different sources, but in Modern English the forms are identical. No verb shows any difference in form ..., not even be. [Thus] we reject an analysis that has gerund and present participle as different forms syncretised throughout the class of verbs. We have therefore just one inflectional form of the verb marked by the -ing suffix; we label it with the compound term 'gerund-participle' ..., as there is no reason to give priority to one or the other of the traditional terms. [...] This grammar also takes the view that even from the point of view of syntax (as opposed to inflection) the distinction between gerund and present participle is not viable, and we will therefore also not talk of gerund and present participle constructions [...].

Tony Snow is not the first right-of-center flack to take up this issue -- I covered William Safire's encounter with gerundology back in 2004 ("To pass into a certain condition, chiefly implying deterioration", 6/30/2004).

But to me -- I'm a phonetician by trade -- the most striking part of the December 18 briefing was not the syntax but the stammering. Tony Snow is a professional broadcaster, who makes his living by talking, and normally exhibits a high level of verbal facility. In other words, the man is stone glib. But the cognitive stress of his Dec. 18 briefing elicited some truly spectacular disfluencies.

Since official transcripts (sensibly) omit most stutters, false starts, and filled pauses, I've provided an unoffical transcript of a sample passage below -- along with a couple of audio clips where a disfluent passage is repeated so that you can hear it clearly.

Q: Can I just come back to Powell one more time? Just to be clear, one of the points of disagreement, we are losing, you disagree with that?
MR SNOW: Again, the President has said before that we are winning.
It- it- look- what- pre- it- what- Colin Powell is saying, we're not winning, so therefore we must be losing,
and then he says, all is not lost.
So I'm just- I'm not gonna- what I am saying is
that we will win and we have to win,

and that's- that's the most important- that's the most im- hah?

Q: You're not disagreeing with him?
MR. SNOW: I'm just- -- I'm not playing the game anymore.
It's just- it's one of these things where - you end up-
y- it- it all ends up trying to be- (( b- )) i- y- you trying to summarize a complex situation with a single word
or gerund,
and uh-
or even a participle.

An audio clip of a larger portion of the exchange is here. Here's the official transcript (with video link that should start you out in the right place -- the White House transcript page gives a link to a video of the whole briefing).

I'm not trying to pick on Tony Snow here. At least not much. The fact is, everybody does this kind of thing, when the communicative terrain gets rough. And the linguistic phenomena of disfluency deserve (and have gotten) deeper study. We can learn a lot about the psychology of speech production in general, and perhaps also learn something about the beliefs and motivations of individual speakers on particular occasions. We still don't know enough, however, for automatic speech recognition to transcribe such disfluencies accurately, or even to add them accurately to a conventional edited transcript from which they've mostly been omitted. (This is a problem that I've worked on a bit, and plan to return to in future research.)

[Additional commentary by Jon Stewart is here.]

Posted by Mark Liberman at December 27, 2006 09:06 AM