August 11, 2004

On Condoleezza Rice Confirming Disclosure of Pakistani's Identity

The exchange below took place Sunday, August 8, in a CNN interview between Wolf Blitzer and Condoleezza Rice:
  BLITZER: Let's talk about some of the
  people who have been picked up, mostly in
  Pakistan, over the last few weeks. In
  mid-July, Muhammad Naeem Noor Khan. There is
  some suggestion that by releasing his
  identity here in the United States, you
  compromised a Pakistani intelligence sting
  operation, because he was effectively being
  used by the Pakistanis to try to find other
  al Qaeda operatives. Is that true?

  RICE: Well, I don't know what might have
  been going on in Pakistan. I will say this,
  that we did not, of course, publicly disclose
  his name. One of them...

  BLITZER: He was disclosed in Washington on

  RICE: On background. And the problem is...

According to an article in yesterday's Boston Globe, "Blitzer said this exchange meant Rice had confirmed that the administration released Khan's name to a reporter on background". A government spokesman claims, to the contrary, that Rice's repetition was not a confirmation. But his position is, as my dad would say, baloney.

From the Boston Globe article:

  But Sean McCormack, a National Security
  Council spokesman, said yesterday that Rice
  did not say the leak came from American
  officials.  ''She was in the middle of making
  a point and he interrupted her, and she
  reflexively repeated 'on background,' but she
  was not confirming it and went on to complete
  her thought," McCormack said.

This strikes me as patently ridiculous.

The notion of a reflexive repetition is one with which I am not familiar, and one that I couldn't find with some Web searching, unless perhaps McCormack is suggesting that Rice is exhibiting immediate echolalia (think Rain Man).

Absent some ill-defined claim about "reflexive" repetition, it seems reasonable to suppose that Rice's utterance was, in fact, serving an intended discourse function of some kind or another. What might that function be?

It's true that under some circumstances (usually with intonation indicating, say, e.g. anger) repetitions can be used to question, negate, or contradict rather than to acknowledge or confirm. Consider:

  1     You're a fool.

  2(a)  I'm a fool??
  (b)  I'm a fool.  I'M a FOOL!!
But the context provides no support at all for this interpretation. We can dismiss 2(a) immediately, assuming the transcriptionist correctly typed a period rather than a question mark at the end of the utterance. What about 2(b) as a model? Might Rice have intended to express a dismissive attitude toward the statement, effected by repeating the phrase "on background" in a scoffing way, with a sort of "yeah, yeah" intonation? The recording of the interview would answer that question definitively, but that intonation seems pretty unlikely to me and so I'll disregard it until someone who's actually seen or heard the interview tells me otherwise.

I've suggested that Rice's utterance is unlikely to be function-free, and also that it's unlikely to have been used to question, negate, or contradict. So what are we left with?

Well, how about Occam's Razor? There is a discourse function which repetitions do serve, and frequently: they are often used to acknowledge a successfully communicated proposition, particularly when selecting between several alternatives. E.g. in

  3(a) I'd like my coffee with skim milk.
   (b) Skim milk.
the repetition of "skim milk" by the hearer confirms that the proposition "the coffee should be served with skim milk" has successfully been communicated, and it emphasizes the contrast with an alternative proposition such as "the coffee should be served without milk" or "the coffee should be served with cream".

With that in mind, let's look at the exchange. Rice says that officials did not "publicly disclose" the name of the informant. In logical terms, Rice has asserted

  (4)  (not (exists x (disclosure(x) and public(x))))
       "There was no public disclosure"
Blitzer is clearly concerned with distinguishing two alternative ways her assertion could be true: she could be saying
  (5)  (not (exists x disclosure(x)))
       "There was no disclosure"
or she could be saying
  (6)  (exists x (disclosure(x) and (not public(x))))
       "There was a disclosure and and it was not public".

As it turns out, journalistic parlance provides a term, on background, that describes the latter case: something stated on background is communicated subject to "an agreement between a journalist and an interviewee that the name of the interviewee will not be quoted in any publication, although the substance of the remarks may be reported" (Webster's). So when Blitzer interrupts with "He was exposed ... on background", he is using mutually familiar terminology with Rice to make the assertion in (6), "There was a disclosure and it was not public". When Rice replies with "On background", the most parsimonious explanation of her repetition is not some unusual (I would say bizarre) claim about reflexive repetition, but rather a straightforward, everyday use of repetition. She is emphasizing a contrast between (6) and another contextually salient alternative,

  (7)  (exists x (disclosure(x) and public(x)))
       "There was a public disclosure".
In the process of emphasizing this contrast, which applies to part of what is being said (public(x) versus (not public(x)), she is acknowledging the other part, namely disclosure(x).

Just for the fun of it, here's the same basic dialogue with the topic changed to coffee as follows:

  coffee  ~ disclosure
  with milk ~ public
  like  ~ have done
  black     ~ on background
    (i.e. black = coffee and not with milk
          on background = disclosure and not public )

We get:

  PERSON A:  I don't like coffee with milk.
  PERSON B:  You like your coffee black.
  PERSON A:  Black.  And finding good coffee is hard these days...

Is there any doubt that Person A has confirmed she likes coffee, as long as the listener understands that it's black coffee (not with milk)? Substitute the original terms back in. Is there any doubt that Rice confirmed the disclosure having taken place, as long as the listener understands it was a disclosure "on background" (not public)?

Now (good transitional cue word!), I should confess that I'm not a specialist in discourse and dialogue. I probably have as much grounding in that sub-field as your average computational linguist. (Perhaps even a little more, having once analyzed Abbott and Costello's Who's on First routine from the perspective of centering theory for a term paper. :-}) I'd be happy to hear from any discourse specialists (perhaps someone who can speak to this from the point of view of Rhetorical Structure Theory?) who would like to weigh in, correct my terminology, or provide a good linguistic argument that I'm wrong.

As for political arguments, I'm not sure what the implications are of Rice having confirmed that Khan's identity was disclosed on background. I would infer that this confirmation was viewed by the Bush administration as a mistake, or else the NSC's spokesman wouldn't have been denying it took place, in the face of evidence obvious to linguists and non-linguists alike. FWIW, I think mistakes are best acknowledged and dealt with, not brazened out. Linguists don't need their own technical term for the denial of a proposition known by the speaker to be true -- it's called lying.

Posted by Philip Resnik at August 11, 2004 11:25 AM | TrackBack