'Also there was a man who looked like Jean-Paul Sartre, only uglier, with a limp, and was Jean-Paul Sartre.' That's from Susan Sontag's description of a party in Paris on 19 February 1959, from diary excerpts published in the Guardian (14/9/06). Sontag's sentence raises a linguistic issue that is important in blogs - when and to what degree can one unsay something once one has said it? One aspect of this problem is called defeasibility, the way some categories of statements can be cancelled without contradiction. (One summary is in Stephen Levinson's Pragmatics (Cambridge University Press, 1983). Implicatures can be cancelled. Sontag did not say explicitly that the man wasn't Jean-Paul Sartre, but if we assume she was observing the Maxim of Quantity, then we draw an implicature (that he wasn't) from her statement that he 'looked like' Sartre. She can cancel the implicature, and we (or whoever was supposed to be reading this private diary) won't think she is talking nonsense; we might think that the real Sartre doesn't look quite like what she had imagined.
I noticed this entry because I am particularly interested in how defeasibilty work with verbs of perception. Some of the first eye-witnesses reports of 9/11 to be reported on CNN said they saw a small plane hit the tower. Witnesses of the shooting of what turned out to be an innocent man in the Stockwell underground station in London, after the bombings in July 2005, said they saw a man wearing a padded jacket, and said that they saw him jump the turnstile. Witnesses often get things wrong. Could then then say, 'I saw a little plane hit the tower, but it was a 737' or 'I saw him jump the turnstile, but he didn't'?
The problem is related to the wider issue of self-correction in blogs and other media. Of course newspapers have long had spaces devoted to correcting their errors; there is a web-site devoted to these corrections at Regret the Error. Newspapers correct the next day, or later, often after a complaint, or a word from lawyers. Blogs would seem to have the advantage that they can correct immediately, in the same place as the error, and still show the error, for instance by putting strike-throughs on the now outdated or corrected text. The strike-through reminds us of the provisionality of facts, while the print of newspapers may suggest a definitive account. But I wonder if the provisionality accorded to facts in blogs has its own dangers. A statement can be quoted in other blogs, linked to, commented on, and even if it is taken back an hour after it was posted, it continues to circulate.
But don't quote me, because I'm not sure about this.