I had occasion to be standing next to a police officer late one night on the weekend. Outside a motorcycle backfired as it drove by. The officer then said into his radio: "Possible shots fired, Wickenden and Brook. Harley backfiring. [pause] That's right, a Harley backfiring."
Allan argues that this can't be an epistemic modality, since the policeman knows that it's just a motorcycle, and observes "Surely we ought to parse the officer's remark as: 'A sound that could be mistaken for shots bring fired, but which is in fact the sound of a motorcycle backfiring, could be heard a moment ago at Wickenden and Brook.'"
He's correct about the reading--and the dull truth is probably that the police use "possible shots" as a term of art for "sounds that could be mistaken for shots"--but there is a reading on which the modality is epistemic. Namely, that the modality refers to what is possible, given what someone than the other speaker knows.
This would be an especially interesting case, because the people whose knowledge is relevant are unknown to the speaker, and might not even exist. I think that those people would be, any police officers who learn of the sounds (either by hearing them or by having a concerned citizen report them). For those officers, the shots are indeed possible shots, and Allan's policeman is telling the dispatcher that those shots that are possible for the officer calling them in weren't really shots.
(Brian has a lot more about epistemic modals; this post seems especially relevant.)
Posted by Matt Weiner at March 10, 2004 07:12 PM