On "Sylvan's Box"
Of Stories and Impossible Things
In discussing the logic of fiction, one absolutely essential article is “Sylvan’s Box: a Short Story and Ten Morals” (Notre Dame Journal of Formal Logic, Vol. 38, No. 4 [Fall 1997], pp. 573-582). This is a story by Graham Priest that he uses explicitly to explore the question of the appropriate logical system for discussing fiction. (You can download it for free through Project Euclid.) In the story, Richard Sylvan (the logician, also known in earlier days as Richard Routley) has recently died, and the narrator, Graham, travels from Canberra to Bungendore to meet with another character, Nick Griffin, to go over Richard’s papers and effects. As Richard Sylvan was a prolific philosopher and writer, there are many papers, but between a pile on Meinongianism and another pile on paraconsistent logic, Graham finds a small box on which Richard has written a label: Impossible Object. Graham opens the box and finds a small, carved wooden statue. It completely changes everything, because the box both contains the statue and doesn’t contain it. Graham and Nick, of course, have to figure out what to do with the Impossible Object. Besides being an excellent tribute to Richard Sylvan, one of the greatest (but often overlooked) logicians of the twentieth century, it’s a charmingly written story.
Priest goes on to draw a set of conclusions from the story. Four are specifically geared to countering assumptions about fiction made by David Lewis and others: (1) The story is coherent despite including, as an essential element, a contradiction; (2) using the principle of charity to try to get rid of the contradiction misunderstands the story; (3) the appropriate logic for reasoning about the story cannot be classical, but must instead be paraconsistent (that is, it has to reject the principle that contradictions imply everything); (4) the story has to be handled as a whole, so the logic cannot be nonadjunctive (that is, you can’t handle the contradiction by rejecting the principle that having A and having B gets you both A and B, which would let you ‘chunk’ and separate different parts of the story so that they don’t interact). All four of these morals are both plausible and highly disputed.1
Priest then discusses six more morals, which have to do with the need for impossible worlds semantics, i.e., a way of handling modal logic that does not just make use of possible worlds but impossible worlds, as well.2
We intend over the next few months to discuss a number of things relevant to all of these issues.
Our own view is that (1) is right, (2) is much more questionable, (3) is certainly right, and (4) is doubtful. But there are arguments that have been made both for and against each.
Our view on these morals is that Priest is right as to the essential point — it is in fact necessary to have an impossible worlds semantics, not merely a possible worlds semantics, not merely for fiction but for a wide variety of other things. Priest, however, is a dialetheist, who holds that there can be true contradictions in the actual world. We are not, and think that Priest often moves too quickly from the need for a paraconsistent logic or an impossible worlds semantics to dialetheism.