On Wed, 15 Jun 2011, Pavitra wrote: > On 06/15/2011 05:53 PM, Pavitra wrote: > > This is not at all obvious to me. Backwards-disclaimers *might* work if > > the disclaimer was in the same atomic message as the disclaimed text, > > but in the case of any kind of confusion, I would strongly lean towards > > evaluating in chronological order, with the context of each statement > > determined solely by its predecessors. > > I've just come across CFJ 1971, which indicates that a disclaimer can > affect earlier parts of the same (atomic?) message. Hrm.
It's probably a mistake to consider "qualifiers" within an atomic message (e.g. post-action disclaimers) as ordered actions, but rather action + context. Context within a message doesn't have to be "in-order" as long as the reordering doesn't create a paradox or self-contradiction. For example, a past CFJ found that the context could be in the message envelope (e.g. subject line or headers) if the message body explicitly pointed out that the headers contained relevant context - this didn't take "ordering" into account. If split across multiple messages, then, the question is whether the context is "close enough" to apply to the message. This set of atomics should clearly work: [The following action is invalid] [I register, subject to the context of the earlier message]. The backwards version shouldn't work as you point out: [I register] [The previous message is invalid] This version is very questionable: [I register, subject to a context to come in a later message] [The previous message is invalid]. If it works, at the very least the time of the action would be the second atomic message not the first. Your version (where the action is after the context, but the action doesn't explicitly address the context) is probably reasonable along the lines of your earlier argument but just worthy of clarification. -G.
