View source for Talk:conditionals

<div>Posted by [http://www.lojban.org/tiki/user11 pycyn] on Sat 10 of July, 2004 19:55 GMT posts: 2388</div><div>
Apologies. The suggestion for subjunctive conditionals: {da'i} ... i ... ({da'inai}, will not work and a major reasomn is only a paragraph earlier: there is no way to negate such a construction, as "might if" conditionals require — not to mention just ordinary use. So, we need two things only: an afterthought connective and the corresponding forethought. BUT this connective is liable to a lot of use, so should certainly not be longer than the material conditional {na'a}. And there is no such word availablr don't think we need to experiment with this so I exclude {VV}). Since lerfu shifts are up for discussion, can we discuss getting rid of some of them and recycling the words? (a remark at the beginning of a passage explaining what the alternates will be — maybe even three or four at a time — would seem adequate for any purpose I can think of.)

The probability notions of conditionals get short shrift in th discussion. The useful ones are pretty simple (apparently), since we have the notion of conditonal probability available. We can modify (if need be) the notion of probability from descriptive distribution in history to descriptive distribution over worlds to get a broader concept that includes unrealized cases. And we can add an acceptability function (defaulting to .5+d, but adaptable to situations, e.g. "significant"). With a modified acceptance function, we can get another shot at "generally," now as = "probably:" "Lions are generally vicious" = "For all x, probably (if x is a lion, then x is vicious) (where "probably if ... then ..." is a single operator, a special consitional.
</div>