| We now have a very peculiar situation.
I take it that everyone agrees that for the general case "Only S is P" means "All P is S" and thus does not entail "Some S is P" or even "Something is P". If there is some doubt about this, consider the following. For humans it is universally true that only females are pregnant. So, in particular, it is true that only female inhabitants of the Carmel of Sts Tereesa and Therese are pregnant. But, even though there are female (and only female) inhabitants, it does not follow that any of them are pregnant. Similarly, only female inhabitants of Gethsemani Abbey are pregnant. It does not follow from this that any of these men is pregnant, indeed, from the fact that they are all men it follows that none of them is pregnant. Of course, you could say that it is not true of these groups that only female members are pregnant, but that entails that they are not human, contrary to all the available evidence. However, when the S class gets small enough or specific enough or is mentioned in a certain way (I am unclear just what the condition is here), this rule no longer holds: "only s is P" means something else. I am not perfectly sure what, but it seems to be at least "s is P and nothing different from s is P" which simply adds the questioned conclusion to the general solution (the second half is equivalent to "All P is s"), thus guaranteeing that it does indeed follow. Is there any independent reason for thinking that this addition is needed? Generally (logic at work again) we want "s is P" to work out about the same as "Anything identical to s is P" (with suitable adjustments for cases where s is plural). There is one addition: for a certain class of "s" --hard to specify formally outside a particular formal language -- we want "s is P" to entail "Something is P." Presumably, within Lojban, proper names and {le} _expression_ would be "s"s of this sort. But notice that even for these sort of expressions, we do not want to infer "Something is P" from "All P is s," since that is true even when there are no Ps (the existential import rule -- i.e., that universal have none). My conclusion is that the stange translation of "Only s is P" is an ad hoc rule devised to satisfy an unexamined intuition that it should be so. The examination shows (we have had a lot of these lately -- see the cases of the cognition predicates) a strong implicature has been taken for an entailment, to the considerable confusion of what was once a fairly tidy bit of logic. In short, you all are wrong about English, Spanish, Lojban and, as far as I can tell, Chinese as well. Bite the bullet and get on with it. |