| In a message dated 8/30/2001 8:44:31 PM Central Daylight Time,
a.rosta@ntlworld.com writes: "What I have for dinner depends on what's in the fridge" I am not sure I understand, so much less agree with, the "explicit", or even the normal, versions in this list. But I want to focus on this last case, since it is pretty clearly ambiguous and I want to see whether I have sorted things out aright. First there is a general claim about how meal planning is do, more or less. I'll get back to that. The second is about what for dinner today or what shirt I will wear today. In that case, it seems clear that it le meal/shirt depends on loi fridge content/le weather, with no indirect question games at all. However, the indirect question form turns up again behind this dependency, as a general rule that now gets applied. So, in one sense, the relative clause versions turn out to be instantiations of the indirect question versions. And that looks right: if the indirect question version involves a set of propositions (or whatevers -- see the extensions to at least {ka}) then the relative clause version brings the {makau} down to a single particular answer. So back to the indirect question version. Each indirect question is a set of answers, propositions that match the paradigm forms in meaning even if not in form. What then of "depends on"? It appears to mean a selection of conditional sentences taking (at least) a subset of the crossproduct of the two sets -- but quite a bit more, I think. "Depends on" says that there is an satisfied subset of this sort, such that if P is "What's in the fridge" set and Q is in the "What's for dinner" set then "If P the Q" may be in the set that makes up the dependency -- and will be if it is non-vacuously true. It is important to note that the connection here is "depends on", not "is determined by", so other factors may enter in -- what I had for lunch, what our guest cannot eat,...., and also does not require that what we eat be from what is in the fridge (if the answer about content is "Nothing," then the answer about dinner may be "canned beans" or "take-out" or "eat out" or...., as indeed it may always be if the contents don't suit). This leaves the question of how to say this. The gismu list is not very suggestive of where to start. Working with the way things change, which I thought a good idea a couple of weeks ago, now seems less advantageous than trying to tackle the issue head-on. The changes are in loi content and le meal, not in the general associations (though I suppose that the associations change some. too -- I may not go for dragon and phoenix every time I have the ingredients, even though the former drawbacks no longer apply -- like I just had it for lunch or yesterday's dinner). |