In a message dated 2/17/2002 12:05:01 PM Central Standard Time, jjllambias@hotmail.com writes:In Lojban, we say {la djan prenu} or {le mi pendo cu prenu}. True, but the fact that it can be referred to by two apparently unrelated expressions is a vital piece of information, so the difference in the expressions is significant, since they generate different sense with the same reference. <xod: >The fancu2, 3 should refer to the axes or real quantities in question. But >these are sets, not single values. So you say, but you don't follow your prescription. The last time you used {fancu} you wrote: >>fancu ro velvecnu le jei mi tervecnu kei lipa >>For every price, the truth value of whether or not I buy it is unity. You had each price as x2, and truth values as x3. You did not have a set of prices in x2 nor a set of truth values in x3. You were using it (with respect to x2 and x3) the way I say it should be used.> Touche' <>You do want the device for specifying a function to specify the >right function after all and also to be informative. You can do that in more standard ways than having two places for the same argument. Using {po'u} for instance. I take it that using the name in x4 and the more informative _expression_ in x1 of fancu is as acceptable as the other way around?> But {po'u} is a very sloppy way of specifying the function you have just decided to name -- presumably that specification is the central act here. I might think moving fancu 4 to 2 made sense, but not putting it in as an incidental. And thus I don't think that putting the new form in fancu4 and the formula in fancu1 is quite right either (even without the explicit reference to the rule for computing) since it is rhetorically inept. <I take it you mean {lo'i numcu}. {lo'e numcu}, the archetype number, is perfectly fine in my preferred way.> Not obviously. It seems to say only that the typical number (but not others) is mapped, or that numbers are typically, but perhaps not always, mapped. Very inexact. <And who says you can't talk about ranges unless you have a place for the range in the place structure of {fancu}? The difficulties come when you don't have a place for the values. I said in a previous post that {lo'i te fancu} was the range in my interpretation, but that's not necessarily so, it is the image of the domain, which only has to be a subset of the range. How do you talk about the image in your interpretation?> Well, it depends. If it is a nice manageable critter, you can use it as the range, since it is (so, for successor, the range is the positive integers (up to isomorphism)). If it is a gappy, wandering thing, then probably you just take a small manageble superset, since fancu4 will sort it out correctly, and, of course, it is always lo'i te fancu, circularly enough. Or the set lo'i something vo'o (there was a discussion this a while ago, but I forget the niceties of the trick). And, as I said before, ranges are nice in that I may not know the image (sin maps angle measurements from 0 to 90 onto the closed real interval 0-1, but I am damned if I know anything more detailed about that image -- and I may even have that wrong, come to think of it). <Instead of {fy fancu lo'i namcu lo'i namcu} you can say {fy fancu ro namcu pa namcu}. Which also allows you to say {fy fancu li pa li cici} if you need to single out that value. How do you say that with the domain/range definition?> But why would you want to say this last, unless 1 is the only argument this function takes and 33 is its value for that argument? This is what you say using "whizbang" once it has been specified: {li cici uizbangi li pa}, say. Fancu is about the nature of the whole function, not about one case -- cases are what we have the function itself for. |