From xod@thestonecutters.net Wed Oct 02 09:42:40 2002 Return-Path: X-Sender: lojban-out@lojban.org X-Apparently-To: lojban@yahoogroups.com Received: (EGP: mail-8_1_1_4); 2 Oct 2002 16:42:40 -0000 Received: (qmail 37729 invoked from network); 2 Oct 2002 16:42:39 -0000 Received: from unknown (66.218.66.217) by m8.grp.scd.yahoo.com with QMQP; 2 Oct 2002 16:42:39 -0000 Received: from unknown (HELO digitalkingdom.org) (204.152.186.175) by mta2.grp.scd.yahoo.com with SMTP; 2 Oct 2002 16:42:39 -0000 Received: from lojban-out by digitalkingdom.org with local (Exim 4.05) id 17wmdD-0004yL-00 for lojban@yahoogroups.com; Wed, 02 Oct 2002 09:45:51 -0700 Received: from digitalkingdom.org ([204.152.186.175] helo=chain) by digitalkingdom.org with esmtp (Exim 4.05) id 17wmcd-0004y0-00; Wed, 02 Oct 2002 09:45:15 -0700 Received: with ECARTIS (v1.0.0; list lojban-list); Wed, 02 Oct 2002 09:45:14 -0700 (PDT) Received: from [66.111.194.10] (helo=granite.thestonecutters.net) by digitalkingdom.org with esmtp (Exim 4.05) id 17wmcY-0004xq-00 for lojban-list@lojban.org; Wed, 02 Oct 2002 09:45:11 -0700 Received: from localhost (xod@localhost) by granite.thestonecutters.net (8.11.6/8.11.6) with ESMTP id g92Gfvh91231 for ; Wed, 2 Oct 2002 12:41:57 -0400 (EDT) (envelope-from xod@thestonecutters.net) Date: Wed, 2 Oct 2002 12:41:56 -0400 (EDT) To: lojban-list@lojban.org Subject: [lojban] Re: ka ka (was: Context Leapers) In-Reply-To: Message-ID: <20021002120303.G90967-100000@granite.thestonecutters.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 1823 X-ecartis-version: Ecartis v1.0.0 Sender: lojban-list-bounce@lojban.org Errors-to: lojban-list-bounce@lojban.org X-original-sender: xod@thestonecutters.net Precedence: bulk X-list: lojban-list From: Invent Yourself Reply-To: xod@thestonecutters.net X-Yahoo-Group-Post: member; u=110189215 X-Yahoo-Profile: throwing_back_the_apple On Mon, 30 Sep 2002 pycyn@aol.com wrote: > In a message dated 9/29/2002 8:29:37 PM Central Daylight Time, > xod@thestonecutters.net writes: > > << > > If you insist that the difference in semantic level between "comes and > > goes" and "changes over time" is comparable to that between "appears blue" > > and "a mixture of gases in such-and-such a ratio", I can only assume > > you're not arguing honestly. > >> > As you well know, that is not what I am arguing, but rather that in > ""Example 5.4 conveys > > that the blueness comes and goes, whereas Example 5.5 conveys that its > > quantity changes over time." Specifically, the "whereas" makes it mean > > ingless because there is no difference between the two clauses. (The high > > quality of the rest of the CLL makes the conceptual chaos of this > > notorious chapter all the more noticeable.)" > blueness (a quality) is different [the] quantity [of blueness] (a quantity, > duh). So we're back to where I started: "comes and goes", and "quantity changes over time" are two strings of characters that describe the very same reality. They don't even focus on distinct aspects of that phenomemon either. > << > Now that you understand the terminology, perhaps you can go back to my > last two applications of it and make sense of my thoughts on ka. > >> > Presumably about {ka}. OK: > "> ka + ce'u describe tergi'u, not sumti. It is well-defined, whereas your > > usage of ka without any ce'u is ill-defined, very subjective including any > > feelings anyone has about the fact that da is in broda1, and I believe it > > was trounced, a casualty in the last gang bang of ka. It's also been > > abandoned by usage as far as I see, all users now sticking to the doctrine > > that every ka has at least one ce'u, and they write it explicitly." > > The structure {ka + ce'u} is about places in a predicate, not about the noun > phrses that fill them. Nope, no clearer. Sorry, but that English rendering seems perfectly sensible to me. > At a guess you mean that where the > {ce'u} is in the bridi after {ka} tells what place of the relational > predicate is to be occupied by the NP in constructing with that predicate, > but that the {ka} followed by a {ce'u} less bridi is not an obvious extension > of that notion, since the notion gives no clue about what happens once the > {ce'u} slot is filled. This is almost true; as the case of {du'u} shows, we > might expect {ka} with a full bridi to represent a proposition. Sure, but nobody has ever suggested that. > The property-of-event reading is unrelated, but seems to be what the > examples call for. But *which* property? Which property of an event is being selected is well-defined if there is a ce'u in the event sub-bridi (and I will not entertain a tangent on the issues surrounding ka nu ce'u broda), but without a ce'u, any conceivable property might be intended by the speaker. This is, of course, why I already declared ka without ce'u ill-defined. If you still don't get it, then take the bridi "la godziras. cu cadzu", and use ka to extract "earthshaking" from that. > > << > > He's tall, but everyone who calls him tall know there are things taller > > than he is. Thus, "tall" never meant "infinitely tall" and everything > > remains consistent. > > >> > > Yes, there are probably men taller than Kareem, but that doesn't mean that > it > > is more true that they are tall. Suppose there is a fellow, Jake, who is a foot taller than Kareem. It is possible for an observer to define the boolean "tall" as true if the person is over the height of Kareem + .5ft. Now Kareem is short and Jake is tall. And in this way, when moving to a fuzzy system, it is truer that <> than >. > > But that is exactly how I interpret fuzzy logic, and how I use jei. And > it's quite a bit more useful than any competing interpretation or usage. > >> > > Well, the people who have been using fuzzy logic theoretically, and -- more > importantly -- practically, for thirtymyears would disagree. I'll let you know when I run into a problem using jei this way. > << > In any case, if we used a boolean to describe height, we would be > describing only two heights. If we used a three-state variable, we would > describe three heights. It is true that in these two cases, the actual > heights don't need to correlate to the values used for the logical > variables with a simple function. But when we use a real variable for the > logical value, simplicity dictates that there be a simple function between > the height and the logical value. At the very least it dictates that the > function be monotonic, and that's all I need for my original point. > >> > This seems one long and iterated non sequitur. How do we use a boolean to > describe heights and how does doing so describe (only -- but also even) two > heights? I made an error here; each boolean state does not need to refer to a single height. > A boolean is just a member of the set {True, False} (however realized). As > such, it presumably applies to claims or sentences. But what the sentence > describes or how is not affected by the truth value it receives (rather the > other way, I suppose); So, "Kareem is tall," which seems to describe exactly > one height, Kareem's would get T or F. If we move to trivalent logics, the > sentence could take any of three values, but it still describes on height. > And similarly if the logic (whether probability or just infinite valued) > takes any value within [0,1]. I suppose your point is that the thee truth > values tend to divide the continuum up into different segments: two values > divides height into > tall and untall (turn-around for American men somewhere around 5'8" -- the > fact that it is not clear where the break is is what eads to fuzzy logics, > including, as is reasonable here, fuzzy bivalent logics). Three values gives > tall, moderate and short or some such classes (and, again, really wants fuzzy > analogs). With an infinite number of truth values, we can divide height into > an infinite number of classes, but nothing says that this has to be directly > on truth values. I suppose that most non-metric height classifications are > going to take (for American men, again) anything below 1' at least as in the > 0 class and anything over 8' as in the 1 class (and that probably goes lower > in the second case and higher in the first). I don't know why we'd arbitrarily round all heights between 0ft and 1ft to 0 truth value, since we certainly have enough truth values to go around between 0ft and 00ft (that's infinite to you). Introducing discontinuities into a function which symbolizes a continuous one seems uncooperative although legal. > That the function between measured height and truth value (or qunatity, > for that matter) be monotonic seems reasonable, but that only requires > that a greater height not have a lower value, and that is pretty surely > going to be the case for all these system (well, I'm not sure about > quantity, come to think on it). Admittedly, one of the more famous -- > and difficult -- truth assignment systems on [0,1], the percentage of > people who unhesitatingly (we can fuzzy this by taking hesitation into > account) assent to the claim, might very well be non-monotonic, since it > violates every other intuition at some place. But notice that, in any > case, being monotonic doea not mean that 1 correlates to only infinite > height. Indeed, the bivalent system is monotonic, I suppose (though I > do think that factors other than height sometimes enter into the > judgment of tallness -- skinny people are tall shorter than fat people, > for example). Perhaps monotonic or even continuous isn't enough of a claim, but some sort of Grice/Occam's Razor assumption of keeping the truth value/reality mapping as linear as possible; any deviation requires some sort of justification. -- Before Sept. 11 there was not the present excited talk about a strike on Iraq. There is no evidence of any connection between Iraq and that act of terrorism. Why would that event change the situation? -- Howard Zinn