Date: Sat, 24 Jan 1998 04:50:41 -0500 (EST) Message-Id: <199801240950.EAA13745@locke.ccil.org> Reply-To: Steven Belknap Sender: Lojban list From: Steven Belknap Subject: Re: fuzzy bears X-To: And Rosta X-cc: LOJBAN@cuvmb.cc.columbia.edu To: John Cowan In-Reply-To: X-UIDL: 819ddc0f8f99f276100f5d0f013be17a Status: OR X-Mozilla-Status: 0011 Content-Length: 6939 X-From-Space-Date: Mon Jan 26 12:42:28 1998 X-From-Space-Address: - >Steven: >> I believe that natlangs are fuzzy, and I am pleased to know that lojban is >> also fuzzy, as made explicit here in the grammer. With that in mind, I >> propose a formalism for statements where explicit fuzziness is required, >> which would be simpler than using or . This formalism would >> appear to require no change in the grammer, nor any cmavo other than a >> lojban number: >> >> >> >> "One or more of the things which are fuzzily 4/7 really bears creates the >> story." and: >This means that only 4/7 of a bear exists: all beardom amounts to >no more than 4/7 of an individual bear. I would agree with this translation: "One or more of the only four bears which exist wrote the book." but I was not sure that a fractional bear would follow the same rule. There is some ambiguity about what means. This could perhaps be a polar bear or a bearish type of white, or a whitish type of bear, or many other things. Isn't there also some ambiguity in what a is? Isn't this also a metaphor, just as a is a metaphor? If not, why not? Can't apparently nonsensical, but grammatically correct lojban statements acquire meaning through conventional usage? Certainly a is different from a . >You get the meaning you want you must change to: > > Yes, I've used this construct in previous posts. >> >> >> "One or more of the things which are mostly bearish creates the story." > >The lojban is nonsensical. Sure, and the square root of negative one was nonsensical prior to the acknowledgement of complex numbers. >Again, you need to add {ja`a xi} before {so`e}. > >> Perhaps you are misunderstanding my point. I fully understand that this is not a previously explored lojban utterance, and I acknowledge that it is perhaps nonsensical, given current usage. My question is whether or not these types of statements can acquire meaning (perhaps a fuzzy meaning) if they are: 1. grammatically allowed 2. don't overlap with the meanings of other concepts which would reasonably fill that semantic space 3. are consistent with other acquired meanings I am *not* proposing a change to the grammer. On page 131 of the grammer, it is stated that "inner quantifiiers are permitted on "lo" descriptors for consistency's sake, and may occasionally be useful." I am proposing such a use. It is difficult to come up with reasonable alternative uses for these constructs. Actually, I can't think of any at all. Similarly, may be nonsense, in which case, it is an exploitable niche for assigning meanings, which may be shortcuts for longer more explicit utterances. Isn't that how languages develop? >> "One or more specific things each of which I describe as being almost >> entirely bearish creates the story." > >same as above > >> could have many meanings, of course. >> >> It could mean the remaining 400 pounds of bear carcass after a cougar has >> eaten the other 300 pounds, for example, although that would not be a very >> useful interpretation most of the time. > >We haven't yet concluded the lapsed thread on countability. But >at any rate, almost certainly means 4/7 of a >single bear. I don't think so. This is an indefinite description, no? Shouldn't it turn into: ? Isn't this currently nonsensical? If so, good, it is a niche of the language awaiting exploitation. >> It is already clear that one can say: >> >> le ci cribe pu finti le lisri >> >> "The three bears wrote the book." >> >> The semantic space of fractional bears are unassigned in the grammer; >> describing something as a fractional bear would appear to be grammatically >> correct but meaningless. > >But what about the things that quite clearly are fractionable? >We all agree what half an apple is, for example. Or half a dollar. RIght. But there is no conflict. Half an apple is a fuzzy half in conventional english usage. Half a dollar is a crisp half. If you must specify the fuzzyness or crispness in lojban, you can do so by using a longer utterance, such as . Otherwise, it is ambiguous whether or not you are being crisp or fuzzy. >> My proposal fills this semantic space of what a >> fractional bear is by building on the already acknowledged fuzziness of >> lojban utterances and setting a convention by which fractions between 0 and >> 1 when applied to a gismu (for example) are making explicit the fuzzy >> extent of that gismu. > >There already exists a method of doing what you want, with ja`a xi, >and it has the virtue of not being limited to sumti, which your >proposal is. With time, Zipf would shorten this more precise fuzzifier, I would think (and thus there would be some ambiguity as to whether the shorter utterance was fuzzy or crisp). I do not see why my use of numbers could not be extended beyond sumti. >> Although some consider my conceptualization of objects to >> be eccentric, I believe this view more accurately reflects reality than the >> artificial contrivance of an arbitrary threshold above which persons are >> diabetic and below which they are not. > >I agree with you on the absence of definite boundaries, but I am not >sure whether we could ever agree on meaningful scales of x-ness. >Rather, {ja`a} means "above a contextually relevant threshold" >and {na} means "below a contextually relevant threshold". > >If you really care so much about making fuzz explicit, then I >suggest using ja`a xi all the time. Lojban has been designed with >certain prejudices making somethings shortwinded and other things >longwinded. But it is the responsibilitity of the the users to >ignore this, and be exactly as longwinded as we need to be. >This would make it apparent where the prejudices in the design >erred in making the wrong things longwinded and shortwinded. >Then, once enough people get fed up of having to be so longwinded, >then there'll be enough of a mood for change, and profligately >wasted cmavo can be reassigned to better uses for things it >takes too long to say. I believe that my suggestion is consistent with the approach you suggest. But I don't want to suggest a change to the grammer, so I am using an approach which is both shorter than and doesn't require the reassignment of cmavo, or other changes in the grammer. If the statement in the grammer that sumti are fuzzy is to be taken seriously, then quantifying the fuzziness directly seems eminently reasonable, don't you think? The default is fuzzy, so the are fuzzy bears, and the numbers are fuzzy by default as well. -Steven Steven Belknap, M.D. Assistant Professor of Clinical Pharmacology and Medicine University of Illinois College of Medicine at Peoria