[YG Conlang Archives] > [jboske group] > messages [Date Index] [Thread Index] >
Lojbab: > At 03:46 AM 6/3/03 +0100, And Rosta wrote: > > > I disagree, primarily because of by experience with Russian. By English > > > standards, parts of Russian seems terribly long-winded and non-Zipfean: > > > their word for "use" is "izpolzovat" (sp?), their word for "artistic" is 5 > > > syllables, they express dates in years the LOOOONg way: one thousand, nine > > > hundred, ninety seven. They still seem to use these quite a lot despite > > > their length. On the other hand, they encode a lot of tense/aspect > > > information in their perfective system, which English expresses only > > > cumbersomely with extra words. Still English isn't tenseless and uses > > > those words, longwinded though they may be > > > >What is the motivation for 'Zipf' if not practicality? > > For Lojban design: because JCB said so > > In real life, I think Zipf is overblown. In "high social registers" > verbosity is considered a virtue; talking in monosyllables isn't going to > win any respect Just because there are countervailing factors to Zipf doesn't mean that Zipf is overblown. Counterzipfean tendencies are explicable by other principles (e.g. (i) to say things more lengthily is to make more effort, and to make more effort is to be polite; (ii) to say things more lengthily can have incidental but valued poetic effects; (iii) ordinarily the Gricean maxims require us to use no more words than we need to to express our meaning, so if we use extra words then the addressee may, in the absence of other evidence, infer that we have extra meaning; etc.).. > >I wonder whether "izpolzovat" is as frequent as "use" > > It would seem so, and there are other words and phrases of significant > length that are common as well. For example, that which we call WW II in > Russian is called such things as the "Great Patriotic War" only taking > twice as many syllables as the English - it seemed to be talked about just > as much despite the long form It is obvious that average word length varies across languages, but I presume you are making a stronger claim that in Russian there is no discernable tendency to shorten longer high frequency words? I am skeptical about that, but I have no way of knowing if it's true. Still, cross-linguistically the tendency seems to be so glaringly obvious that it will take more than the citation of apparent counterexamples to get me to revise my views. > >Also, in saying that "the practicality of an expression partly involves how > >longwinded it is", I also have in mind that the speaker makes a > >calculation of an expression's cost in terms of effort to the speaker > >and hearer and benefit in terms of communicating information > > I don't think that speakers make such calculations. Well linguists do think so. And that is all I'm prepared to say on the matter: if you want to disagree & be heeded by me, then you will have to read up. > I think that language > AS A SYSTEM, undergoes some optimization for efficiency in communicating > information, wherein we can trade precision and artistry (and formalism) > for brevity when speed is important. At first glance, that sounds the sort of thing that might be true, but in fact I can't think of any obvious examples except for ellipsis. > If Lojban is to value precision and > formalism, it WILL have greater length I don't know exactly what point you're making. Lojban is a longwinded language, partly because no effort was made to design it to be otherwise. My point -- which you dispute in the face of abundant evidence from both lojban usage & natural language -- is that Lojban speakers will be precise and explicit partly in proportion to how longwinded the lojban is. Hence the design matters. A distinction coded in the gadri system will be used a lot. A distinction that requires the use of, say, MOI will be used much less. So if you're deciding whether to encode a distinction in LE or in MOI you know from the off that this is going to have big ramifications on how much the distinction actually gets used. > >For > >any bit of information (by which I mean to include small bits, such > >as whether a reference is singular or plural), there will be a degree > >of cumbersomeness beyond which the informativeness is not worth the > >effort > > But, if one studies for example 19th century conversational styles, one > would find that the degree of cumbersomeness changes with the social > context. It is the peculiarities of the modern American (and perhaps > British) English social contexts that are placing great value on > brevity. I address this in the first para of this reply. > But I can't imagine that people who want to learn Lojban do so > for its brevity No indeed, and as you know, that's one reason why I think Lojban cannot suffice as a practical logical language. But that is not the point under discussion. The point under discussion is the one I reiterated in the para before last. > >I would have thought this point was blindingly obvious, were > >it not that so many Lojbanists seem oblivious to it > > > >You can see this point illustrated today on Lojban-list, by Pierre > >and Craig discussing ways to avoid zo'u, with Craig suggesting that > >the effort of using zo'u is not worth the benefit of saying what > >one means (in casual speech) > > Yep. That is indeed a tradeoff. Lojban allows sloppiness. It is the job > of the teachers to make it clear that it is sloppy. After all, in English > we don't see breeds of dog either, > "we see 3 dogs, each representative of a breed which is distinct from the > breeds of the others" It is not certain that English "I saw three kinds of dog(s)" does not encode the meaning that Pierre intended. > > > >Furthermore, > > > >we regularly come up against stuff that is easy to say in English but > > > >that nobody can find a way to say in Lojban > > > > > > I think that this is partly lack of fluency. I think they can be said in > > > Lojban, but we haven't thought things through always > > > >I hope Nick will give this one of his splendidly scornful tirades > > > >If I may echo Jordan's rhetoric, this notion of "lack of fluency" is > >a load of shite. Sure we lack fluency. But there is nothing there to > >be fluent in. If they can be said in Lojban, the bits of Lojban they > >can be said in are the bits of Lojban that haven't been created yet > > Correct. And once we agree upon a pattern for expressing that sort of > thing, fluent language habits would spread its usage to all manner of like > situations. I contend that the reason we don't agree is because there is > too little fluency, too little experience trying to communicate. If we had > gained that experience, we would be able to agree, and likely there would > be somewhat less contention about how to say it either formally or sloppily I invite you to consider the fact that those Lojbanists who have invested effort into travelling the road towards fluency fall into two groups. Those who don't give a shit about sloppiness. And those who throw up their hands in frustration when they find out how flimsy and rudimentary the prescription is. > >So when I say "nobody can find a way to say in Lojban", I mean > >"nobody can find a way to say in the bits of Lojban that have been > >created so far" > > The solution to that is xod's and Jorge's: start creating more Lojban Come again? This is *Lojbab* speaking? Or only his evil twin? Unfortunately the broader community rejects this solution, so we cannot embrace it, though you know I would have dearly loved to. > > > I contend for example that all of the nuance that people are trying for in > > > the gadri are NOT especially expressible in English > > > >Different grammarians have different views on the extent of polysemy > >in grammatical categories. Generally, the more formalist one is by > >persuasion, the more one favours monosemy, and on that view it is > >the case that not all the nuances are encoded in English. But we > >do conceptually distinguish the nuances when communicating in English > >Since Lojban won't tolerate polysemy, then Lojban must either find > >different ways of encoding each nuance, or leave some nuances > >unencoded > > So long as the cmavo are a closed set, every word amongst them will cover a > range of meaning and hence be to some degree polysemous John corrected this in his reply, so I won't repeat it. > But I don't think we really disagree, except that I think we can encode all > nuances if we so choose We can't do it with existing gadri. As for whether we can do it with new gadri, that depends on how the BF vote and then on how the wider community vote. > >(Personally I don't think it vital that every nuance be encoded. But > >I do think that the most important distinctions should be encoded, > >and that that is not what the current gadri system does.) > > I don't think we know or agree on what are the most important > distinctions. It is that disagreement that led in 1994 to the current > formulation which I can defend only in saying that "it is baselined", but > never approved of Our knowledge is always incomplete and susceptible to revision, but if we take that for granted, then we do know what are the most important distinctions, and we almost all agree. The more a person understands the issues, the more they tend to agree with others who understand them. Evidence for all these claims was apparent in the great gadri debate of 02. > Meanwhile I note that TLI and JCB never seemed to be bothered with only > having around half the gadri that we have and using them even more sloppily How do we know how sloppily they use them in comparison to Lojban? Also, I can imagine how having fewer gadri might yield fewer usage errors. --And.