[YG Conlang Archives] > [engelang group] > messages [Date Index] [Thread Index] >
selpa'i, On 01/10/2012 02:04:
Am 01.10.2012 01:42, schrieb And Rosta:selpa'i, On 30/09/2012 14:50:I'm not sure it's a good idea to have tense markers at all in Xorban.I guess, Selpa'i, that you haven't read the whole thread? I can't see how, if you'd read my contributions, you could be writing your message. You write as if you were challenging a consensus that there should be unary operators expressing tense.I think I have to apologize for misrepresenting the intentions of some of Xorban's contributors and I probably wrote this message in the wrong thread. Upon rereading this thread, I realized that you were agreeing with me. My post was the result of many small things I thought I had noticed in seperate threads which to me seemed like Xorban was introducing loads and loads of new words and constructs which in turn seemed to introduce the same sort of chaos that we find in Lojban. If that is not what is actually happening, I'm glad to know that. Maybe it was just an illusion, which is very possible. =)
As Mike pointed out, a lot of the discussion is brainstorming, i.e. just tossing forth ideas for discussion.
I also think the roots could be shorter such that every word/root is monosyllabic.
Do you mean things like "mpl", where the schwas are easily elidable? Or do you mean that there should be a morphology/phonology that guarantees monosyllabic roots?
But that's up to you (pl) to decide.
Not "up to *us* to decide"? You don't want a hand in it? I think you should make criticisms and suggestions.
which some here like to call a failed attempt at a loglang,I don't think anyone here likes to call it that. It certainly does not succeed in being an adequate loglang, but it wasn't really ever an attempt at a loglang, certainly not a wholehearted one.I thought I'd remembered you saying that once.
I might have in a moment of inattentiveness or careless wording.
Of course now you're saying that it wasn't ever trying to be a loglang. Which is of course weird given its name, but I don't know enough of early loglan's history.
I suspect that right from the beginning, being a form of speakable logic *was* one of Loglan's aims, albeit one competing with other aims, and it's surely always been something that has attracted a significant proportion of adherents. But the official primary goal was the Sapir-Whorf test, whose efficacy somehow required fluent speakers, and the founders of Lojban held steadfastly to that goal. On the one hand I greatly rue that steadfast adherence to a goal I find deeply silly, and the consequent repudiation of any serious attempt to create a loglang, but on the other hand I admire the Lojban leaders' dogged and relatively successful pursuit of their avowed aims.
I guess I'm very wrong in finding Xorban not too different from Lojban. I do think gua\spi does better than Lojban in certain regards, for instance in its root word definitions and in how much simpler than Lojban it is.
What are Guaspi's root definitions like? Is there anything from Guaspi that's you'd recommend to Xorban?
Ah I remember now, one thing I didn't like was how you (pl) decided to implement compounds in Xorban. Having no regularity in them and only using the roots for mnemonic purposes seemed like a cheap solution, and less compositional than I would prefer.
It was me who was most insistent about this. If there's a way to create fully compositional compounds, why would we not just count this as syntax? For Xorban, compounds by their very nature are noncompositional -- a mnemonic way of creating new stems.
If the definitions of the roots were adjusted slightly, a compound system akin to the one used in gua\spi would be much more to my liking, and feels more logical to me.
Tell us more -- the kind of system you'd like, and why.
However, I seem to not understand what "logical" even means.
Different things to different people, but for my part I try to avoid using it in any sort of conlang-specific meaning. (But I use "loglang" with a specific meaning.)
I've enumerated the faults of Lojban qua loglang so many times that I can't really muster the energy to recapitulate.it is extremely verbose, able to avoid verbosity only by not encoding important information;Yes and no, I'd say. I wouldn't call it extremely verbose. A Lojban text is almost exactly as long as the same text in English, depending on the subject matter. Most of Lojban's verbosity comes from it's long words and all the additional words that carry no meaning, e.g. terminators (which gua\spi did away with) and a handful of other cmavo, as well as gismu being bisyllabic, which could also have been avoided.
Yes, and brevity is avoided by means of throwing away information -- mainly by having an implicit "zo'e" instead of "su'o da xi mu".
its so-called grammar is a work of utter fatuousness, only exacerbated by it being labelled a 'grammar';Sure, it allows you to say nonsense. It's not a grammar in the classical linguistics sense. But at the same time, that's what makes Lojban a nice experiment. Somethings might turn out to make sense in the future that now seem nonsensical. I take it you would prefer things that make no sense to be ungrammatical.
Not really. The grammar generates a set of sentences. A sentence is a pairing of a phonological form, interpreted phonetically, with a syntactic form, interpreted semantically. The Lojban pseudogrammar generates structured phonological forms, not sentences. I'll say more about this in another message.
Isn't that possible in only two ways? Have more selma'o or write a semantic parser. The former is not desirable I think, the latter just hasn't been done yet.
If by "semantic parser" you mean the rules that convert phonological forms into semantically interpretable logical forms, then yes, those are the rules that would make the grammar complete; but those rules haven't been agreed on, let alone codified, and they're really the core of the language.
it has no coherent, comprehensive and codified rules for defining correspondences between phonological forms and logical forms,I'm not sure what this business about phonological forms is about. Would you mind giving an example?
The phonological form /su'o da broda/ corresponds to the logical form "Ex broda(x)". A phonological form is pronounceable.
and to the extent that some conventions have emerged to try to compensate for that lack, this has been achieved by Jorge's brilliance and perseverance in the face of dogged opposition from the mainstream community and its leadership.Defintely, no argument here. And I am certain that Lojban will continue to improve in this direction, getting simpler and simpler and removing all those stupid little things you mentioned above. We'll have only one TAG selma'o, cmevla and brivla might merge merge, I'm certainly for it, and using {cu} already, and xorlo will become the standard throughout (there are still a surprising number of people who are opposed to it, but give it some time). I completely agree that the leadership of Lojban is scared of improvements, very much so in fact, but it can't and won't stay that way forever. These conservatives are a dying species, even if you and I won't be around to seeing it happen (though I think it won't take that long).
Wow. I do see a cultural shift among Lojbanists -- am not infrequently astonished by it indeed; but I still see plenty of conservatives who don't want to unlearn or discard anything they already know. Surely it's far easier and quicker to start again, e.g. with Xorban,than to go through a painful process of rebuilding Lojban?
Pretty much everybody in the Lojban community has always been very clever -- the language is not at all the product of human idiocy. There are three main reasons why it falls so short as a loglang. The first is that it wasn't fundamentally and wholeheartedly designed to be a loglang.=). I'm probably being annoying, but what *is* and what *isn't* a loglang? What are the objective criteria?
For me, it's an ergonomic language that unambiguously encodes logical forms, which are structures of predicate--argument and operator--variable relations. For others, it also involves logical, e.g. model-theoretic, semantic interpretation rules.
Unfortunately, I wasn't around to see the beginnings of Xorban, so I don't even know the motivation behind the current system, in particular the formula + variable idea. It seems a lot like FA tags to me, and the rest is just a simplified Lojban (I'm oversimplifying her). But maybe what I call Lojban is not exactly what you have in mind when you criticize it.
I see cosmetic similarities to Lojban, because it was Jorge who came up with the "skin" (the look and feel) and he evidently has a fondness for Lojban. It was also Jorge suggested using vowels for variables -- that was the germ of Xorban. Apart from that, the rest of Xorban is pretty much just the simplest possible syntax for encoding logical formulas -- probably lots of us who have thought about loglangs over the years have independently converged on something pretty close to Xorban. (That doesn't mean we all think it's optimal. I think it's not ergonomic enough, but compared to my more ergonomic alternative, Xorban is far more suitable to being understood and learnt easily, so is arguably more suited to being an auxlangy loglang.) --And.