[YG Conlang Archives] > [ceqli group] > messages [Date Index] [Thread Index] >
Rob: > On Tue, Apr 30, 2002 at 01:18:26AM +0100, And Rosta wrote: > > I have the impression that Rex and his fellow collaborators want something > > like a blend of English and Chinese, while preserving the virtues of > > Lojban. What exactly does that mean? Well, for one thing Lojban has a > > lot of obscure complexity that can be dispensed with. For another, > > the goal of brevity and terseness is added to the mix -- the telegraphese > > "Man bites dog" idiom; this goal was never one of Loglan's. > > I had the impression that it was going in that direction for a while, > but no longer is. The guy who had all the expertise in Mandarin is no > longer on the list. > > Ceqli already has brevity from reducing 2-syllable words to 1. More > brevity would require the drastic changes from Lojban grammar that you > mention. I really don't think we can make these drastic changes; too > much work has gone into the Lojban grammar, which is very good except in > a few annoying areas. My view is that Lojban grammar is crude and unnecessarily verbose. The virtue of starting with Lojban is that it is explicitly defined and well-understood, and unambiguous. But it is a long long way from being to any degree a design triumph. > I see changes like making single words for le + > NU, making NIhO and I the same, putting NAI and CAI into UI, fixing the > 'kujoi' mess, getting rid of unnecessary parts like separate words for > modals, etc. (Of course the words and names for the classes would be > different.) Well that is certainly a different and more modest project than what I had thought Ceqli was about. But do let me make it clear, I don't want to steer or influence the direction Ceqli takes -- I have my own conlang where everything works in the way I think things should work. Rather, I'm just interested in observing the preferences and goals of other engelangers, and in advising where I feel I have advice to offer. > > > Though terminators are unnatural, I think the alternative is worse. > > > Requiring every place of every predicate would cancel out the > > > conciseness that Ceqli has. The other alternative, of course, is for the > > > grammar to be ambiguous, in which case I don't think Ceqli would fill > > > any particular need. > > > > This is not the only alternative. A better alternative is to have > > a clause-initial particle that tells you which arguments are present > > and which order they occur in. This has all sorts of added side > > benefits. For one thing, it automatically distinguishes between > > 'nouns', 'verbs' and 'clauses': every clause begins with one of > > these particles, and within the clause -- "man bite dog", say -- > > you know that the 2nd element is the verb because the clause-initial > > particle tells you so. > > So first you diagram the sentence, and then fill it in with the words? I don't understand what you mean. A clause-initial particle has 1-4 complements (1 predicate, 0-3 arguments); it specifies which order they occur in and which are omitted. > (Consider that many sentences may be more complex than 'man bite dog'.) No word would have more than 4 complements. The internal complexity of those complements is invisible to the head word. > That's horrid. What natural language does that? I don't know what you are imagining, but what I am describing is when it occurs in natlangs called 'diathesis' or 'voice' -- devices that indicate the order of arguments and which are implicit. English passive, for instance, converts 1P2 to 2P. > Realize that any feature resembling this would nullify one of the most > significant changes of Ceqli from Lojban, which is that bare predicates > combine into compounds. Or would you have another particle for "the next > compound is three morphemes long"? True. I think that feature of Ceqli is not a good one; it gets in the way of saying "Man bites dog" efficiently. So I'd prefer a different strategy for compounding, such as having a 'hyphen word' (like Lojban zei) or inserting a homorganic nasal before an initial strong (so da + da becomes danda) or suchlike. > And do you really want a conlang designed for the purpose of creating > newspaper headlines? No, you misunderstand. Headlinese is perfectly adequate for the needs of most communication, and if you look at natlangs you will find that, say, determiners (analogous to Lojban LE) are far far from universal. As Rex has (rightly, IMO) said, as much as possible should be optional, the effect being that the obligatory element to a sentence is the barest minimum, which can then be elaborated on by the addition of extra words for extra precision. You yourself should be able to see that the information expressed by Lojban LE could very often be inferred from context, so is often redundant. > > Also, don't underestimate the unnaturalness of terminators. If > > something simply doesn't occur in natural language, it is very > > likely going to be a very problematic feature except for people > > with computer-programmer-like minds. > > ... > > Erroneously omitted terminators are a frequent problem in Lojban > > text. > > I've seen that, when terminators are required in places that make sense, > people don't have a problem using them. They're just like commas, > really. It's the bizarre stuff like requiring 'ku' before 'joi', or > 'boi' between a number and a letter, that throws people. Those are the > minor parts of the grammar that can be fixed. No, people also lose count of how many terminators are needed when there is a lot of embedding. > The situations which require terminators like 'ku' and 'kei' should be > internalized as rules of grammar with a reasonable amount of practice. > Most people would not learn them from 'thinking like a computer > programmer' but just from examples. Maybe, but I am claiming that the mental faculty people would use with terminators is the faculty used by computer programmers and not the language faculty, since this does not come equipped with the necessary ability. An empirical prediction: whereas linguistic ability correlates very little with any kind of intelligence, ability to use terminators would correlate with a certain kind of intelligence. --And.