Leonardo Castro, On 14/11/2012 11:52:
> 2012/11/12 And Rosta<[log in to unmask]>
>> Leonardo Castro, On 12/11/2012 15:08:
>>> If understand it well, Lojban allows us to be as generic as we want. I can
>>> say "cat is black" without specifying if it "a cat", "the cat", "some
>>> cats"
>>> or "all cats". So, the more generic sentences can be short in loglangs
>>> too.
>> Yes, but that's uninteresting.
> Well, that's something I like in Lojban, because I think that this is a
> feature that allows a conlang to be loglangish and auxlangish
> simultaneously, as natlangs differ a lot on their level of mandatory
> specificity. The opposite direction is to convey maximum information in
> minimum phonemes. Maybe Ithkuil is the best example of this type of
> language (I'm not sure because I haven't studied it seriously.)

The issue under discussion -- namely the adequacy criterion for a loglang -- is not one of information density in general. It is about encoding the predicate--argument structure (including operator--variable structure) unambiguously and in full without that extra info making the average loglang sentence longer than the average natlang sentence of equivalent information-density, where equivalence of information density is to be measured without including information about the predicate--argument structure.

An example of encoding predicate--argument structure: Compare (1) & (2)
1. He wants to go. Vuole andare.
2. He wants that he goes. Vuole che va.

(1) unambiguous encodes that the wanter is the goer. (2) doesn't. The loglang must offer a (1)-like solution everywhere arguments have the same value.
> 2012/11/13 Alex Fink<[log in to unmask]>
>> Fair enough.  I don't envy working to meet at once the disparate goals of
>> Logic and Auxiliarity!  (Though to be honest auxlangery leaves me
>> uninterested on the whole, especially where it leaves "be accessible" and
>> enters "be accessible to speakers of Mandarin, English, yadda yadda".)
> And yet this is my challenge. I think these are goals of different worlds:
> to create an auxloglang to be used inside the nerdosphere and to create a
> logauxlang to be easy to be used by anyone. For the latter, I think that
> most logical features should be optional. It's true that even a natlang can
> be as "logical" as we can by adding specifiers, so the main difference of a
> logauxlang would be its regularity and easyness.

What's the rationale for having a logauxlang rather than just an auxlang, and how would a logauxlang differ from an auxlang?

If you really cared about the log bit, but wanted your easy phonology, one simple solution would be to take a simple loglang such as Xorban (which is in fact currently about the only candidate) and just change its phonology and lexical forms.

> I think that it's impossible creating a completely easy phonology for the
> whole world. As I have said before, even native English speakers that are
> accustomed to complex phonotactics have problem with a simple thing as
> pronouncing final /o/ and /e/.

Good observation. CV is (IMO) not the basic syllable structure for English for nonschwa vowels.