[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Lojban is *NOT* broken! Stop saying that! (was Re: [lojban] Re: Vote for the Future Global Language)

And Rosta wrote:

Bob LeChevalier, President and Founder - LLG, On 09/01/2011 11:39:


And Rosta wrote:


Natural languages are defined by what their speakers know (or do).
An invented language may be defined either (A) explicitly, by means
of formal grammars and suchlike, or (B), like a natural language,
by what their speakers know (or do).


[...]


The impressive thing is the vitality of the user community, and the
amount of labour folk have invested in it, not the specification.
It would be easy -- with the benefit of experience -- to improve on
the specification enormously, i.e. easy to design a language better
in every conceivable way. But it would be nigh-on impossible to
achieve a lojban-scale user-community for it.



But of course it would be impossible to actually complete such an
"improved specification" BECAUSE the user community wouldn't be there to
complete it. We've had a hard enough job completing (A) even with the
extremely robust user community and 20-odd years to play the game.
Doing a better (A) is likely no longer "easy", because Lojban's standard
for (A) is a pretty high bar to top.
I completely disagree. The Lojban design embodies several decades' worth of design ideas, all of which are too entrenched to change. Anybody doing the language over from scratch can keep the bits they can't improve on and replace the bits they can. Neither of us would claim that the language remains in its current state because its design cannot be improved on. (By "its design cannot be improved on" I mean "it is not possible to find a design that better satisfies the design goals of a logical language".) 

Perhaps by (A) you are thinking of CLL and its thickness,
That is the established standard for "complete specification" (being the only example thereof %^)
I'm sure lots of people could devise "improved Lojbans" if all they had to do was replicate Esperanto's 16 rules, or present a YACC or PEG grammar. But your complaints include a lack of semantics specification, and I would contend that such an "improved Lojban" would thus probably have to be MORE than CLL, having the equivalent of CLLs content as well as the missing specification. 

> whereas I am
thinking more of the design itself. I certainly agree that it would be hard to replicate the monumental effort that led to CLL; it's the underlying design that I think could so easily be improved on.
CLL *is* the design. You may be envisioning some sort of ideal as being the true design, represented in text by CLL, but ideals aren't much use in the real world if not communicable (at least not in the matter of language).
And it would be possible to come up with a design so much simpler that the documentation required for it could be much less. (Actually a lot of CLL is instructional rather than specificational, so its thickness is a bit of a red-herring.)
I don't disagree, but simplification of the design isn't necessarily easy. And writing more briefly takes quite a lot of talent (as anyone who reads my prolificity knows too well).
What is forgotten is that CLL itself is the product of the community. John Cowan started writing individual papers that eventually became chapters in 1993 (and I had already written the Negation paper and some additional material before then). That stuff was hammered on by the community (especially by xorxes) for a couple of years before the final writing.
If someone were to write a new specification with significant redesign (as opposed to the tuning that CLL2 will presumably consist of), whether simpler, shorter, or whatever, it would still need the communal review that CLL had. And the experience of byfy deliberations shows that even a community as robust as Lojban's doesn't accomplish such review quickly nor necessarily with much consensus. 


We have already won that prize: Lojban is the most precisely,
formally specified language that there is, for any language with its
number of speakers or higher. Period. I challenge anyone to find
anything even *remotely close* to the CLL in terms of covering every
*possible* grammatical combination. Even if you can find such a
thing, the formal grammar takes it so far ahead of everything else
they can't possibly hope to catch up.



which is an ahistorical claim. To say that "far ahead of everything
else" is not "excellence" seems to be trying to define "excellence" not
only ahistorically, but unrealistically.
I'm not sure, from your comments, if you understood the points I was making. To restate them, in elaborated form:
1. The building of the Lojban speech community is a huge achievement. The building of the speech community necessitated design-by-committee, which also was a huge achievement. Of all languages with a comparable or larger user community, Lojban comes closest to being a logical language.
But the design-by-committee necessitated the building of a robust user community, as well.
I started with the product of such a community over 20+ years under JCB and 10+ under pc (JCB made sure he got all the credit, but pc was the one who edited The Loglanist and thus significantly led the community). The design improvements we made were pretty much ALL fashioned by some sort of community effort. God knows *I* certainly didn't have the linguistics knowledge to do it - I just did the spec writing until Cowan took over. To get to CLL probably took at least 3 dozen people I could name, on top of the work that had gone into TLI Loglan (and pc could perhaps name dozens whose ideas had gone into that version).
I contend that any non-trivial attempt to redesign Lojban would similarly requires dozens of people, or the result would not even be as good as Lojban.
2. Technology moves on, as each generation of designs improve on the last. At one time, the Lojban design was cutting-edge techology, but nowadays we have learned enough to do much better. Lojban is rather like the qwerty keyboard layout; the mere fact that the design can be substantially improved upon is not enough to make a significant number of folk switch to an alternative (and certainly not to make them all switch to the same alternative, which fact makes it so much harder to decide to switch).
What is missing from this analogy is that significant advances in technology tend to require increasing investments in R&D by ever larger teams of developers in order to achieve them. 


In short, until someone actually DOES produce a demonstrably more
excellent (A), claims that it would be "easy" to do so are indeed both
hollow and insulting.
The claim that the design of Lojban could be substantially improved upon is so transparently true
It isn't to me. I can imagine that ideas for improving elements of Lojban may be easy to devise. But producing a language design is a lot more than thinking up new ideas. 

Esperanto's "16 rules" is NOT a language design.

>that it is completely unreasonable to find it
hollow or insulting. Indeed, the idea that, given the goal of creating a speakable predicate logic,
Actually, I suspect that writing a book that *completely* specifies formulaic predicate logic (without making it speakable) would be non-trivial, based on what pc has said in the past about how many areas of the subject remain contentious after decades or even hundreds of years. 

> the design of Lojban could be thought a

near-optimal solution strikes me as near lunacy.


I didn't say it was near-optimal in the purist or idealistic sense.
It *may* be near-optimal in the pragmatic sense that the collective investment needed to do a *significantly* better job may exceed what is plausible given human capacities (and interest levels).
This does not mean that slow evolutionary improvement isn't possible. But CLL2 is just such evolutionary, and will have taken 14 years by a supposedly robust community, if it is finished this year. It isn't a redesign, and it is getting done largely because "improvement" has been largely unwelcome.
If you were thinking that I was claiming that it would be easy to write another CLL, then I agree that would have been hollow and insulting.
I repeat. Esperanto's "16 rules" is NOT a language design. You keep talking about it being easy to produce a "better language design", but no one has a clue of how a language design could be "better" without being as well-specified as CLL (if not better specified, since much of the complaints going into CLL2 stem from the existing versions lack of specification of some things. And you seem to require an additional layer of specification on top of what is planned.) 


The language itself could not
have been substantially improved without great detriment to the user-community. 

If you invent a "better language" but no one can/will use it, I contend
that it is not a "better language".
I expect we can agree that the success of designs for things can be assessed both by how well they achieve their design goals (e.g. fuel efficiency for an engine, etc.) and by how much they get used. 

If a language is not used, it has not achieved the goal of being a language.

> By the
one sort of criterion, Dvorak is more successful than Qwerty; by the other sort of criterion, Qwerty is more successful than Dvorak.
Dvorak isn't a language, and therefore has lower standards to meet its design goals, but in any event, it IS used. But Dvorak has never been "more successful" than QWERTY, even if it might be a "better" design.
Just as we should not say that Lojban is more successful than Esperanto, even if we think it has a better design. 

lojbab

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to lojban@googlegroups.com.
To unsubscribe from this group, send email to lojban+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban?hl=en.