The proposed 'goals' of lojban
白松 Oren
out a fun interpretation of the goals listed here:
http://teddyb.org/robin/tiki-index.php?page=Lojban:+You're+Doing+It+Wrong
Rephrased:
* Lojban should be Logical
* Lojban should be a Language
* Being both logical and a language makes Lojban unique
[and then the non-trivial bit]
* Lojban should be backwards compatible
I'm actually confused by the seeming contradiction here:
"[backwards compatible] in the sense that users of the language should
not have to re-learn anything; preserving the meaning of old
utterances is of very little interest to me, obviously"
It seems like formalizing 'versions' of Lojban would only make things
backward compatible in the sense that 'users of a previous version
*would* have to relearn something.'
On another note, I would distill the goals down to this one sentiment:
Lojban: The Logical Language
Mission: Full expressive freedom: Any meaning expressed in a natural
language can be expressed in Lojban, and at any desired level of
ambiguity or precision.
"logical" means no exceptions to rules, and no ambiguous syntax.
"language" means it should be effective in human communication
(clear*, brief and simple)
*This obviates separate definition of 'cultural neutrality,' imo,
which seeks to reduce a type of non-clarity.
co'o mi'e korbi
Kevin Reid
> Mission: Full expressive freedom: Any meaning expressed in a natural
> language can be expressed in Lojban, and at any desired level of
> ambiguity or precision.
We are not interested in supporting ambiguity. We are interested in
supporting vagueness.
(Ambiguity is discrete; vagueness is continuous. Ambiguity means the
listener has to choose among distinct and possibly very dissimilar
meanings; vagueness means the listener gets a broad but contiguous
cluster. Ambiguity, if possible in the language, has to be fixed by
rewriting; vagueness can be fixed by adding elements.)
--
Kevin Reid <http://switchb.org/kpreid/>
Matt Arnold
meet the goals set down in the essay. It does meet them. Of course
nothing is perfect-- nor can it ever be-- but Lojban leaves other
languages in the dust, since they do not even make the attempt.
The reason we need the goals formalized is to guide in clarifying rare
words and edge-cases. Given a set of choices on a specific issue,
these goals act like weights on a scale to influence the decision.
With that in mind, these goals will not cause the BPFK to dither
endlessly, trying to perfectly meet the goals; rather, the goals will
let the BPFK quickly resolve questions for which there are no answers,
the minimum necessary to settle on a release of Lojban version 5.
Goal #1 currently reads: "Lojban should be as logical (internally
consistent, semantically complete, ??? I don't really know how to
define this well) as possible".
My suggestion for goal #1 is as follows: "Minimize rule exceptions and
maximize how predictably the components combine." What did I miss?
-Eppcott
> --
> You received this message because you are subscribed to the Google Groups
> "lojban" group.
> To post to this group, send email to loj...@googlegroups.com.
> To unsubscribe from this group, send email to
> lojban+un...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/lojban?hl=en.
>
>
Oren
Epcott's first three [consisting goal 1, "logical"]:
1.1 Maximize regularity (machine readable)
1.2 Maximize predictability (internal consistency)
1.3 Maximize interoperability (flexible grammar)
Then my attempts at the remaining three (from the article):
[consisting goal 2, "usable"]:
2.1 Minimize semantic ambiguity (formalized definitions)
2.2 Minimize length (brevity)
2.3 Minimize core vocabulary (simplicity)
[consisting goal 3, "unique"]
3.1 Minimize conformity to any particular natural language's
conventions (linguistic cultural neutrality)
3.2 Minimize non-empirical specification (culturally unbiased)
3.3 Minimize cultural context ('cultural neutrality' -- recursive
definition, inherently part of 2.2?)
[consisting goal 4, "backwards compatible"]
4.4 Minimize changes to the language
...I understand now the line about not caring much about backward
compatibility-- it's the last tie-breaker. I agree with it keeping
that position too. =p
co'o mi'e korbi
Oren
neutrality' in any sense not covered by:
> 3.1 Minimize conformity to any particular natural language's
> conventions (linguistic cultural neutrality)
> 3.2 Minimize non-empirical specification (culturally unbiased)
co'o mi'e korbi
Oren
top priority - 'freedom of expression'
0.1 maximize the stuff we can say (testing the Sapir-Whorf hypothesis)
...without which the only optimal solution to our ordered constraints is {}
Christopher Doty
Chris
Matt Arnold
Currently newbies come to Lojban and ask "How do I say this?" and
sometimes the answer is "we don't know yet." We need official rulings
on those questions, and don't have them yet.
This could go on and on forever. When I say "the minimum necessary", I
mean the BPFK will quickly make the minimum number of rulings which
are necessary to declare a new version of the language. I don't want
to see it go on and on forever in pursuit of more questions to give
rulings on, so I specified that I want it to be the minimum necessary
with which to do the things that are on hold waiting for the BPFK to
finish.
-Eppcott
Oren
as a linear programming problem would.
Linear programming takes a function you wish to maximize and then
applies a series of constraints. So the 'edge-cases' are the ones that
find one or more possible eqivalent maxima-- for example: 'this way is
more usable, but this way is how we've been doing it.' The point of
ordering things strictly hierarchically, as proposed by Powell, is to
prevent such 'local maxima' from having ties.
And in my last email, I said the function to maximize is 'what can be
said,' subject to Powell's constraints of remaining
1) logical
2) usable
3) unique
4) backwards-compatible
see the content of my previous email for my suggested precise definitions.
co'o mi'e korbi
And Rosta
> http://teddyb.org/robin/tiki-index.php?page=Lojban:+You're+Doing+It+Wrong
> Rephrased:
> * Lojban should be Logical
> * Lojban should be a Language
> "logical" means no exceptions to rules, and no ambiguous syntax.
> "language" means it should be effective in human communication
> (clear*, brief and simple)
& Matt Arnold writes:
> I hope no one misunderstands, to think that Lojban does not already
> meet the goals set down in the essay. It does meet them. Of course
> nothing is perfect-- nor can it ever be-- but Lojban leaves other
> languages in the dust, since they do not even make the attempt.
Lojban could certainly be tidied up a lot, by a more muscular BPFK, by a scheme such as Robin proposes, or by -- as I would haha-only-serious-ly advocate -- Jorge being given the cathedra and its powers of fiat; and so like others I support Robin's new initiative. But the only way to really reconcile the "logical" and "language" goals would be to start the language from scratch all over again. Such a starting all over again could be done under the aegis of the Loglan--Lojban project -- John Clifford calls such a third-generation Loglan/Lojban 'LoCCan3' -- but it would be perverse to call LoCCan3 'Lojban'.
Matt is right that Lojban leaves "other languages" in the dust only if "other languages" include only those with a community of users. From a design perspective, it would nowadays be extremely simple to effect enormous improvements on Lojban's design; and the starting from scratch would not in itself entail much new labour. The real challenges for LoCCan3 would be in reaching consensus on a design and in the very difficult search for a design that satisfies the "language" requirements (essentially, the amount of logical form expressed per syllable) enough to make the language better than natural language for general-purpose use.[*]
So, IMO, changes to *Lojban* should be confined to regularizations, simplifications (e.g. eradication of certain cmavo or selma'o) and specifications of semantic interpretation. You can relax the requirement of backwards-compatibility somewhat, but if you completely subjugate it to the Logic and Language goals, you'd end up starting again from scratch and producing LoCCan3. (And to repeat, LoCCan3 would be an immensely better language, but would not be Lojban.)
[* As I see it, the design problem has two parts. Both have to do with finding ways to logically precise meanings concise enough to be worth the effort of verbalizing. One part is to find a more concise way of of encoding variables than standard predicate logic notation and Lojban offer, given that in most propositions we express (in natural language) there are many variables and each variable tends to be argument of many predicates. The other part is to devise an inventory of predicates that expand to more complex logical structures.]
--And.
Oren
--
Jonathan Jones
mu'o mi'e .aionys.
.i.a'o.e'e ko klama le bende pe denpa bu
Timo Paulssen
> So then...
>
> Baselined Lojban = Perl
> Cleaned-up Lojban = Perl 6
> Hypothetical LoCCan3 = Lisp
Erlang (here the quality that makes the comparison work is, that lisp is
still lambda-calculus based, but Erlang is a pi-calculus language.)
mu'o mi'e timos