Hallo!

On Sat, 24 Jul 2004 15:27:02 -0500,
"Mark P. Line" <[log in to unmask]> wrote:

> Jörg Rhiemeier said:
> >
> > On Fri, 23 Jul 2004 09:08:42 -0700,
> > Gary Shannon <[log in to unmask]> wrote:
> >
> >> To my warped way of thinking
> >> it's more interesting to build a conlang knowing
> >> nothing about how languages are built.  After all,
> >> every natlang in existence, past and present, was
> >> initially created by people who didn't know the first
> >> thing about linguistics.
> >
> > True.
>
> I'm not sure what "create" is intended to mean when applied to natlangs here.
>
> It occurs to me that maybe y'all are referring (by the word "create" that
> I was wondering about above) to the *production* of language (of
> utterances), not to the invention of a natlang code (that is, "parole" is
> being created, not "langue"). Then again, maybe not.

I understood it to refer to the (rather unconscious
cultural-evolutionary
process of) creation of _langue_; but I am not Gary Shannon.

> >> Linguistics is NOT the study of how to build a
> >> language, but is, rather, the study of how to describe
> >> languages taht spontanesouly came into existence and
> >> evolved in the absence of any planning or design.
> >
> > Yes.  Linguistics, as it is taught in universities, is not
> > about *building* languages.  It is about understanding
> > how languages work - languages that are usually the product
> > of cultural evolution.
>
> Similarly, musicology (especially ethnomusicology), as it is taught in
> universities, is not about composing music. It is about understanding how
> music works -- music that is usually the product of cultural evolution.

And one doesn't have to be a good musicologist to make good music,
nor is a good musicologist necessarily a good composer (let alone,
a good performer).  And indeed, most composers owe much more to
previous composers than to theory of music.

> I wonder if it's relevant to conlanging, by analogy, that some composers
> know a lot about musicology while some don't, and that there is an
> infinite variety of ways to evaluate these composers' products, most of
> which ways probably do not involve the composer's knowledge of musicology
> (although some might).

Yes.

> Personal note: My beef has usually been with conlangers who insist that
> they are operating well inside of natlang evolutionary space,
> notwithstanding any amount of evidence to the contrary (e.g. deep center
> embedding, phonological conditioning of open-class suppletive allomorphy,
> pure ergativity, etc.). I'm becoming more mellow with age, however, and
> now almost always leave everybody alone. :)

What exactly do you mean?  Fictional human languages that are not
plausible human languages?

> >> That's how I like to build my conlangs; just forging
> >> ahead blindly, without plan or design, to create
> >> something that can only be described with lingusitic
> >> terms AFTER it has reached a certain level of
> >> maturity.
> >
> > Well, I think some linguistic knowledge is indeed helpful
> > in conlanging, but the bearing linguistic education has
> > on conlanging should not be overrated either.
>
> I couldn't agree more. Most linguistic education sucks, as nearly as I can
> tell.

I haven't undergone formal training in linguistics, but I have worked
through several textbooks, and picked what I found useful in conlanging.
What I have noticed (but what didn't surprise me) was that linguistics
is far less of a rigorous, exact science than, say, physics, and much
bad theory is in circulation.

> >> The best conlangs are the result of the same kind of
> >> linguistic chaos and anarchy that forged all the
> >> natlnags.  IMHO. :)
> >
> > I think it cannot be generalized either way.
>
>
> I have to agree with Andreas on this one.

Andreas who isn't Andreas, in fact ;-)  But you are not the first
to call me `Andreas'; I know a guy who consistently calls me
`Andreas' because it is somehow engraved in his mind.

>       There are just two many ways to
> give meaning to Gary's phrase "best conlangs". If we choose to evaluate
> conlangs on the basis of apparent erudition of their creators, then the
> anarcholangs might lose. If we choose to evaluate them on the basis of
> disregard for natlang evolutionary space, then the anarcholangs might win.
> If we choose yet another strategy to evaluate conlangs, then it might not
> make any difference whether or not any linguistic knowledge was used in
> their creation.
>
> This begs the question of whether or not there's a best way to give
> meaning to the phrase "best conlang". I think not, because we create our
> conlangs  in pursuit of a much too widely disparate choice of goals.

I whole-heartedly agree.  I think the only meaningful gauge for
a conlang are its own design goals.  For example, if one was to
present a philosophical language with self-segregating morphology
and a typologically unlikely phoneme inventory as a near-extinct,
pre-Indo-European minority language of the southeastern Alps,
I'd say that he has done a pathetic job of it.  But that doesn't
mean that the language is not excellent from some other point of view.

Or Old Albic would probably not be a good machine translation
interlingua, for example.  But that is meaningless because that is not
what it is made for.  It is a fictional language of pre-Celtic Britain,
and what would be a meaningful criticism would be an whether it is
a plausible pre-Celtic language of Britain or not.  Criticizing
its lack of, say, self-segregation, is meaningless because it wasn't
a design goal, nor implied by some other design goal.

The discussion about what makes a good conlang has happened here
several times, and the conclusion was always that there is no single
gauge that meaningfully applies to all conlangs.

Greetings,

Jörg.