Re: Wittgenstein on conlanging [was Re: relative weirdness]

"Thomas R. Wier" <trwier@...> wrote:

> Quoting Jörg Rhiemeier <joerg.rhiemeier@...>:
> >
> > Well, I have no children to tell me that I waste time with
> > conlangs, but my (elder) brother does the job.  When I recently
> > made the mistake mentioning the Conlang Translation Relay in
> > a conversation about translation problems, he thrusted
> > Wittgenstein's infamous argumentation against "private
> > languages" at me, stating that the translations were inevitably
> > meaningless and thus the whole thing a waste of time.
>
> Not only was that unfair of him, I think he rather missed
> Wittgenstein's point.
So do I.

>        Wittgenstein conceived of rules as
> kinds of signposts that have merely made official the
> tendencies and customs of past usages, rather than rules
> being abstract laws (he analogizes these flaws to that of
> a railroad).  As such, language is a matter of putting your
> uses into the context of what everyone else before you has
> done.  This would seem to rule out conlangs by their very
> nature, but the question is not one of private versus public,
> so much as logically private versus empirically private.
And conlangs are at most empirically private, and if someone else
than their creator takes the trouble learning them, not even that.
As soon as someone else learns it, the conlang is no longer private;
on the other hand, a natlang may become empirically private when
no new speakers learn it any more and all old speakers die or
forget the language save one.  Esperanto started as an (empirically)
private language; now it is used by hundreds of thousands of people,
and it is even reported that there are native speakers of it
(though they are only few, and all bilingual).  Klingon, Quenya,
Sindarin and several other conlangs also have their communities of
speakers (even if few of them use the conlang in everyday-life
situations, but rather write poetry in them or whatever), and hasn't
_Hamlet_ been translated into Klingon?  Such empirically
private languages are thus not substantially different from
public languages, only that there are fewer speakers, namely only one.
(The number of speakers of a language may even go down to *zero*
without making a real difference as to Wittgenstein's argument,
as long as there is a dictionary and a grammar still sitting on some
bookshelf, allowing later generations to learn the language again.
Thus, a language may return to life after having gone extinct, as
happened with Cornish.)

But Wittgenstein's argument against private languages is all about
*logically* private languages that are closed to others, and nothing
else.  W. cannot have believed that constructed languages are
impossible; Esperanto and a few others were alive and well at his
time, and he *must* have been aware of the fact.

> I'll quote A. C. Grayling's comments on the subject at length,
> since he puts it so much better than I can:
>
> [lengthy quote snipped]
>
> Wittgenstein's point, then, is that the kinds of things we do
> here in this group are perfectly possible, because all the
> languages that we create here are *in principle* capable of
> being learnt by all the other members of the group; in fact,
> many of us *do* learn the others' languages.  If they were
> not, that would beg the question of how *any* language could
> exist at all, much less constructed ones.
Yes, there is no reason whether it should make a difference on the
learnability of a linguistic system whether the grammar and semantics
are the outcrop of millenia of cultural evolution, or were designed
by a single individual from scratch, as long as they are compatible
to the modes of human thought, which is the case with most conlangs.
The fact that the meaning of something has been set by a single
individual, rather than being agreed upon by generations of native
speakers, is irrelevant.

The results of the various translation relays that have been played
here demonstrate that conlangs are accessible to people other than
their creators and thus aren't "private languages" in the
Wittgensteinian
sense.  Sure, the semantics drift quite a lot, and what comes out
after 20 or 30 translations is quite different from the original text,
but even after so many translations, there is still a resemblance left.
The drift is due to the same problems that are encountered in
translations from one natlang to the other: imperfect knowledge of
the language translated from; words and grammatical categories
that do not correspond semantically in a 1-by-1 way; words denoting
concepts unique to the cultural tradition associated with a language
(whether it is real or invented); different connotations with
words that have the same literal meaning (to give one example from
the current relay: the Skerre text I received contained the phrase
_hahak-oto_ `waterways', which I *could* have translated literally
into Germanech as _strazes a ach_, but the latter I imagine to
have connotations of barges, canals, regulated rivers and port
facilities - it means `waterways as traffic routes' - in Germanech,
which makes it unsuitable in the sort of mythological text
the relay text clearly was, so I changed it into _reivers_ `rivers').

If one was to do the same experiment with natlangs from all corners
of the Earth, the drift would probably be of the same order of
magnitude.

> (You should tell your brother about this.)
My response was less founded, as I am not very familiar with
Wittgenstein's theory, but it went in a similar vein: I pointed out
that conlangs are hypothetical languages, models of linguistic
systems that could be *imagined* to be spoken by people; thus,
no matter that they are creations of single individuals, each
element of them has its meaning set by the creator, that is as well
defined as the meaning of a correspondent element in a natlang.

He did not object to that, he only said, "Hmmm, yes".
And when I pointed out that we are doing it just for the fun of it
anyway, he said nothing anymore.

(I have a vision of Wittgenstein's ghost haunting my brother at night,
telling him, "You got me wrong when you lampooned your brother
about the `private languages' he and his friends play with.
Those are only empirically private languages, not logically private.
Of those I didn't talk at all.  Did you ever take a look into the
CONLANG mailing list, anyway?  Have you seen what your brother
and his friends are doing there?")

His wife, BTW, asked me only a simple question on the workings of
translation relays, namely, "How does a player know the previous
player's conlang?", and I explained to her how we do it with
interlinears and the grammar sketches we post on the list.

> > My parents, however, know nothing about my conlanging.
> > I have learned not to tell anything about my bizarre
> > worldbuilding and related projects quite early, when all
> > I got was utter indifference from my father and sharp
> > disapproval from my mother.
>
> My father only learned that I conlang when he read my website
> a while back, on which I admitted taking part in the "secret
> Vice".
My parents have no Internet access and don't care about it at all,
so they don't know about my web pages.

>       He asked me about it, and seemed to keep an open mind,
> but did take a great interest.  Like you, I tend to keep my
> conlanging to myself and to those few whom I trust not to
> belittle it in a serious way.
Yes, that's a healthy strategy.  Avoids wasting energy trying to
explain it to people who'd probably never get it.  I'm doing the same
with fully clothed swimming and other personal weirdnesses.

> [...]
>
> > Well, I know that I am somewhat weird -
>
> All human beings have oddities about them, and none are
> based on that fact alone warrant for persecution.
Nik Taylor has a nice quote on this matter in his .sig:

"There's no such thing as 'cool'.  Everyone's just a big dork or nerd,
you just have to find people who are dorky the same way you are."

(I hope you don't mind me quoting it here, Nik.)

Jörg.