Hallo conlangers!

On Thursday 05 June 2014 09:52:22 R A Brown wrote:

> As most of the reply below has nothing directly to do with
> Bālaybalan, I've brought all under this new thread.

Fair.
 
> On 04/06/2014 20:15, David Peterson wrote:
> > On Jun 4, 2014, at 11:21 AM, R A Brown
> > 
> > <[log in to unmask]> wrote:
> >> But this is like arguing that Quenya or Sindarin are
> >> a_priori, which IMHO is nonsense.  If a language is
> >> a_priori then it is *built from scratch from _first
> >> principles_, without any reference to any natural
> >> languages*.   Clearly Quenya and Sindarin are not so
> >> built.
> >> 
> >> The grammar of both languages are original (though with
> >> aspects that recall several natlangs that I will not
> >> detail here) and their vocabularies are entirely
> >> original.  But we know that Tolkien fashioned these
> >> languages out many years of exposure to natlangs.
> > 
> > This is, in my opinion, an absolutely useless
> > classification,
> 
> 'Absolutely' is a strong word.  It clearly misrepresents
> what I was saying.  I did, in fact, say that IMO the
> categorizing of conlangs as a_priori_ or a_posteriori_ makes
> little sense and can be misleading when applied to artlangs
> and engelangs.  They are, however, a useful categorization
> of auxlangs.

Yep.  While it is often said about artlangs such as Quenya or
Klingon that they are "a priori", this misses the _first
principles_ bit which is important in the original definition
of the term in interlinguistics.  There are no "first principles"
in those languages; their form-meaning relationships are as
arbitrary as in natlangs.  We are dealing with two different
definitions here:

Auxlangers: developed from a set of first principles
Artlangers: developed from scratch

As with "naturalistic", two different usages.

> > and I would say it is one that few if any conlangers
> > agree with right now, nor would agree with—nor should.
> 
> Presumably most conlangers would at least agree with me
> about a_priori and a_posteriori auxlangs.

Yes.  There aren't many auxlangs which are built from scratch
but without first principles, mainly because such languages
would have no justification in their design - nothing would be
won that way over using a natlang (except, perhaps, "cultural
neutrality" - but an African or East Asian could, for instance,
complain about Quenya as an IAL that it is clearly European in
its makeup, as it is built on Indo-European and Uralic models).
 
> > By this classification, the only theoretically possible a
> > priori language is one built by machines, or by humans
> > who have never been exposed to language.
> 
> ... which is contrary to the facts. [...]
> 
> What these people had in common was the notion that natural
> languages were imperfect and flawed.  They eschewed
> natlangs, and attempted to build a perfect "philosophical"
> language from scratch, from first principles.  The languages
> are in that respect a_priori; and I suspect everyone on this
> list would agree.

Just that!  And a language such as Quenya wouldn't have satisfied
them at all.
 
> This approach contrasts strongly with that of people like
> Schleyer, Zamenhof and Jespersen who tried to craft auxlangs
> out of natlang material.  I suspect everyone on this list
> would agree that these are a_posteriori.

Of course.
 
> AFAIK the division of conlangs as a_priori and a_posteriori
> was first applied to auxlangs.  There, I have said, the
> division is a useful one.

The division of conlangs into "a priori" and "a posteriori"
languages indeed was first applied to auxlangs, where, as you
correctly say, it is useful and meaningful.  But with artlangs,
it is much less so; hence, Arika Okrent added the category
"other" in the sense of "neither a priori nor a posteriori".
 
> > Any classification system that calls both Quenya and
> > Esperanto a posteriori is, in my opinion, simply of no
> > practical value whatsoever.
> 
> I agree that classifying all conlangs as either a_priori or
> a_posteriori is of no practical value.  I do *not*, however,
> agree that the classification itself is of no value for all
> conlangs.

As you say, it is useful for what it was made for: auxlangs.
It breaks down elsewhere, though.
 
> Both Esperanto and Quenya certainly draw their material from
> natlangs.  If Tolkien had not been an English speaker with a
> knowledge, inter_alia, of Latin, Classical Greek and
> Finnish, Quenya would have been a very different language.

Surely, it would be a different language indeed!  While it is not
possible to state a source for individual Quenya *words*, it is
very clear that the design as a whole is strongly influenced by
the languages you mention, and Tolkien never made a secret of
that.
 
> Therefore, IMO a classification system that calls both
> Quenya and Babm a_priori is simply of no value whatsoever.

Quite certainly, Quenya and Babm are so different from each other
that a classification that puts them into the same category is of
questionable value, or at least misses something important.

Balaybalan and Quenya are strongly influenced by a limited set of
natlangs (Turkish, Arabic and Persian with Balaybalan; Latin,
Greek and Finnish with Quenya), and not based on first principles.
Hence, they are in a sense closer to a posteriori than to a priori
conlangs.

> ============================================================
> 
> On 05/06/2014 03:13, Js Bangs wrote:
> [snip].
> 
> > Aside from that, I quite agree with David Peterson's
> > assertion about the meaning of "a posteriori". To call
> > Quenya and Esperanto a posteriori is to render the term
> > meaningless.
> 
> No more meaningless IMO than calling Quenya and Babm a_priori.

Indeed not.  If you define _a priori_ as "based on a set of
first principles" and _a posteriori_ as "based on one or more
natural languages", Quenya can be said, with some qualifications,
to be _a posteriori_ as it is in a holistic way clearly based
on Latin, Greek and Finnish.  And Balaybalan is a similar case
to Quenya.

> ===========================================================
> 
> On 04/06/2014 22:23, Chuck Haberl wrote:
> [snip]
> 
> > Ray, I agree that the problem is basically definitional
> > and that your points are well taken
> 
> Thank you.
> 
> > (although characterizing other approaches to the schema
> > as "nonsense" and "meaningless" is a bit over the top).
> 
> I did qualify "nonsense" by _IMHO_ (in my humble opinion).
> As you see from the above, I have not been the only one to
> use the term 'meaningless'; I certainly did not categorize
> a_priori and a_posteriori as 'absolutely useless' (quite the
> contrary).  However, IMO unless the terms are used with some
> care they can lose meaning and usefulness

Certainly!
 
> > would, however, dispute your claim that the terms are
> > useful for auxlangs but not for any other forms of
> > language, constructed or otherwise (but please stop me
> > if I've misunderstood this claim).
> 
> I don't understand how the terms would apply to
> non-constructed languages, i.e. natlangs.

They wouldn't.
 
> What I argue, as you see above, is that they are a useful
> categorization of auxlangs.  But if we try to categorize
> _all_ conlangs as either a_priori or a_posteriori then we
> run into problems and the results can be meaningless (see
> above) or downright misleading.

This is clearly the case.  We have had quite some confusion in
the past over the term "naturalistic", which is used differently
among auxlangers and among artlangers.  It is apparently also
true with "a priori/a posteriori".  People use the term "a priori"
to classify artlangs whose words and morphemes are not taken from
a concrete set of source languages, and in *that* usage, Quenya
is indeed "a priori".  But this is *not* the way the term is used
in interlinguistics.  We are dealing with two different definitions
here.

> That does not mean that some cannot be so labelled without
> confusion.  No one would dispute that Brithenig is other
> than a_posteriori. But does that really help?  Brithening is
> a "what-if" language, in this case "What if Britain had
> become Romance speaking."  I fail to see how any "what-if"
> language could be a_priori - they must per_se be a_posteriori.

Surely, Brithenig is a posteriori even by the definition used by
auxlangers.  Yet, it is a posteriori in a different way than
Esperanto.  But we have a word for the specific kind of
a-posteriori-ness of Brithenig: _diachronic_.

For an artlang that is a posteriori in a similar way as Esperanto,
see Mark Rosenfelder's Verdurian, where you can often easily tell
which language a word is taken from.  Yet, Verdurian is still a
different beast than Esperanto, chiefly by virtue of being an
artlang rather than an auxlang.
 
> But as an example of meaninglessness or, at best, confusion
> let me take the example of Lojban's lexicon. We know how it
> was derived; we know the natlangs from which it was derived.
>   The lexicon of Lojban is a_posteriori.  But it is certainly
> not a_posteriori in the same way that the lexicon of
> Esperanto, Novial or Glosa is. Therefore, I argue, to label
> Lojban's vocabulary a_priori is contrary to fact, but to
> label it as a_posteriori is misleading.

It is, technically, a posteriori, but not in the same way as
Esperanto.  It is perhaps like Volapük, which is also technically
a posteriori, though the words are mostly altered beyond
recognition.  Auxlanger call those beasts "mixed languages", but
that does not always actually mean a mix between first principles
and borrowed words (though Lojban may be considered such a beast).
 
> My argument is that where the two terms are helpful, let us
> use them; but where they are incorrect or misleading then it
> is better IMO not to use the terms.

Yes.  But I fear that the term "a priori" is already established
in the weak sense of "designed without drawing on concrete words
and morphemes from natlangs" in the artlang community.  As I said
above, two different definitions in two different fields of study.

When such things happen, it is often futile to fight against that.
Like the way the term "protolanguage" is used in utterly different
senses in historical linguistics and in language origins studies.
You must eventually concede that the same word has different
meanings in different fields of discourse.
 
> > If we narrowly define "a priori" and "a posteriori"
> > according to the epistemic conventions of philosophers,
> > then we obviously run into the problem that David
> > raised.
> 
> ... and that I also raised.  But the two terms are both
> Latin phrases and I am only too well aware what they mean.

Sure!
 
> > All of the so-called "first principles" and taxonomic
> > systems assumed by these languages are environmentally,
> > culturally, and even linguistically determined, and the
> > approach to grammar that their designers take is
> > invariably derived from the contemporary state of the
> > art in the study of natural languages (often rather
> > self-consciously so). Therefore, they are ALL undeniably
> > "a posteriori," at least from an epistemological
> > perspective.
> 
> They are indeed conditioned by the cultural milieu in which
> they were derived.  That includes even the 'late-comers'
> like Ro and Babm.

Sure!  Many of the classifications found in 17th-century
philosophical languages now appear quaint and outdated.  In
one of those languages (I think it was Dalgarno's Ars signorum)
comets are classified as fire phenomena - now we know that they
consist mainly of ice, which is of course as far removed from
fire as one can get within the realm of everyday life.

>       But the important thing IMO is that their
> authors deliberately rejected natlangs and tried in their
> own ways to build the thing from scratch.  Whether they are
> epistemologically a_priori is, as I see it, rather like
> arguing whether Cartesian philosophy is epistemologically
> a_priori, with everything being logically derived from:
> cogito ergo sum.
> 
> I think we can agree to call these languages a_priori.

Yep.
 
> > I guess what is needed is a shorthand for languages
> > whose components are original to themselves rather than
> > borrowed directly from another source or sources.
> 
> Well, yes, _directly_ is the interesting word. Zamenhof (and
> others) took his vocabulary (with perhaps the odd exception
> like the notorious _edz-_) directly from natlangs, whereas
> Lojban derived its vocabulary indirectly by feeding the
> natlang data into an algorithm.

Which means that Lojban occupies a middle position bwtween
a priori and a posteriori, and could be classified as "mixed".
 
> I agree, however, that labeling Bālaybalan as a_posteriori
> is as meaningless as to label Quenya  a_posteriori, and I
> shall be amending the paragraph on Bālaybalan on my website
> as soon as possible today.  But I equally hold, for reasons
> argued above, that to term Bālaybalan as a_priori is equally
> meaningless.

Yes, languages such as Balaybalan or Quenya also fall between
the two poles of _a priori_ and _a posteriori_, but in a
different (somehow opposite) way than Lojban.  We may have
something like a tetrachoric chart here.
 
> Of course Jörg is right:
> "What this highlights is the sheer fact that conlang
> classificatory terms are used differently among
> interlinguists than they are used here."

Yep.
 
> So my classification and Jan van Steenbergen's differ;
> http://www.carolandray.plus.com/Glosso/Glossopoeia.html
> http://steen.free.fr/classificatie.html

Indeed, they are different.  Jan uses the term "a priori" in the
way many artlangers use it, without reference to "first principles".
(He even misclassified Verdurian, which he classifies as "diachronic
a priori" like Quenya, while it is actually an eclectic a posteriori
language where the origins of the words are often easy to tell.)

I have opined earlier that the distinction between "a priori" and
"a posteriori" is of less use in artlangs than in auxlangs, and
this discussion vindicates that.

--
... brought to you by the Weeping Elf
http://www.joerg-rhiemeier.de/Conlang/index.html
"Bêsel asa Éam, a Éam atha cvanthal a cvanth atha Éamal." - SiM 1:1