Hallo conlangers!

On Friday 06 June 2014 21:34:34 R A Brown wrote:

> On 05/06/2014 14:21, BPJ wrote:
> > IIUC the terms _a-priori_ and _a-posteriori_ as
> > originally coined to describe auxlangs referred
> > exclusively to whether their vocabulary was *derived by
> > design, prominently, and consistently* from one or more
> > natlangs or not,
> Are you sure that was once used exclusively of vocabulary?
> Louis Couturat and Léopold Leau in their book "Histoire de
> la langue universelle" (1903) define the terms thus:

Couturat and Leau are of course defining authorities, at least
historically, when it comes to auxlang classification.
> "[les] projets qui, pour des raisons diverses, ne tiennent
> aucun compte des langues naturelles, et qui sont des langues
> originales, construites de toutes pièces."
> (Projects which, for diverse reasons, do not take any
> account of natural languages, and which are fully
> constructed original languages).

While this does not explicitly mentions first principles, it is
quite clear that this definition of _a priori_ applies only to
constructed languages designed without regard of natural
languages.  This applies well to "philosophical" languages as
well as to Solresol and similar schemes, but certainly not to
such languages as Balaybalan or Quenya, which *do* take account
of natural languages.
> "[les] projets qui, prenant pour modèle les langues
> naturelles (particulièrement les langues européennes),
> s'efforcent de les imiter et leur empruntent presque tous
> leurs éléments."
> (Projects which, taking natural languages (particularly
> European languages) as their model,  endeavor to imitate
> them and borrow nearly all their elements from them).

Certainly, Balaybalan and Quenya do take particular natural
languages (Arabic, Turkish and Persian in B; Latin, Greek
and Finnish in Q) as their models and could thus be considered
_a posteriori_ even though they do not borrow nearly all their
elements by the letter.  While the concrete shapes of their
words and affixes may be wholly invented, the way these languages
work, i.e. their phonology, morphology and syntax, is clearly
influenced by particular natlangs and thus a posteriori!
> Doubtless Christophe will correct any errors in translations  :)
> Thus according to these two gentlemen it is not only the
> vocabulary, but all elements of the language, i.e. its
> morphology and syntax.  Thus in their definition an a_priori
> language *takes no regard of any natural languages* but
> constructs its vocabulary, morphology and syntax from
> scratch.   While an a_posteriori language tries to imitate
> natural languages and borrows nearly all its elements from
> natlangs.

And at least by "spirit", B and Q are closer to what C&L define
as _a posterori_ than to what they define as _a priori_.
> I must admit that this was the meaning of the two terms I
> encountered sixty years ago, and I had not been aware that
> there had been any change.
> According to this usage, which has more than a century of
> tradition behind it, Quenya, for example, completely fails
> the  criterion "do not take any account of natural
> languages" and simply cannot be classified as a_priori_.  On
> the other hand, it certainly imitates natural languages and
> does, indeed, borrow elements from them.  The only way it
> might be said to fall short of is in borrowing _nearly all_
> its elements from natlangs.

Yes - it falls somewhere in between.  Yet, it is IMHO, for the
reasons I have laid out above, closer to an a posteriori than
an a priori language.
> > and I think that it's best to keep this as the strict
> > sense of the terms,
> I most certainly think a (fairly) strict definition has to
> be maintained, if they are to have any meaningful use.


>       The
> terms, as we know, are taken from epistemology and Chuck is
> right in saying that no conlang actual fulfills the strict
> epistemological meaning of a_priori.  He is doubtless also
> correct in saying that the authors of a_priori conlangs, in
> the sense defined by Couturat and Leau, were conditioned by
> the cultural milieu of their times and, unconsciously at
> least, by languages they had encountered.

Indeed, no human being could devise a language that is
*epistemologically* a priori - all of us have been exposed to
languages (and other experiences relevant to the perception and
creation of languages), and any language a human being could
design must be influenced, if subconsciously, at least by the
native language of the designer.  There simply is no _tabula

Even the "philosophical" languages of Dalgarno, Wilkins et al.
*do* derive from human language in that they are speakable,
have words and grammatical morphemes, etc.  A language designed
such that it has *nothing* in common with natural languages
would pretty certainly be unusable by humans.  It tells a lot
that none of the _a priori_ schemes achieved any success.  They
were just too difficult to handle.
> Nevertheless, this usage has been around for a very long
> time.

Yes, and it is best not to deviate from it; otherwise, one
gets into a quagmire of misunderstandings.

> If (some) artlangers now redefine a_priori as "not
> borrowing its vocabulary directly from natlangs", then that
> is an innovation and, in view of the continuing use of the
> older definition, it is confusing, as the present thread
> has clearly shown.

Fair.  It is a new and different usage, and a misfortunate one
as it results in misunderstandings.  So we have two definitions
of _a priori language_:

1. A language designed from scratch without regard to any
   natural languages.

2. A language whose vocabulary and grammar are not based in their
   substance on any natural languages.

The first is in use among auxlangers; the second is in use among
artlangers.  They are thus used in two different (though closely
related) fields of discourse.  But because these two fields of
discourse are so closely related, they are easily confused.

This is similar to the different usage of the term _naturalistic_
among auxlangers and artlangers.  Again, the definition used by
auxlangers is the older one, and misunderstandings may ensue.

Yet, the usage among artlangers by now is pretty well established,
too, and there is no practical way of getting rid of it.  One must
live with the fact that the meaning of _a priori_ and _a posteriori_
depends on context, I fear.
> Nor am I the only one to be unaware of this redefinition.
> Arika Okrent defines a_priori as: "_A priori_ languages are
> completely created from scratch."

And languages such as Quenya or Balaybalan are *not* completely
created from scratch by modelled on natural languages.  Of course,
the question is what "from scratch" really means here.

> [snip]
> > ... otherwise e.g. Esperanto would be both at the same
> > time, rendering the terms meaningless. I also think that
> > obviousness does *not* constitute a criterion: Lojban and
> > Volapük do have a-posteriori vocabularies even though
> > that may not be immediately discernible.
> Yes, Couturat and Laeu do add a third category:_mixed_.
> That is, languages where some element(s) is/are a_priori and
> other(s) a_posteriori.  The example they like to give is
> Volapük.

Yes.  And Lojban is another example.  Its grammatical structure
is clearly _a priori_ in C&L's sense (it is not derived from
any natlang but based on a formal logic calculus), but its
lexicon is derived from the vocabularies of a set of major
natural languages and thus _a posteriori_.
> Arika Okrent also accepts their categories of _a posteriori_
> ("_A posteriori languages are based on existing natural
> languages") and _mixed_ ("_Mixed_ languages generally make
> use of roots from natural languages in combination with
> endings and other elements fashioned according to the _a
> priori_ categorization technique).  But she adds a fourth:
> _other_.
> Of this fourth category she writes; "Throughout history, the
> purpose of language invention has for the most part been to
> improve on natural language and create a universal language
> for real world use. But sometimes, and increasingly more
> commonly, inventors make languages for artistic purposes
> only, crafting languages to satisfy a vision of what a
> language should be. These projects may incorporate elements
> of a priori, mixed, or a posteriori strategies, but they
> belong in a category of their own."
> In other words, her "other" category is 'artlangs'!


> Needless to say, she places Quenya there   :)
> When I see Quenya designated as "a priori" I wince because
> it counter to more than a century of use of the term _a
> priori_; then I collapse in disbelieving laughter because,
> as a Latinist, to see Quenya described a _a *priori*_ is so
> bizarre!

I understand that notion, and have felt for a long time that
Quenya is too close to certain European languages at least "in
spirit" to really count as a priori!  Yet, it is an example of
what is often called "a priori" among artlangers.
> Artlangs are, are they not, works of _art_?  It seems to me
> just as bizarre to discuss whether Quenya is a_priori or
> a_posteriori as it is to discuss whether Botticelli's "Birth
> of Venus" or Beethoven's 5th symphony is a_priori or
> a_posteriori.

Fair.  Couturat's and Leau's definitions simply do not apply
well here because the language is of a very different *nature*
than the languages C&L were considering.

... brought to you by the Weeping Elf
"Bêsel asa Éam, a Éam atha cvanthal a cvanth atha Éamal." - SiM 1:1