Re: Defining "Language"

On 7/19/07, And Rosta <and.rosta@...> wrote:
> li_sasxsek@NUTTER.NET, On 18/07/2007 20:35:
> > This may not be a popular thing to say here, but I don't consider most
> > conlangs to be "languages".  I consider them to be plans or blueprints
> > for languages, and then become languages when they come to life
> > through usage.  Until then, they are just concepts.  The "community"
> > may only be two people, but there does need to at least be a speaker
> > and a listener.
> Terminologically, I prefer for "language" to denote the blueprints not the actual human
> behaviour (or brain states) that realizes the blueprints. But the key thing
> is the grammar--usage distinction, and it doesn't really matter whether we
> prefer "language" to mean "grammar" or "usage". ('Grammar' = the language
> code, the system of form--meaning correspondences, Saussurean langue.) What's
> clear is that conlangers invent grammars.
In a September 2005 thread on the AUXLANG list, I said (thinking of the
langue/parole distinction but not using those terms):

> From a linguistic perspective, there seem to be two
> ways of defining a language:
>
> 1. the set of all utterances in the language, written and
> spoken;
>
> 2. the language-specific structures in the brains of
> all the fluent speakers of the language.
(Note the context of our discussion was defining specific
languages, not language in general -- "What is Esperanto?"
--> the total Esperanto corpus or the language-structures in the brains
of Esperanto speakers, not the original prescriptive blueprint for the
language, or the general idea of an auxlang with such and such
characteristics.)

It is in sense #1 that dead languages are still languages
even if they have no fluent speakers, and (perhaps) conlangs
with a corpus, even a small one, might be considered languages
even though they have never had even one fluent speaker.
In sense #2, however, a conlang might could be considered a "real"
language even if it has only one fluent speaker (typically the
language's creator).

I think something is missing here, though, because Dana's definition
requiring at least two fluent speakers to form a community makes
intuitive sense to me.

And And's use of the term to refer to "denote the blueprints not the
actual human behaviour (or brain states) that realizes the blueprints"
has problems even when applied to conlangs.  I think Kalusa, for instance,
was more of a "real language" (whether in terms of langue or parole
or speaker community) than most conlangs, even though it never
had any blueprints -- only a corpus, and (by the end of the project)
several more or less fluent "speakers".  (Even if two of the most fluent
"speakers" had different pronunciations in mind for the name of the
language, as we found out at LCC2 (David Peterson after my talk:
I was surprised to hear you pronounce /ka.'lu.sa/ as /'ka.lu.sa/.))

--
Jim Henry
http://www.pobox.com/~jimhenry

Replies