Hallo conlangers!

On 08.06.2016 09:17, R A Brown wrote:

> On 07/06/2016 18:21, And Rosta wrote:
>> On 7 June 2016 at 01:37, [log in to unmask] wrote:
> [snip]
>
>>> SLI: I'm still confused. According to And, 1. Is a
>>> conlang a language?
>>>
>>
>> Under my narrow and platonic definition of language,
>
> With respect, I give more weight to linguists like David
> Crystal and the late Larry Trask.

So would I.

> Also one might recall
> that J.R.R. Tolkien in his essay "A Secret Vice" refers to
> his creations from juvenile 'Nevbosh' to his mature ones as
> _languages_ and speaks of "language-construction."

Fair. They are languages at least in an informal, colloquial sense.

>> a conlang is a language either (1) if its inner workings
>> are fully specified, in the case of the authored
>> conlang,
>
> How many natural languages meet that requirement?  Indeed,
> as living languages are continually in a state of flux and
> evolving, can their "inner workings" ever be fully
> specified?  Of how many dead languages can we say we know
> the full specification of their "inner workings"?

Indeed. Linguists still find out new things about even the best-known 
natlangs - most of these findings, however, are simply *not relevant* to 
anyone intending to *use* them. All you need is a dictionary, a 
reference grammar and a collection of sample texts, and that is a quite 
overseeable amount of data that fits into a few small books. A conlanger 
*can* achieve that - there are examples.

>> or (2) if it is used, in the case of the pidgin conlang.
>
> But what criteria do we judge whether a conlang is a 'pidgin
> conlang'?
>
>> So in practice this means that most conlangs aren't
>> languages. I don't know what degree or quality of
>> usedness is sufficient for us to conclude that a langue
>> underlies the parole (and hence that this is a
>> language)..
>
> I do not understand why this narrow definition is termed
> 'platonic.'  Indeed, the use of the terms 'langue' and
> 'parole' surely suggest _Saussurean_ rather than platonic ideas.

Yep. I know very little about Platonic philosophy; while it seems to me 
that eternal Platonic Forms do not account well for the changeful 
universe we live in, I must concede that I perhaps understand the 
concept wrongly. We should leave poor old Plato out of the game here.

On 08.06.2016 09:53, David Peterson wrote:

>> On 7 June 2016 at 01:37, [log in to unmask] <[log in to unmask]>
>> wrote:
>> 3. Does a conlang maker invent a language?
>> On Jun 7, 2016, at 10:21 AM, And Rosta <[log in to unmask]> wrote:
>>
>> Only if its inner workings are fully specified (which is currently feasible
>> only for loglangs).
> In this discussion, unless I’ve missed it, there’s only one thing worth stating that hasn’t been stated yet. It goes roughly thus: If a painter is painting a mural on a wall but dies before it’s completed, were they painting a mural, or a model mural?

A mural.

> There’s no question that the mural itself is incomplete, and so isn’t the best example of a mural to show someone, but we’re talking about what the muralist was actively engaged in doing before they died. Same goes for an unfinished novel like Dead Souls, or an unfinished symphony, or an unfinished anything. There’s no argument about what it is once the artist has left off, either willingly or not. If it turns out that the work is never finished, though, it seems neither charitable nor accurate to say that they weren’t engaged in the creation of a full and complete work. If one includes the vast potential vocabulary of a language, it’s a task that likely is too time-consuming to complete for a single person living an ordinary life. But I’d argue that most language projects are sincere attempts to create a language even with the multitudinous vocabulary that will likely never be completed. Consequently, what the creators are engaged in is the creation of a language.

Right - unfinished works are commonplace, and some are counted among the 
best achievements in their arts, and it has been argued that a conlang 
may never be really complete; but there are conlangs that are 
well-rounded and usable for general conversation.

> A conlang, like any other work, though, has certain levels of achievement after which time it can be appreciated. Dead Souls, for example, is one of my favorite works, despite the fact that it was never finished. There’s enough text there to be read and enjoyed. An unfinished novel that only has two sentences and maybe a couple of notes is probably not worth reading, even though it wouldn’t take but a minute (probably only of interest to a grad. student studying the author). Conlangs are the same way. A sketch that’s a day old has no meat. A conlang that someone has been working steadily on for several years is getting closer to something like Dead Souls, which can serve the purpose of a full novel in every sense of the term, despite its not being a full novel. Everything beyond that just brings the conlang closer to completion.

I have always felt that our art, like any art, has its large and small 
forms. A well-rounded, fairly "complete" conlang such as Mark 
Rosenfelder's Verdurian, with a detailed reference grammar and thousands 
of words, would be the equivalent of a novel or a symphony. My main 
conlang Old Albic aspires to become such a work, though it is unfinished 
yet. But there are also smaller projects which are never meant to be 
"completed" but merely to test out an idea or give an "image" of a 
language. An extreme case is this vignette of mine:

http://www.joerg-rhiemeier.de/Conlang/attidian.html

Just seven words, and very little grammar. There is a bit more behind 
the scenes, though - the words (except the name _Hirculua_, which is of 
course a Latin loanword - but the changes it has undergone shed a light 
on the phonology of the imagined language) and inflections are all 
cognates of Old Albic words and inflections, connecting to the latter by 
regular sound change. This is not really a language, merely an image of 
an ancient language of which only a tiny fragment is known. Yet, is this 
an "invalid" work of conlanging? I don't think so.

Most conlangs fall between these extremes. The typical conlang consists 
of a grammar sketch, a few hundred words, and some short sample texts. 
Not enough to actually use it in everyday situations, but enough to give 
a good idea of what the language is like. Like the sketches one finds in 
books such as Comrie's _The World's Major Languages_. Yet, if such a 
conlang is well-designed, it can be expanded into an actually usable 
language.

I have plans for several descendants and relatives of Old Albic, but 
most of them will probably never be developed into languages as complete 
as I intend Old Albic to be; for most, I will just write a grammar 
sketch with a list of a few hundred words and a few short sample texts. 
Yet, I think of these projects as "languages", because they are meant to 
be *models* of imagined languages.

> Within the community, there are certain unofficial benchmarks we’ve recognized when it comes to stages of completion of a conlang there are a bit haphazard. Completing the Babel Text is (or was) one. Celebrating arbitrary word counts is another (coining the 500th word, 1,000th, etc.). Being able to translate the 207 word Swadesh List. Participating in a conlang relay is one. Though it was a minor one, participating in an *inverse* conlang relay was a bigger benchmark.

These are all good tests of the workability of the conlang. Not as 
demanding as everyday communication, but good ways of trying out and 
demonstrating the language.

> Aside from that, though, the standards are purely notional. We kind of rely on the conlanger to say, “Okay, you can use this for serious translation, outside of the need for specific vocabulary here and there”.

What domains a conlang should be usable for of course depends on its 
setting. The language of a starfaring race in a space opera universe 
needs words for starships, their parts and other concepts pertaining to 
interstellar travel. The language of a Bronze Age culture does not. The 
language of a reindeer-herding tribe would have a more differentiated 
vocabulary for reindeer than one of a temperate-zone urban civilization 
which may have only one word in that field, namely 'reindeer'.

> It might be useful to actually come up with some criteria by which we can indicate how complete a conlang is, though I don’t know what those criteria should be. Gary Shannon had a list of graded sentences, I recall. That’s one thing, though it might lead to conlangers “teaching to the test”, so to speak (like how conlangs from a certain era had words for “mortar” and “confuse”).

Sure - some conlang word lists easily give away which texts the author 
has already translated. I find a thematic vocabulary like that of Verdurian:

http://www.zompist.com/thematic.htm

a nifty means of coming up with a well-rounded vocabulary. You can 
easily see where your language already has a good vocabulary and where 
more words are needed.

--
... 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