Hallo conlangers!

On Mon, 2 Apr 2012 21:37:29 +0100 Sam Stutter wrote:

> On 2 Apr 2012, at 20:46, Jörg Rhiemeier wrote:
> 
> > Hallo conlangers!
> [...]
> > A smooth flow of a conlang is indeed not easy to achieve.  Many
> > beginners' conlangs feel as if they were assembled from LEGO(R)
> > bricks or the like - the morphology is cookie-cutter regular
> > with junctures wide open, the words bite each other, the whole
> > thing sounds like gibberish or Pig Latin and not like an
> > organically grown language.  I admit that avoiding this and
> > coming up with a coherent, beautiful conlang is difficult.
> > Languages such as Quenya or Kēlen were not cobbled together in
> > a fortnight; they were honed and refined over *years*.
> 
> We've all been there.

Sure.  My first conlangs were pretty newbish, modeled on what
natlangs I knew then: German, English and Latin.  I did not even
make up proper phonologies, I just used what kinds of sounds I
knew and felt right to me.  Also, "diachronic developments"
basically broke down to random changes of vowels and sometimes
consonants, and languages separated by 12,000 years of "history"
resembling each other about as closely as German and English.

It was only later, in the early 2000s, when I read my way through
linguistics textbooks, that I learned to do fictional languages
"properly".  Basically, the impulse for this self-study of
linguistics was a phase when I perused Ardalambion to acquaint
myself with Tolkien's languages, and saw how much linguistic
sophistication was in there.  That was early in the year 2000.
At about the same time, I began following the discussions on
CONLANG, and I saw what kind of sophisticated stuff people were
doing there.

That was the point when I decided to go to the library, and read
every textbook they had.  I also worked my way through dozens of
grammar sketches from all over this world and several others in
order to get a sense of the vast diversity of languages, both
natural and constructed.  However, I avoided the temptation of
building everything I found into a single language, and made a
decision about what to use in what was to become Albic and what
not.  I also soon found out that doing my very own thing was
infinitely better than basing it on Sindarin, as I tried when I
started with it.

Also, by the wayside, I discovered the fascinating world of
historical linguistics, which has become a hobby of mine, both
as a source of inspiration for my conlangs and as a "thing of
its own".

> > You once wrote that you design languages without defining a
> > phonology, perhaps that is part of the problem?  It is rules
> > of phonology that give a language a coherent sound and smoothes
> > out the seams between the morphemes.  Even such highly regular
> > languages as Turkish have a coherent sound because there are
> > well-defined phonological rules.
> 
> I find "hearing the speakers" before actually coining any words really 
> helps. I know that there *is* the right sound out there, because I can hear 
> it being spoken, it's just like trying to remember that dream you had last 
> night. It's more of a process of holistic reconstruction rather than top-up 
> construction (even though that's what you're *really* doing).

Yes.  It is reasonable to start with a rough idea of what the
language should sound like; but casting this notion into an
actual phonology is not always easy.  It was actually quite
straightforward with Roman Germanech, but required more fine
tuning in Old Albic, where I still discover new phonological
rules - only a few months ago, I found out that Proto-Hesperic
*/s/ has become /d/ after nasals and liquids, for instance.

> [...]
> 
> > Right.  And the best way to emulate this in a conlang is to
> > start with a "protolang", i.e. another conlang that is then
> > subjected to a simulated linguistic history with sound changes,
> > semantic shifts and the like.  At least that is the way to go
> > if you want something naturalistic; it makes much less sense
> > with an auxlang or a loglang.
> 
> Well, maybe not, I've never done that. But going through a dictionary and 
> just randomly coining words is only going to get you so far.

Indeed, it doesn't get one far.  I tried it; it did not work well
for me.  I know that some people program scripts which churn out
roots which they then assign to lists of vocabulary entries, but
that's not for me.  I want my words to sound "right".  Working on
Old Albic or Roman Germanech feels more like *discovering* an
existing but unknown reality than like *inventing* something with
no prior existence.

> [...]
> 
> > Yes, that is true.  Poetry is a good performance test for a conlang,
> > better than primers.  So far, I have not written actual poetry in
> > Old Albic, but a prose version of a creation myth is under
> > construction, and it turned out that writing such texts is a good
> > way of rounding out a conlang.
> 
> I would say not just poetry, but any sort of writing which might have 
> particular concultural significance, like creation myths, folk stories and 
> declarations of independence.

Absolutely!  I found that working on the creation myth is much
more satisfying and helpful in working out the language than
just translating Bible portions or other texts that have nothing
to do with the conculture.  Such original texts give an insight
into the minds of the speakers of the language that translations
just cannot give.

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