Re: vocabulary

Hallo!

On Wed, 15 Dec 2004 11:57:42 -0700,
Dirk Elzinga <dirk_elzinga@...> wrote:

> On Dec 13, 2004, at 5:46 PM, Roger Mills wrote:
>
> > # 1 wrote:
> >
> >> How did you make to create a complete vocabulary?....(snip)....How do
> >> you
> >> create a whole lexicon? Particulary if you want it to not be as
> >> another.
> >
> > Quick and dirty method: pull words out of the air. You should have at
> > least
> > some idea of what constitutes a permissible "word" in your language, so
> > you're allowed e.g. /bik/ and /brik/ but not /bnik, bzik/ etc.
> > Make an alphabetical list of these, with the associated meaning.
> >
> > Vastly better: create the phonology and rules of syllable structure.
> > Then
> > generate a big list of possible words-- there are said to be ways of
> > doing
> > this with computer programs, and I hope others will suggest some--
> > otherwise
> > you can do it by hand.  I recently hand-created a list of all ±40,000
> > forms
> > for a CVCV ~CVCVC language; it didn't take that long. Of course if
> > you're
> > creating Neo-Georgian with a complicated phonology it might be a
> > little more
> > difficult.
> > Then you can assign meanings at random and as the spirit moves you.
>
> I think it's an open question which method is best. I also generated a
> list of phonotactically acceptable forms to use as roots in
> Miapimoquitch. I pick and choose as I need them. The advantage is that
> I know ahead of time that the forms "fit", but the disadvantage is that
> I'm pretty much locked into those phonotactic patterns (if I want to be
> consistent). Of course, there's nothing to prevent me from going
> outside the canon and making up oddly formed roots, but if I'm going to
> do that often, there really isn't much point to generating the list in
> the first place.
>
> The first method, pulling them out of the air, *I* think has the
> potential to be vastly better, since it allows the language to grow
> organically, at least phonologically. When a sufficient number of
> vocabulary items have been created, you can then see what the
> phonological patterns are. It seems to me that this can be very
> self-revealing, letting you know where your own phonological
> inclinations lie.
Yes.  I don't even know exactly what the phonotactic rules of Old Albic
*are*, so I simply cannot generate a list of "acceptable" word-shapes.
I just make up words and intuitively decide whether they sound "right"
or not.  What has begun to crystallize out is that most words are
derived from CVC roots, with some occasional CCVC (mostly CRVC,
R=liquid or semivowel) and CVCC (mostly CVRC) roots strewn in.
What I definitely have, though, is a well-defined phoneme inventory
and a handful of rules (such things as Grassmann's Law, metathesis
rules, rhotacism or umlaut patterns) operating on them.

>       And has said before that some of the best projects
> out there are done by people who are linguistically very sensitive, but
> otherwise not formally trained.
Yes.  I agree that a feeling for language is useful in conlanging, but
formal training in linguistics is far less important.  It is the same
as with music, for example: a good musician needs a feeling for music,
but he doesn't need formal training in music theory.  It has been
claimed that musicologists, when they compose at all, compose soulless,
overintellectual music (though it has also been claimed that this
is in no way the case; I don't know who is right), and it might be
similar with linguistics and conlanging.

>       The reason, as I understood him, is
> that there are no preconceived notions of allowed or disallowed
> structures to artificially constrain the language. I think the same is
> true for vocabulary generation out of thin air; there are no
> preconceived phonotactic patterns which must be adhered to, only the
> emergent patterns which are discernable with sufficient items. Of
> course, it takes a long time. I've been working on a new project off
> and on for about two years now, and I have about 20 words to show for
> it. But they are gems!
20 words in two years - and I thought I was slow in building
vocabularies.
(I have about 300 Old Albic words so far.)  But I know what you are
talking about; it is one thing to randomly assign meanings to a list
of word forms, it is something entirely different to coin words that
sound "right" - but the latter is infinitely more rewarding!

> > There are also vocabulary lists, usually grouped together in some
> > logical
> > way (Carsten Becker recently posted one; there are others-- IIRC the
> > Langmaker site links to the "Universal Language Dictionary"); but it's
> > just
> > as simple to think of, say, colors, kin terms, verbs of motion, or
> > household
> > items, and just start free-associating. One thing leads to another.
> > It's
> > basically how I created the Kash vocabulary, now approaching 5000
> > words or
> > so.
>
> This is probably a good tool for vocabulary generation ex nihilo as
> well.
Yep.

Greetings,

Jörg.

Reply