A sincere welcome to the list, monsieur Babb!  Hope you enjoy it here.


I see Jim has already called me and my random conlang generation project
out.  The guiding principle of my project is naturalism.  It moves slowly:
the components that are extant are listed at
  http://000024.org/conlang/random-language.html .
Right now I'm working on Gleb, the phonology component.  After that gets to
a respectable state, I mean to make the morphophonology component (see
http://en.wikipedia.org/wiki/Morphophonology ) and work on the morphology
component some more.  

There's nothing I wouldn't like to do with the project, eventually; I'd love
it if it could open its hatch and roll out a complete well-documented
conlang, just as you said.  But lexicon seems by far by far one of the
hardest aspects of language to do justice.  At the root of the problem is
simply the richness of the world.  Having a good enough understanding of the
world that you can generate realistic semantic spreads and contrasts and
derivations and fixed combinations and ... seems tantamount at least to
having some sort of ontological concept database better than any out there
at present, if not to completely solving strong AI.  Worse yet, lexicon
reflects culture in a pretty strong way -- you may believe Sapir(-Whorf)'s
hypothesis more or less, but certainly a culture that didn't know the
vehicle would be unlikely to have a name for the spacecraft, and everyday
English doesn't really have words for kin relationships defined in the
Warlpiri system of sixteen skin-groups -- c'est a dire, naturalistic lexicon
generation presupposes naturalistic culture generation!  
So I may not ever get there.  

In any event, it would be awesome if anything you write ended up being
compatible with things I write or have written.  


To add a few cents to the previous responses:

On Tue, 25 May 2010 17:16:02 -0600, Dave Babb <[log in to unmask]> wrote:

>I see some areas of being a conlang that can be open to human error,
>namely, the construction of the vocabulary. Lots of room for conlang
>developer overlap.

As Jim points out, whether and why this is a problem is quite a different
matter for naturalistic languages and languages engineered to care about it
(let me subsume auxlangs under the latter).

Personally?  When I discover in my own artlangs that I've created a
homophone (whether by sound change or just unmotivated reuse), I mostly just
let it pass.  If it's contextually awkward (moreso than a bat (the mammal)
and a bat (the implement)) I'll change one of them.  In any case, (to the
best of my knowledge!) keeping a single master lexicon with the words in a
canonical sort order has addressed this completely for me.  

If the question is synonyms, well, then, that's one to pay attention to.  A
language in (actual or fictional) use generally _can't_ retain exact
synonyms.  The synonymous words will pick up distinct senses and
connotations from the other words or constructions they often occur with, or
the segments of society that use them, and push against each other 'till
they have a reason for being different.  If not, one of 'em will probably
just be dropped as redundant.  

That said, any naturalistic language will have a healthy number of
near-synonyms, particularly in areas where there aren't clear-cut boundaries
between the senses (e.g. characteristics of people).  Take as your image of
the division of semantic space, instead of a neatly tiled floor, a floor
strewn haphazardly with pancakes until you mostly can't see the linoleum
anymore :-)

>For step #1: vocabulary
>only. Casing (if any) tenses, and possesive's will come later. Any
>gender specifics will be included in round #1, be that pre, insertive,
>or postfixed indicatives.

To say what I think you mean in the jargon, this is the distinction between
_inflection_ and _derivation_.  To take a stab at differentiating them for
the layman, derivation is changing a word into one of a different meaning
(for instance adding the meaning "a masculine X"); inflection is just adding
some mandatory information that the word needs to be able to serve in a
grammatical sentence.

>No I am not comfortable using all the "genitive, nominative, accusative"
>terminologies. I don't understand the conlang jargon completely. I
>understand the concepts only in everyday terms. I've had a small stroke
>so give me some defilade here.

Of course, no pressure.  But I do hope you'll warm to them.  Languages do
many mighty weird things, and there simply aren't ordinary names for all of
the possibilities...

>Right now here is the feature set in my head:
>
>a)    Have bound into the software package a 2000 word vocabulary. Think
>of the 850 word "Basic English" vocabulary fleshed out. This file will
>be user editable.

For a wordlist, you could do worse than to use one of the ones that has
already been compiled and looked over by many eyes off the conlang list.  I
like the wordlist of the Intercontinental Dictionary Series of the Max
Planck Institute, Leipzig:
  http://lingweb.eva.mpg.de/ids/

Incidentally, as Jim points out, everyone who comes across Basic English
should be aware that
| *Ogden and Richards cheated.* They were able to reduce the vocabulary 
| of Basic English so much by taking advantage of idioms like "make good" 
| for "succeed". That may save a word, but it's still a lexical entry that
| must be learned as a unit, with no help from its component pieces. Plus, 
| the whole process was highly irregular.
( http://www.zompist.com/kitlong.html#howmany )

>b)    The ability to enter in the conlang consonants, vowels, dipthongs
>in Roman character set, or in a developed conlang loaded font.
>
>c)    Control of the target word's syllable count. As in "same",
>"changed/specified", "random". "Changed/Specified" will have a user GUI
>user definable field. AKA source = 3 then target = 5; source = 2 then
>target = 7, etc. "Random" is exactly what it means at face value. There
>will be a user definable "min" and "max" fields.
>
>d)    Control of the target word's stress or accent. Same options as "C"
>above. In the Roman character set, stress will be indicated by all caps,
>or in a user loaded conlang font, by being underlined.
>
>e)    A user definition of what a sylable can be. It will be a multiple
>entry drop down box. Definitions will use the letter "c' for consonant,
>and "v" for vowel. AKA "ccvc" for one entry, and "vcc" for another
>entry, and "cvc" for yet another acceptable entry.

All of these, except for the native writing system part of (b), are (or will
be) done in a bundled way by my Gleb, and I believe the same can be said of
Jim Henry's Boris.  It would therefore be nice if you provided an option
that just delegated the whole ball of wax.

More later, perhaps.  

Alex