Re: Help for a new conlang

Roberto wrote:

<<But I've got no idea of how to do O:-) And I'd like to know how
semitic languages (or whatever are similar to what I'm proposing; maybe
I'm unknowningly referring to another kind of language) behave on this
aspect.>>

The very first language I ever did (before I knew that other people actually
created languages for fun) had a triconsonantal root system.   This is the
meat of it (oh, the 1, 2 and 3 are the first, second and third consonants in
the root.   I'll list the roots beside the entries, where relevant):

NOUNS:
1.) Natural Noun: 1e2I3, as DezIZ (writer)   ROOT: D-z-Z, "writing"
2.) Verbal Noun: 1i2ej3at or 1i2ej3It, as DizejZat (writing)
3.) Utility Noun: 1&2I3, as D&zIZ (writing utencil)
4.) Place Noun: 1i2E3, as DizEZ (a place in which one writes, office, study)
5.) Object Noun: 1A2o3, as DAzoZ (written item, document)

ADJECTIVES:
1.) Natural Adjectives: 1a2i3, as maZil (beautiful)   ROOT: m-Z-l, "beauty"
2.) Nominal Adjective: 1U2u3, as mUZup (butterfly-like)   ROOT: m-Z-p,
"butterfly"
3.) Verbal Adjective: 1a2i3ad, as Zi DaziZad devId (the writing man)   NEW
ROOT: d-v-d, "male"
4.) Object Adjective: 1u2&3, as Zi Duz&Z mikejlat (the written word)   NEW
ROOT: m-k-l, "speech"
5.) Place Adjective: 1&2E3, as h&lEf DezIZ (a heavenly writer)   NEW ROOT:
h-l-f, "happiness"
6.) Color: 1i23a, as vitja (orange)   ROOT: v-t-j, "orange (the color)"

Then verbs have 1a2a3 for imperfect, 1I2I3 for perfect, and you can add tense
suffixes.   There're other forms, but these were the main ones.   The whole
thing is hopelessly regular.   Real root systems aren't as simle.   Example
from Arabic:

If you were to look at the words /waalid/ (father) and /t_?aalib/ (student),
you might guess that the human infix set is 1aa2i3.   Not so.   Take the
words /ibn/ (son), /s_?adiiq/ (friend) and /Zundii/ (soldier).   [I actually
know a different word for "soldier" that fits the pattern, but this serves
our purpose.]   And the same goes for other classes of words.   The
prototypical adjective pattern is 1a2ii3, so you have /saGiir/ (small),
/Zadiid/ (new), /qasiir/ (short), /kabiir/ (big), etc.   Then you get kinks,
though.   You have the word for "good" which is /'Za.jid/, rather than the
expected, *Za.'jiid   Also, if you see the word /Zaziira/, it'd probably look
like an adjective with the feminine ending /-a/ (you add that to any
adjective when it modifies a feminine noun), but instead it's a noun--the
word for "island".   Then you also get adjectives like, /Gajuur/ (jealous)
which come seemingly out of nowhere.

These are just two patterns, mind.   There are many, many, many.   And while
there are diachronic explanations for just about everything, they're
synchronically irregular.   Dig?   There are prototypes, and you can say
which types of words are prototypical members of the group that any pattern
or wazan represents, but you can't say for certain, for always.   It's a
continuum, kind of like noun classes in languages that have them.   So my
advice is to keep that in mind when designing your language.   If I had mine
to design over again, I'd find some way to build the irregularity into it.

-David

"imDeziZejDekp2wilDez ZejDekkinel..."
"You can celebrate anything you want..."
               -John Lennon

Reply