Re: What criteria do you have for your own or others' languages?

>On Wed, Dec 06, 2006 at 12:27:18AM -0800, Sai Emrys wrote:
>[...]
> > Please list, in your own words (preferably detailed) what criteria you
> > apply to conlangs to judge them to be (in your opinion / for your
> > purposes) "better" or "worse", or more or less "likable" or
> > "impressive" etc etc. They can be subjective, objective, or both.
>[...]
>
>This may sound vague and indefinite, but I like conlangs that have
>"character". By this, I mean that it should have a "flavor" or "feel",
>(be it sound flavor, or idiosyncratic choice of phraseology, etc.) that
>is consistent, and unique. I dislike conlangs that feel like they are
>mass-produced, e.g., auto-generated vocabulary or completely rigid
>grammatical rules. (Not that I am not guilty of the latter myself.) Good
>music composers write in a "style" or "flavor" that often immediately
>identifies who the composer is, whereas mediocre composers write
>mediocre pieces that leave no lasting impression. Similarly with
>conlangs.
I agree: "flavor" is very important.


>On a more objective level, I like novel features in conlangs, such as a
>novel phonology or syntactic feature, say an unusual typology, or a
>clever way of expressing something that can only be expressed in a
>verbose way in another lang.
I share this sentiment too.


>But the over-arching criteria over all
>these is that these features must "fit", must form a coherent, workable
>system (not a hodge-podge of cool ideas the author chanced upon).
>Moreover, this system should bear the mark of its uniqueness,
>characteristics which immediately identify it, as opposed to some
>standard average non-descript twice recycled system.
>T
As far as hodge-podges of cool ideas go, I've just kept filling the
idea-pool as simply that, an idea-pool, not assigning everything to a lang
right away; but rather waiting to see which features will start suggesting
that they naturally "belong" to a common project. It might turn out
eventually that my pool get filled too fast, or slo, for this method to
yield successful results, but I'm not _yet_ officially abandoning anything
much, or putting too much thought into a "missing" part of some project.
'fcors this is not a design criteria, but a method... but I feel explaining
it might make the next parts clearer.

Now, naturally no characteristic can be necessarily uniq to a single
language - e.g. even the most valiant and original syntax will never imply
anything about the existence of an exclusiv 1P dual pronoun. Theoretically,
all combinations are probably just as valid, tho many would not be
naturalistic or even plausible. Still, even if accepting those tighter
limits, there would be a far greiter amount of possible languages than do
exist; a gazillion variations on each theme, each flavor. IMO it's not worth
the time to creäte more than one variation of something, however. Note that
this does not refer to diachronic close relativs, major re-editions & the
like, but rather doing the "same thing" over and over again in different
guises. It's hard to pin down what exactly constitutes "the same thing" -
the border will necessarily be fuzzy - but anything that's a more or less
direct clone of any existing lang (nat or con) will certainly count, at
minimum.

....I dunno about recycling hi'er level ideas; maybe it is possible to make
650 distinct and original languages of the type "a lang of natfamily A
influenced by a substratum of natfamily B", maybe it isn't. Whether these
kind of things start to seem tiresome must be a matter of personal taste
anyway, and I'm new to the scene & unable to predict what I will and what I
won't eventually get tired of, so I'll just shut up on this now.

Lo'er level ideas... generally anything concrete, from single features
upwards... as noted abov, the set of possible languages, even those that do
adhere to hi'er level design preferences, is still vast. Taken isolated,
case system X might in the designer's mind fit together with any of the TAM
systems 1 thru 5 (but not necessarily 6 thru 20) (and this part I find
fundamentally fairly arbitrary, so I won't bother explaining which is my
fave case or tense or rhotic or concultural setting.) My first solution to
narro' this down is to turn to the world of existing languages: I seek
"holes", combinations that do not exist yet but rather plausibly could. For
a quick example from phonology... To my kno'ledge, aspirated labial-velars
and similar co-articulated stops do not occur in any natlangs. This is
however not because such phones would be inherently more marked or unstable,
but simply just out of chance: they happen to be part of such Sprachbunds
that do not include aspiration. LabialIZED-velar aspirates *are* attested,
also. Hence, enter a labial-velar aspirated stop in one phonology sketch.
Nothing particularily groundbreiking required, in other words, but still a
dash of novility.

This naturally requires lot of research to pull off, and I likely might at
places end up going with a different roote after all, especially with
features that are not easily reduced to having vs. not having some discrete
properties, but it is what I *wish* I could do.

I have implied it already, but to be explicit - yes, I like naturalistic
languages. Diachronics should also play a part - IMO a history-less
naturalistic conlang is an oxymoron. I do like tossing in some engelangy
goals too, but at second priority to naturalism. Auxlangs I don't care about
at all.

There's also a fourth element I do like: xenolangs. Alienness. Poking around
to see what universals could be broken, and how, while still maintaining an
"internal naturalism". Obviously this is bloody easy with phonology and
semantics, no prob to toss in fifty vowels of seven phonations, or fourteen
distinct lexeme sets covering variations of "salt" (even if pulling it off
convincingly and internally consistently would still be impressiv); but
alienness in grammar, and I don't mean alien due to being too logical and
simplified, seems like a very intriguing concept to me. Not surprizingly, my
projects in this regard are progressing even slo'er than the naturalistic
ones.

When it comes to judging other's languages... in theory I would rate them
with largely the same principles I hold to myself; altho a well-done
principally logical engelang I do appreciate too. But I also second the
issue of clarity in representation. In practice, I rarely have the time to
concentrate on the lang's notes and tend to end up skimming a few discrete
and/or usually-presented-first features (phonology & script; case or pronoun
systems; etc.) and gessing from them, usually pessimistically, whether it's
worth to read more. I'd appreciate if design principles, preferences and
foci were stated right away, so as to not accidentally skip an intriguing
language on the basis of a section that was added only as a necessary evil.

John Vertical

_________________________________________________________________
Uutiset ja kasvot uutisten takaa. MSN Search, täyden palvelun hakukone.
http://search.msn.fi