Re: OT: What makes a good conlang? (was Re: Super OT: Re: CHAT: JRRT)

T wrote:

<<Be that as it may, I do not believe that human language is as easily
systematized as some linguists would have us believe, so I am not
apologetic in the least in saying that Ebisédian's case system is actually
quite intuitive, even if my first attempt at it may have gone a little too
far.>>

There have been a lot of of comments in this vein (e.g., a linguistics major 
doesn't make the best conlanger, people who don't know anything about 
linguistics often make better conlangers, people who follow their intuition make more 
realistic languages, and the above), which I wanted to say something to.   
Now, I don't disagree.   In fact, I agree with what you wrote here--especially 
when you consider universalists, who've taken the view that not only can all 
languages be systematically described, but they can all be systematically 
described in basically the same way, with a handful of language-specific stipulations 
(which usually amount to, IMO, the language itself).   The counterexample to 
this would be a linguist who assumes that a language can only be understood 
within itself, and so the linguist attempts to discover the systematicity of the 
language on its own, rather than on comparing it to other languages, or 
trying to make it conform to universal constraints.   While I value this method 
more highly (for a natural language), I still agree with Teoh (I'm sorry, but is 
this your name?   I get the impression that it's your last name...   :(   I'm 
sorry!), and say that if such a linguist assumes that every aspect of the 
language can be systematically explained, from within or without, I think s/he'd 
be assuming too much.

However.

From a design point of view, I think one should consider realistic 
anachronisms.   Just because something can't be explained systematically doesn't mean 
that it should be totally off the wall.   So, let's say you have a language that 
allows only CV syllables; there are no codas whatsoever.   Then, however, you 
say that to form the plural for one particular word (and this is the only 
word of its kind), you add an /-f/.   This would be a totally unrealistic, 
unsystematic anachronism.   Now, though, let's say you had a whole class of 
words--small, though it may be--that all take a plural in /-f/.   Let's say that this 
class is predictable, based on one feature (whatever that feature may be: 
phonetic, morphological, semantic, whatever).   Further, let's say that with some 
of these words, the pattern isn't even very strong--some speakers have /-f/ as 
a plural, some have the standard form, whatever that may be.   Now you've got 
a rather unsystematic (fairly), completely nonproductive, universal-violating 
anachronism in your language, but it's more believable because it gives the 
hint of a previous status, i.e., it may be indicative of a productive process 
from the past.

A better example, I think, would be with noun cases.   In your standard 
nom/acc language, the nominative case is the one assigned to subjects.   Sometimes 
you can get a genitive subject; sometimes a dative.   Generally, though, 
nominative is the subject case.   Then let's say, though, you have one verb where 
the subject gets the accusative.   Only one.   Now, a linguist can speculate as 
to why this is (a good guess would be to examine the semantics of the verb), 
but with only one form, you can't say anything about this form save that it's 
an exception.   And when you have performative verbs, ECM verbs, control 
verbs, etc., you can get these weird cases where there's like one verb out of all 
the verbs in the language that assign a weird case to its subject.   Yet, it's 
more believable if the case that this verb assigns is one of the core cases 
(Nom., Acc., Dat., Gen.), and if there's at least the hint of an explanation, 
even if there isn't really an explanation.   If you had a language that had the 
above four cases, plus an instrumental, a locative, an allative, an ablative, 
an oblique and a vocative, though, it'd be very strange if you had a singular 
verb that assigned allative case to its subject.   This isn't to say you 
couldn't have a class of verbs that assigned the allative case to its subject (I 
can dream up a somewhat plausible scenario), but it'd be really weird if you 
only had one, whereas it wouldn't be so strange if it assigned one of the core 
cases.

With a phonemic inventory, let's say you had...

p t k
b d g
m n N
f s x
v z G
w l j

It'd be particularly strange if you had, in addition to these, a voiced 
lateral fricative.   It'd be totally out of left-field.   It'd be less strange if 
you had /q/, though.   First of all, there are plenty of languages that have 
/q/ but not its voiced counterpart.   Along with this, there are lots that have 
/q/ but not its nasal counterpart as a separate phoneme.   And, given the 
system above, you can imagine how, if you had /X/ and /R\/, theycould merge with 
/x/ and /G/.   Now, both the addition of /q/ and the addition of a voiced 
lateral fricative would be unsystematic, but the addtition of /q/ as opposed to a 
voiced   lateral fricative would be a more realistic unsystematic addition.

So, what I would say is that a knowledge of linguistics and typology can 
still give you an idea about realistic unsystematicity.

Joerg wrote:

<<People who follow their intuition often create better and more
realistic art than people who try to be exact.>>

The only caveat I'd add to this is that, with visual art, for example, 
there's little chance of someone being influenced by anything other than their own 
intuitions if they choose to follow simply their own intuitions.   In 
conlanging you run the risk of emulating your L1, or any other language you know.   So 
I'd say it's important to draw a distinction between the instinct of what 
sounds right/makes sense, and what sounds right/makes sense *within* the framework 
of the language one's inventing.   After all, I think we've all probably seen 
examples (and, indeed, I've *been* an example) of someone doing something 
because they think it sounds/feels right, and what they do ends up emulating 
English, or some other known languages, almost exactly.

-David

Replies