Re: Unilang: the Morphology

On Fri, 20 Apr 2001 18:10:06 +0000, Raymond Brown <ray.brown@...>
wrote:

>No, no - wouldn't work!  That's why if we don't have traffic lights at busy
>junctions, we have a roundabout instead (forgotten what roundabouts are
>called in the US).  You got to have something, even if it's the simple old
>French rule: "Prioritée à droite".
Roundabouts - a good point in this analogy :)

>>Can syntax do the job?
>
>The answer is obviously "yes" - there are natlangs that to prove that.
>
>>Can prepositions and other similar items do the job?
>
>Yes - take a look a real natlangs used by millions of people every day.
>The majority do not use case endings.
>
>>Are those any easier?
>
>Why are they the predominant type in pidgins & creoles?
Perhaps you misunderstood the general tone of my letter :) I don't actually
want abstract and pedantic endings flying around; my questions were more to
show that we need to ask ourselves these things first, rather than being
intended as challenges.

I pretty much agree with you, actually. I've usually taken the analytic
character of pidgins and creoles as proof that that must be the simplest
and easiest grammar.

>Ok - forget the dogmatics of Auxland, but I would advise taking a good look
>at what actually does happen in pidgins and, IMO more importantly, in
>creoles.
Certainly. This list is sometimes one of the best sources of information in
these matters, but I should see what I can find in the library. Anyway,
comments from a "creolist" would be very constructive. Still, I wouldn't
say that I know especially little about the character of those langs; I
have quite workable knowledge of general Chinese grammar (which may not be
creole, but we might call it "creoloid").

>Personally, I've found case-less Welsh easier to get on with than Gaelic
>with its four cases.  It seems to me that particles like the Japanese ones
>I gave above and the modern English possessive particle, which can attach
>itself to whole phrases, do the job just as well as and are more flexible
>than case endings, cf.
>"That's the guy next door's dog."
Bull's eye! :) I have no interest in IE-style omnipresent case-marking;
simple clitics that mark grammatic information concerning whole phrases is
definitely good; whether it be case, number, gender, tense, mood, aspect,
etc, anything would be well marked that way.

But I certainly think, also, that syntax should do a good share of the case-
marking job. Nominative ~ accusative, for example, is probably best taken
care of by a fairly rigid SVO order, as per English and Chinese.

I'd like to phrase here an "underlying" working rule of mine: the Unilang
should represent an intricate _balance_ of features; it should not be
characterized excessive reliance on any one feature, e.g. syntax or
morphology, nor by excessive phonetic simplicity nor complexity. Balance is
the key-word, as I see it.

Regards,
Óskar