Print

Print


Respondante al plifrua mesag^o mia,=20
je 05:03 AM 9/29/97 -0700, Mark P. Line skribis:

>But my knowledge of conlang IAL candidates is hardly at issue here.

No, only to the extent that after recovering from my surprise I wondered=20
if you judged yourself to have a sufficient basis for comparison of=20
"simple" natlangs and conlangs.  As I mentioned, I've heard so many=20
opinions from people who hadn't even looked at the question, I wasn't=20
really very interested in challenging them, there was nothing there to=20
challenge!  But it interests me a great deal that you have experience=20
with Esperanto, yet feel that certain natlangs are comparably=20
uncomplicated, or even *less* complicated.

>James had stated that natlangs can be discounted categorically as
>possible IAL's, because they're too complicated. Although he has backed
>off from his hyperbolic formulation, his (and presumably your) basic
>assumption remains unshaken.=20

Well, previously I had no valid evidence to question the assumption,=20
although I recognize that lack of evidence does not constitute proof! =20
Based on your information I certainly no longer think that it=20
necessarily is true that a natlang must be complicated and exception-
rich.  However, that is how I would describe English, French, German,=20
Spanish, Greek, Russian, Hungarian, Tamil, Sanskrit, and incidentally,=20
Interlingua, which retains many complexities of its Romance source=20
languages. =20

Just to be sure I understand:  in comparison to Malay, Tok Pisin, or=20
Quechua, do you find any of the ethnic languages listed above to be=20
of comparable simplicity?  Of course, ideally I should go study those=20
languages myself, but I lack the time to do so in the foreseeable=20
future.  I may even have to cut back on my time on auxlang list.  :-(

>My intention has been to show that this assumption is untenable to the=20
>extent that there are some natlangs which are quite uncomplicated.=20

I understand.

>If I can show this assumption to be untenable,
>then lines of argument must fail which discount the possibility of
>natlang IAL's categorically because "they're too complicated".

That would indeed follow.  There are, of course, other reasons some=20
people might prefer a conIAL over a natIAL, but let's not open that can=20
of worms yet!

>> I guess my point is this:  I've heard people who know
>> nothing about Malay or Esperanto claim that Malay is "just
>> as easy to learn" as Esperanto.  So may I ask, which
>> conlangs have you studied, and to what degree compared to
>> the natlangs you mention?  You see, I do remain surprised
>> when you say that Malay, etc., are as simple as Ido or
>> Esperanto, or ....  Yet if I understand you correctly, I
>> think this is what you have said.
>
>I hope that I have never said anything as stupid as "Malay is just as
>easy to learn as Esperanto".=20

I'm sure you didn't say it that baldly.  Please view my comments=20
as a sort of conceptual short-hand.

>At the very least, such a claim would have
>to be relativized for numerous demographic factors.=20

Yes.

>And besides, "easy
>to learn" is a whole nother discussion -- we were talking about
>"relative complicatedness here". Ease of learning would involve
>similarity of vocab to one's L1 and L2's, whereas complicatedness is
>something that (we assume) would be captured as a property inherent in
>each language by itself -- how it compares to anybody's L1 wouldn't
>enter into it.

Yes, I too was talking about ease of learning because of simple patterns=20
and few exceptions, *not* about familiarity with vocabulary, which=20
depends on the languages one already speaks.

>I guess
>we could define "complicated" in terms of the amount of information
>necessary to describe the grammar of a language, using a descriptive
>mechanism (formal, semi-formal or prose) that allows comparison across
>typologically divergent languages.=20

Yes, this is the kind of thing I was thinking about.  I feel that there=20
*should* be some way to speak of absolute complexity of languages,=20
unrelated to any student's previously learned languages.  Whether the=20
universe fits my world-view is, of course, another issue.

>In that case, I would have to say
>that I find Malay to be less complicated than Esperanto.

Wow.

Well, this is precisely the type of complexity measure I had in mind,=20
although I could not by any means rigourously define it.

All I can say is, you surprise me!  And I can't help wondering if=20
perhaps we are still not quite on the same wavelength.  You mentioned:

>My knowledge of Esperanto is similar to that of the natlangs I mentioned
>(and of some natlangs I didn't mention, of course). I once had the
>opportunity to study in detail the results of a corpus-based grammar
>project at the University of Saarbruecken, in which things like
>preposition usage and verb valencies were exposed descriptively.

I realize that there are areas in Esperanto where current usage selects=20
one of many allowed possibilities, and it has occurred to me more than=20
once that it would be harder to define Esperanto as it is used than to=20
define the range of allowable statements within Esperanto.  This hasn't=20
bothered me much, because I feel I have every right to diverge from the=20
"normal" usage without being criticized, precisely because Esperanto is=20
a constructed language.  I rather like the idea that Zamenhof left much=20
of it unspecified, so within a certain range one can imitate natural=20
languages without leaving Esperanto.  So I wonder if your "more complex=20
than Malay description" of Esperanto takes into account idiosyncrasies=20
of actual usage whose presence or absence is immaterial to the core of=20
the language?  Based on my experience with the language I'd have to=20
say that I have a very difficult time imagining a language which could=20
be more succinctly described.  But that may indicate a lack of=20
imagination on my part rather than any information concerning languages.

>> But I notice your careful phrasing in both paragraphs
>> above:  "more complicated than they need to be".  It may be
>> impossible to define this concept.
>
>That is _James'_ careful phrasing, not mine -- I've merely been quoting
>him.

Ok.  Then I'm uncomfortable with _his_ phrasing.

>Note that I responded to James with my suggestion of Malay, Quechua and
>Tok Pisin/Bislama only after he made the more objective request
>concerning possible natlangs which I believe to be no more complicated
>than Ido.=20

I understand.  I was also very interested in your answer.  Oh, by the=20
way, I myself find Ido and Esperanto to be of comparable complexity, viz.=20
much less than natlangs I have studied.  =20

>> But could we agree that a language with separate "formal"
>> and "informal" vocabularies would take longer to learn (as
>> an L2) than a language having one vocabulary with an extra
>> "honorific" word to be placed in a formal sentence, or
>> perhaps an "unhonorific" for informal sentences?
>
>I suppose you could approach the question empirically by creating a
>minimally-differenced conlang to compare with a natlang that does its
>honorifics in a particular way. I suspect that you would find that your
>results were significantly dependent on the learner's L1 and L2
>background. The same would be true if you created a
>minimally-differenced conlang, to be compared with English, in which
>animacy is marked by a prenominal particle.
>
>In other words: No, we can't agree that a language with lexically coded
>honorifics would take people of every conceivable L1/L2 background
>longer to learn than a language with analytically coded honorifics.

I may be missing something obvious here, because it seems to me that=20
whether or not the language grammatically resembles previous learned=20
languages, a reduction in memory load of vocabulary to be learned (which=20
I'm assuming is unfamiliar in both cases) would result in a gain of=20
(relative) ease in learning.

>I think that formulation must be equivalent to the idea of relating
>"complicatedness" to the amount of information necessary to describe the
>grammar of a language. Clearly, a universal pattern with no exceptions
>takes less information than a less-than-universal pattern plus a list of
>exceptions (and probably less than two patterns which, taken together,
>are universally productive).

Agreed, with the understanding that (1) even equally productive patterns=20
could be more or less difficult to describe, and (2) I'm unsure of the=20
extent to which a pattern's ease of description corelates directly with=20
the mind's ease of assimilation of the pattern.  Still shortness of=20
description is a good, concrete basis to start from, a concept hallowed=20
by use in information theory, relating to low entropy, etc.

This is fun!

Bonan tagon,
Ken

  *-----------------------------------------------------------------*
  | Ken CAVINESS           Physics at Southern Adventist University |
  | [log in to unmask]             ESPERANTO =3D Lingvo internacia |
  | http://southern.edu/~caviness/     E-o/English/Fran=E7ais/Deutsch |
  *-----------------------------------------------------------------*