OT: Realism? Re: Super OT: Re: CHAT: JRRT

(Note: I have read your whole response, but I'm only responding to this
part.)

And. wrote:

<<I think verisimilitude is a major ingredient of what I most value
in an artlang, too. How do you judge 'realism'? To me, it's mainly
a matter of complexity, of scale, and of completeness. The more
complex, the more large-scale, and the more complete it is, the
more realistic it is. If this is enhanced by something like
the Joseph/Alma Walker framing of Tepa, the delightful effect
is intensified. This means that I find a conlang sketch that
fits neatly into orthodox natlang typological patterns less
realistic than, say, Teonaht, which is fantastical and is created
by someone who by present-day conlanger standards knows
comparatively little about linguistics and doesn't understand
fully at a conscious analytical level, as opposed intuitively,
how her conlang works. On the other hand, languages that seem
to go beyond what is plausible for a human language, such as
Ebisedian and Ithkuil, I don't find realistic (though they
have other attractions).>>

Ahh...   I see, then, that we actually disagree on this point.   Taking your
three criteria...

1.) Complexity
2.) Scale
3.) Completeness

...let me judge my first language, Megdevi, by it.   While it was, by and
large, regular, it was complex, with its triconsonantal roots and the way you
used the different forms to create different words.   In other words, it's no
Esperanto (and even that, as has been stated before, isn't exactly intuitive for
every person on Earth).   As for scale, it's still, to date, my largest
language, having somewhere between 5,000 and 10,000 words (it's hard to count them),
and the list can be added to effortlessly, since all I need to do is come up
with a new triconsonantal root and a new semantic idea.   As for completeness,
well, I have claimed, and still claim, that I can translate anything into it,
which I can't say about any of the languages I've created since.   I even
finished translating much of The Tempest, and had started on Coriolanus when I
realized something: The language (by the standards that I developed) sucked.
Here are the reasons:

1.) Though it had a large "phonology", it wasn't a phonology at all: It was
just a list of sounds.
2.) Though I had created a very complex method of creating words, the words
will still basically English.   The example I always give is the root relating
to "fortify" (I forget what it is).   By making it a verb, you made "to
fortify".   By making it into a verbal noun, you got "fortification".   By making it
into a human noun, you got "fortifier".   By making it into a utility noun,
you got "(a/the) fortification".   And all these words were defined *exactly*
as they were in English.   So, when you got right down to it, it was a very
complex, very flashy-looking coding experiment.   (In its defense, it did have an
Engelangy element [I used my new word!], in that my intent was to create a
triconsonantal root for every possible semantic idea.   I don't think I realized
that this was impossible at the time.)
3.) The language has prefixes, suffixes, infixes and circumfixes, and by
looking at them, they just don't fit together.   If you look at all the various
affixes, it looks like you had a different person create each affix, and that
these people didn't talk.
4.) The language is kluge-filled.   Whenever I encountered some concept I
wasn't aware of at the time, I just added another affix, without thinking about
how the concept should be realistically built into the language.
5.) I gave no thought at all to syntax.   Word order was completely free.

In short, I made no attempt to work within the identity of the language, or
to imagine how the type of language I had created would solve the problems I
encountered: I just round-peg-square-holed them.   I no longer work on this
language at all; I just keep it around because it was my first (though the script
I created lives on, as it's an Arabic-style script that happens to be
perfectly suited to English.   I write notes in it in class all the time [though
damned if I can read them later.   I wonder why that is...?   I'm a perfectly
literate writer (or speller), but I can't read it at all]).   Nevertheless, by your
three criteria (as I understand them), it would probably be a pretty good
language.   Now, as I parenthesed above, you can probably further define your
criteria to rule out Megdevi, but I still contend that the spirit of the criteria
is wholy different from a criteria I'd come up with, which might be something
like...

1.) Does language X have an internal logic?
2.) Does language X follow this internal logic?
3.) Could language X plausibly exist on Earth?

I think my third criteria is somewhat the same as something you stated in
your original reply, but I think we'd probably judge it differently.  For
example, for (3), I'd rely on typological information.   This doesn't mean that a
language has to obey every universal (most natural languages don't [or do any?]),
but it should be passively aware of them.   By that, I mean a realistic
language will violate some universals, not none or all.   Most importantly, it will
have the look and feel of a language, and, admittedly, this is hard to
quantify, and will differ from person to person.   However, if we're considering
natural languages, I don't think anyone would doubt that a language that did the
following...

(1) Singular nouns end in: -i
(2) To pluralize a noun, add: -embarkadoristo

...would be unrealistic.   In fact, this kind of intuitive knowledge of what
is and is not realistic was the whole basis for the recent "Wacky Language"
thread.  If we all had no idea, or radically different ideas, of what a
realistic language looked like, none of the posts would have been amusing, because we
all would have looked at something like the above, and said, "Yeah?   So?   My
language pluralizes nouns by changing a singular noun to its corresponding
singular noun in Finnish" (hee, hee...   *That* would be fun: A language that
depends on extensive knowledge of another, totally different language in order
to be spoken).

Anyway, complexity can be relative.   What about a created pidgin?   They'll
be markedly less-complex, but will be *appropriately* complex.   (One could
add that to the criterion, but then there's the question, "What's appropriate?")
  Scale is a difficult to define term.   I think I defined above as
"largeness", but maybe you intended "the intent to be large"?   I think this would
favor language with conculture over languages without--and, who knows, maybe this
is a good thing--but I still don't think scale would be a good way, since (I
assume?) a large scale language would automatically be better than a small
scale language.   Depending on how you define "scale" with relation to a conlang,
this might not be a good idea.   And, as for the last criterion,
"completeness", I admit that this is an admirable quality, and something that probably
everyone attempts to achieve.   But it takes time.   A *lot* of time.   So I might
change this to "the intent to complete", until the language is done (which
would mean that the creator no longer adds to it, for whatever reason).   Even
so, completeness can be applied in two ways: Completeness of grammar, or
completeness of language (which would mean that every word that is going to be in
the language is in the dictionary somewhere)?   One could conceivably complete a
grammar (and there are some of us that have), but word creation is a lifelong
process.   So whether or not I'd use completeness as a way to measure would
depend on how it's defined, again.

So I'd be curious to know what you think about this.   Do we have two
completely different opinions about what is realistic in a conlang?   If so, I think
that's pretty interesting.

-David

Replies