Re: Lojban (was Copy paradox)

Lee Daniel Crocker (lcrocker@mercury.colossus.net)
Mon, 17 Nov 1997 13:43:54 -0800 (PST)


> ...This does not mean that the fundamental fact that Chomsky
> recognized is invalid, namely, human minds are built for language, build
> it in childhood using the terribly poor, incomplete examples around them,
> and invariably build it along identical lines. This has earned him the
> eternal wrath of cultural relativists who want human nature to be
> alterable; of course, we want our nature to be alterable, too, but
> transhumanists recognize we need to change the hardware to do so.

I'm certainly not big on ducking reality; I have no doubt that our
minds are wired by evolution in specific ways, and that it is good
to study those ways. I have little patience for those who believe
otherwise for whatever reasons. But what better way to study them
than by deliberately attempting to flout them? Indeed, how else
/can/ one study limits and capabilities except by trying to exceed
them (and learning from our failure)?

> The acid test for Lojban would be, can children learn it and not creolize
> it? If children are exposed to Lojban as a first language and what they
> create based on the evidence available to them is basically the same, it
> conforms to UG. If they alter it radically or minutely, then the
> Lojbanists have made some fundamental mistakes about how to create a
> language usable by humans which makes logical relationships explicit.

If those children do alter it, then /that fact/ will tell us more about
the nature of the fundamental capabilities of the human mind than any
amount of study on existing languages. Far from being a failure of the
Loglan project, it would be a great success (though it would be a
failure of the language itself, perhaps). It is precisely because Lojban
is so alien to the way things are done in evolved languages that makes
it a useful data point. I also think it would be a useul language--
perhaps not as a primary, but as a secondary one. Both of those
conjectures are subject to experiment. I participate in the latter;
I'll have to wait until I have children to participate in the former.

> Age of a science is not a requisite for the accuracy of its theories.
> Yes, linguistics is a young science. Our universe has existed for more
> than 15 billion years; is it irresponsible or hubristic to draw
> conclusions about the way it behaves because physics has existed for a
> tiny fraction of that?

Touche. That was a rather stupid thing for me to say. The idea I
was trying to get across is that for the sake of studying how the
human mind is wired for language, the set of all human languages is
a rather small sample. I'd like to see more outliers in this set:
Lojban is a good one, but I can imagine even more strange ways to
construct a language that might tell us more.

> > I am not one who takes a such things as
> > degrees or books sales seriously; I judge the final product, and
> > for my money, the Lojban group has done about 10 times the valuable
> > work in linguistics than any modern Chomsky-worshipper I know.
>
> Why do you have such animosity for Chomsky or researchers working in his
> wake? What, exactly, have Lojbanists actually done? Made a (possibly)
> natural language that computers can parse better than English or Finnish?
> That says more about the limitations of contemporary computers compared
> to the parsing (and, more importantly, interpretation) ability of any
> human being. Lojban aims to minimize ambiguity. Check. English contains
> sentences like "Time flies like an arrow," which can be parsed any of a
> number of ways by a computer. Check.
>
> However if, in speech and not writing, I say to you "Hey, time flies like
> an arrow," I sincerely doubt you'll take it as "'time-flies' enjoy a
> particular arrow" or any other bizarre sense. Why? Because intonation
> plays a very important role in communicating meaning...

I'll go even further than that: without intonation of any kind, any
normal English reader would read that sentence as I just did, and parse
it correctly. The failure of a computer to do so does indeed say more
about the limits of the computer than of the language. But there are
other cases where the needed specificity is actually /difficult/ to say
in English without great effort, where it is much simpler in Lojban.
Perhaps that is a small subset; but it is definitely non-empty. And
what is the evidence that intonation /must/ play a part in communicating
meaning? Are we not communicating now? Wouldn't it be nice to use
a language in which we could express the things usually done by
intonation in text instead?

My only argument with Chomsky is that he is treated as a God by much
of the linguistics world. Specifically, his dismissal of the Sapir-
Whorf hypothesis is taken as if it had been discredited, when it fact
nothing of the kind has taken place. Many experiments have failed to
show strong effects, and many /have/ show weaker effects--clearly
showing, for example, that language can affect cognition even to the
point of perception. These phenomena go largely unexplored today
because Chomsky followers don't think they'll amount to anything, and
that is a great loss.

> Phonology is a tricky area; we hear phonological distinctions that aren't
> there in the accoustics. Until we can get a computer to make the same
> context-sensitive distinctions and interpretations that any neurologically
> and anatomically normal 5-year-old can, we don't know enough about speech
> perception and production to try and alter the basics of language.

Exactly my earlier point about the SWH: ask anyone here if they
actually /hear/ word breaks in spoken English. Most people will
unquestioningly say "yes, of course". You and I both know that
those breaks don't exist in the acoustics, so their perception is
being influenced by their learned language. These effects deserve
more study. For example, would a child who grew up with Lojban
phonetics "hear" word breaks where Lojban makes them optional
(which is almost everywhere)? Would two children who grew up
speaking Lojban to each other over email, but with two different
native languages, end up using the same intonations, or might they
even use Lojban text in different ways?

The existence of a language like Lojban makes it possible to
study questions like this in a way that would really not be
possible any other way. Even if the language /as a language/
is deeply flawed--and it may well be, though I haven't found
so yet--it is still a great contribution to the science.

--
Lee Daniel Crocker <lee@piclab.com> <http://www.piclab.com/lcrocker.html>
"All inventions or works of authorship original to me, herein and past,
are placed irrevocably in the public domain, and may be used or modified
for any purpose, without permission, attribution, or notification."--LDC