Inkwell: Authors and Artists

    <scribbled by mitsu Fri 25 Jun 99 20:29>
  

    
Regarding Washoe, I think that is correct.  They are able to create fairly
complicated compound-modifier constructs (i.e., adjectives, or even clever
inventions of words, like "DRINK FRUIT" for watermelon), but this sort of
nested tree structure I believe eludes them either completely or nearly so.

I agree that the question of a specialized hardwired structure for language
is indeed an orthogonal issue.  However, I tend to think that what we have
in humans is not a detailed heavily specialized structure, but rather just a
slight augmentation or specialization of an area which already exists in
other animals.  For example, apes and chimps have a brain structure which is
similar to our Broca's and Wernicke's areas, but they do not use them for
language (at least not human-style language).  What I tend to think is that
these areas are "almost" good enough to serve the purpose, and the
difference between those areas and the corresponding areas in the human
brain are relatively small.

As an example of what I am getting at, consider an artificial neural network
that was trained to conjugate verbs by being shown examples.  Though the
network was fairly general-purpose and clearly quite different from any
natural network, it turns out it followed the same learning pattern that
human children exhibit; that is to say, it made similar mistakes in the same
order.  What's interesting about this is what it suggests about the
computational structure of the problem: one did not have to have an ultra-
specialized network for it to exhibit the same behavior as a human child,
even though it is obvious that the two systems (child and ANN) are vastly
different.  This suggests that there is something about the problem itself
which happens to have certain characteristics that manifest themselves in a
similar way just by virtue of the nature of the problem itself and the
problem-solving mechanisms (natural and artificial neural networks).  This
of course is not proof of anything; I merely present it as an interesting
case which might be relevant here.  The point is that the mere fact that
human languages tend to have the same deep structure across the planet does
not to my mind prove in the least that this deep structure is somehow hard-
wired in any detailed sense; it just means that the human brain seems to
solve the problem of language generation in a similar way, which could be in
some sense for computational reasons rather than because language is somehow
encoded into our DNA.  That is to say, we have the ability to make these
recursive trees, and that may be just a matter of computational power,
rather than a lot of hard-coded language structure per se.  The deep
structure may simply be a side effect of the architecture of our brains and
the problem of language, juxtaposed; it may well be that slightly more
evolved chimps that could handle nested tree syntax would also generate
similar deep structure to their language as well; in fact, that would be my
expectation.

This is all, of course, merely my personal opinion, my intuition as it were.
 If I were to advance this as a serious hypothesis naturally I would do a
lot more research first.  I am merely spouting my personal feelings on the
subject here.
  

    
(I don't doubt that once the facility for human-style language arose, it
became a strong evolutionary pressure, and thus language facility became
more and more pronounced.  I simply doubt how much detailed structure is in
some sense hard-wired.  I don't doubt that we have some specialized language
facility in *some* sense.)
  

    
The thing is, we know that's what they can do in human-invented sign
language; we don't know what they do in their own.  I certainly wouldn't
want my abilities in English judged by my knowledge of German.

wg
  

    
good point, Wendy
  

    
Hey, Andrew!  Where did the paleolithic graphic go?  Is it coming back?  Did
the Skeptic get you angry email?

wg
  

    
I decided it was putting people off.
  

    
I *liked* it...
and I didn't save a copy.

baaaaaaaaaaaaaah.

wg
  

    
There is a genuine paleolithic picture added to the site, though. It only
seems to show up in IE: look at the background to the translation of the
Harry Martinsson poem.
  

    
Would I be using IE?

I do think the front page needs a little something if you're not going to
bring back the graphic, though.

wg
  

    
Today's Guardian reports that the Kansas Board of Education has voted to ban
the teaching of evolution on the grounds that it is "not a valid scientific
principle."  The article goes on to say that they intend to rewrite the
science curriculum to allow a literal interpretation of Genesis; it will be
permissible to teach that evolution takes place within species but not that
one evolves out of another.

wg
  

    
Yeah.. Heil Right....
  

    
The swing vote was cast by a man who has gronw, according to hius local
paper, a 173.5 lb watermelon. No wonder he doesn't want to belive in the
transmutation of species.
  

    
Holy shit!
  

    
Wonderful the stuff you can find on the Web these days.

wg
  

    
I wonder how much the infighting in the scientific community has played into
this anti-evolution popular sentiment.
  

    
Very little, I would imagine. No one wrote about it (at least among
evolutionists) until I did and I still haven't got an American publisher.
Since creationists only matter in the USA I think they probably still feel
attacked by a monolithic conspiracy (not only scientists, but Hollywood,
pointy-head liberals, etc etc)
  

    
I spent twelve years attending school in a rural Kansas district (#361)
and two years at Kansas State University, and I think I can safely say
without fear of knowledgable equivocation, HOW THE HECK DID THIS HAPPEN?

This isn't representative of the Kansas that I knew.  Granted, it's been
26 years since I moved away, but, at least at that time, Kansas wasn't
anything like that clueless.

I suspect stealth school board candidacies...
  

    
Certainly the swing vote with watermelons (his name is Harold Voth) was
pinned by the newspapers as being a compromise man until the last moment. 

What exactly is a schol board, by the way, and what does it do?
  

    
A school board is a group of people who have been elected by the local
population to oversee the doings of any particular public school
district. Most school boards are unpaid volunteers, though I think some
of the larger districts pay their board members.

The school board generally has a role in determining certain aspects
of the school district's curriculum (taking into account the curriculum
requirements set by the state, too); it can hire and fire the school
district's superintendent (a person who heads the hierarchy of the
whole group of people employed in the school district) and the other
members of the district's cabinet.
  

    
Kind of like a school governor, Andrew,. but elected.  AIUI.

wg
  

    
And, unless those elections are combined with general elections, they're
notoriously poor for turnout, making it a relatively simple matter for an
interest group to stack the board.
  

    
I read in the paper today that the school board is extraordinarily powerful
in Kansas - like a fourth branch of government.  Not sure how that works.
But the legislature would have to introduce an amendment to the state
constitution to reduce the Board's power, and they don't convene again until
January??
  

    
Sounds more like the that'd be the State Board of Education, not the local
school boards for each school district.
  

    
Today's SFChron has an article saying that the governor and some legislators
are putting together a plan for the legislature to call a referendum to
eliminate the state board of education over this mess.  There is no
initiative process in Kansas for amendments to the state constitution.  The
legislature is not scheduled to meet again until January.