-
Shocks and Instabilities:
Climate is like a drunk.
If left alone, it sits.
Forced to move, it staggers.
Coming on stage now is a stunning example of how civilization must
rescue itself. It dwarfs the three big scientific alerts from the
1970s about global warming, ozone loss, and acid rain. But until the
1990s, no one knew much about abrupt climate change, those past
occasions when the whole world flipped out of a warm-and-wet climate
like today’s into the alternate mode, which is like a worldwide
version of the Oklahoma Dust Bowl of the 1930s. There are big
alterations in only 3-5 years. A few centuries later, the drought
climate flips back into worldwide warm-and-wet, even more quickly.
Unlike greenhouse warmings, the big flips have happened every few
thousand years on average, though the most recent one was back
before agriculture in 10,000 B.C. The next flip may arrive sooner
than otherwise, thanks to our current warming trend. The northern
extension of the Gulf Stream appears quite vulnerable to global
warming in four different ways. An early warning might be a decline
in this current. And according to two oceanographic studies
published this last year, this vulnerable ocean current has been
dramatically declining for the last 40-50 years, paralleling our
global warming and rising CO2.
The 2004 version of this talk, for the Adamson Annual Lecture on
International Studies, is
here.
-
The Evolution of Human Minds:
The Ice Age Emergence of Higher Intellectual Functions
The suite of higher intellectual functions includes syntax,
multi-stage planning, structured music, chains of logic, games with
arbitrary rules, and our fondness for discovering hidden patterns
(the search for coherence). It's likely that they share some neural
machinery for handling structure and judging coherence. But the
archeological record suggests that they are late-comers -- that the
three-fold enlargement of the ape brain into the human brain was
complete about 150,000 years ago, but that they were intensely
conservative, doing little that Neanderthals didn’t do as well. The
"behaviorally modern" aspects were seldom seen before the Creative
Explosion about 50,000 years ago. So the big brain is not all about
intellect. What happened to reorganize the brain after 100,000 years
at its present size, to make it more creative and versatile, back
during the middle of the most recent ice age?
The version of this talk for WCBR 2005 is available as webbed
slides with narration.
-
Think Ahead (But How?)
 How do we generate on-the-fly novelty of high quality? Which we do every time we speak a sentence that
we’ve never spoken before. Or a contingent plan for the weekend, an alternative in case it rains.
We are often right the first time, even when novel. So HOW do we weed out the nonsense? Gradually improve our plan? Avoid getting stuck? Recognize a coherent answer, the perfect fit to our current situation?
The Dusseldorf lecture.
-
Cerebral Circuits for Creativity:
Bootstrapping Coherence using a Darwin Machine

The problem with creativity is not in putting together novel
mixtures – a little confusion may suffice – but in managing the
incoherence. Things often don’t hang together properly – as in our
nighttime dreams, full of people, places, and occasions that don’t
fit together very well. What sort of on-the-fly process does it take
to convert such an incoherent mix into a coherent compound, whether
it be an on-target movement program or a novel sentence to speak
aloud? The bootstrapping of new ideas works much like the immune
response or the evolution of a new animal species — except that the
neocortical brain circuitry can turn the Darwinian crank a lot
faster, on the time scale of thought and action. Few proposals
achieve a Perfect Ten when judged against our memories, but we can
subconsciously try out variations, using this Darwin Machine for
copying competitions among cerebral codes. Eventually, as quality
improves, we become conscious of our new invention. It's probably
the source of our fascination with discovering hidden order, with
imagining how things hang together, seen in getting the joke or
doing science.
The version of this for Stanford University is available as
webbed
slides.
-
Planning ballistic movements as an
evolutionary setup for syntax
For slow movements, progress reports can update the plan and correct
an approximate intention. But for ballistic movements that are
over-and-done in 1/8 sec, the feedback is too slow to correct the
movement; you have to make the perfect plan during get set.
We know that our ancestors were eating a lot of meat by about 1.8
million years ago. They had probably figured out how to bring down
big grazing animals, and with regularity. But accurate throwing (as
opposed to, say, the chimp’s fling of a branch) is a difficult task
for the brain. During “get set” one must improvise an
appropriate-to-the-target orchestration of a hundred muscles and
then execute the plan without feedback. While there are hundreds of
ways to throw that would hit a particular target, they are hidden
amidst millions of wrong answers, any one of which would cause
dinner to run away. Planning it right the first time, rather than
trying over and over, has real advantages. Just use the ballistic
movement planning circuits for other similar tasks in the spare
time. And what fits are the novel structured tasks of higher
intellectual function, such as syntax, contingent plans, polyphonic
music, getting the joke, and our search for how things all hang
together (seen in crossword puzzles and in doing science). Yes, some
of them “pay their own way” subsequently, but the free lunch seems
to be alive and well in the brain, where novel secondary uses
abound.
The Seattle University keynote slides are
here.
|