Tackling a Killer Disease
Turtle Mysteries Hatchings
Art of Science - Panspermia
Common Sense Computer
A race beneath the sea is the ultimate test for submarine
designers. Jump aboard for sub sea adventure on Scientific
American Frontiers. Also...will a boy's courage help beat
a killer disease? Can a computer cope with common sense? And
how do endangered sea turtles find their way home? Don't go
away - it's all next on Scientific American Frontiers.
GTE brings you more than the power of telecommunications,
light and precision materials. A grant from GTE also brings
you the power of a new world in Scientific American FRONTIERS.
At GTE, the power is on.
NARRATION Hi, and welcome to the second season of
Scientific American FRONTIERS. I'm Woodie Flowers, your host
for this series. In my day job I'm professor of mechanical
engineering at MIT - that's why I have a special interest
in the rather unusual craft these guys have built What's brought
us from Boston, and these other folks from all over the country,
to Florida's Riviera Beach? Not just the great weather and
the waves. These teams are here to push the envelope of sub
sea technology, with some 35 creative designs for human-powered
submarines. The immediate challenge is an underwater race,
but the long-term goal is new ways to explore the last frontier
on Earth -- the ocean. Checking out new ideas in engineering,
and having fun along the way, is just the kind of thing we
do on FRONTIERS. So here we go -- let's find out who has got
the style and the stamina to win this contest.
Here's the local favorite, Florida Atlantic University, nicknamed
FAU. Remember, this is a contest of submarines which run only
on person power. Up front, one crew member navigates the course.
The propeller is driven by the second crew member, a human
engine who does the pedaling. The well-trained Florida Atlantic
crew, their shark sub, and its conventional propeller drive
set the contest standard. Across the country at California
Poly Tech in San Luis Obispo, the Submarine Club is struggling
to create a whole different kind of propulsion system. Actually,
though this concept has never been used in sub design, it
has been thoroughly field tested ... by the whale. It's an
ambitious scheme. And it's a first-time project for these
young engineers. They're finding out how complicated building
your ideas can be. But with each mistake these eager undergrads
learn something new and keep on improvising.
POLY "You design as you go, that's kind of the motto. And
then you just modify as things don't work or as they work.
That's the whole idea: learn the hard way or the easy way
depending on how good you designed the first time."
At the other end of the spectrum, professional engineers at
the Battelle Institute in Columbus, Ohio are also pursuing
a unique design. But their experience told them that complexity
DEROOS "The simplicity of the device was the real key. We
kind of thought to try and get something to swim like a fish
would be akin to trying to fly with wings on your arms. It
can be done I think but it's a very large research project
In Ohio, this pool is as close as it gets to ocean conditions.
But that doesn't dampen the excitement of the first underwater
test. The sub's body is slender, to cut easily through the
water. But that makes the inside a tight squeeze - so once
the navigator wedges himself in, the peddler has to climb
right on top. On the maiden voyage, you can see Batelle's
idea. It's modeled on how a frog or a human takes to the water.
It seems to work!... The driving will take practice. The frog-kick
design transmits the energy of the human peddler efficiently.
But speed is a problem they're only making one and a half
miles per hour. Their design goal is twice that - and the
extra speed must come from the peddler himself. So the Battelle
team will need some heavy training to make the frog sub fast
"If we can get a stroke a second for ten minutes we'll be
in good shape."
"Right now, we're at a minute, 17 seconds."
Race week dawns on Florida's Atlantic coast. The action starts
here on the beach, where the submarines are launched and then
towed to an off-shore staging vessel. They'll compete in a
tournament - seven days of one-on-one races around this quarter
mile-long racetrack set thirty feet below the waves. Doing
all this underwater is tricky. There are divers to position
the boats. Underwater lights to signal the start. Buoys every
ten feet to mark the course. And safety is a critical concern.
Each sub must tow a line to a topside safety float that can
be tracked at all times. It's an engineering nightmare! But
the challenge has inspired 34 teams from all walks of life:
from high school students working on a shoestring, to large
corporations with deep pockets. The subs come in all shapes
... from the pencil sharp ... to the wide bodied. Most have
a few features in common.., such as a rudder for steering
left and right.., and dive planes for heading up and down.
But the flair of new engineering ideas really comes out in
the propulsion systems -like this flexible tail that both
drives and steers -arms that push without churning up the
water -a six-bladed paddlewheel -twin propellers that rotate
in opposite directions -and the adjustable blades on this
flying bomb. These designs are strong contenders for the innovation
award, but the real excitement is the speed race.
STUDENTS "Alright, F.A.U.! Whooee!"
And in the first round of the tournament, it's Florida Arianfids
shark sub against a Pomona California team. The systems have
been tested a hundred times, but the nervous drivers just
can't resist one final check.
NARRATION Out of the gates Florida Atlantic is off to a blazing
"The second boat's coming by now, running smooth. But they
seem to be pretty far behind the other boat that came by."
And five minutes later, it's an easy win for the shark, even
with some minor engine trouble.
"Beautiful finish. We have a finish!"
THE BIKE "And I pulled my foot out of the pedal so I had one
foot the rest of the way."
"You couldn't get it back in?"
"Well, I didn't want to try because we're really concerned
about drifting into this buoy. So, I always wanted to make
sure that we had control. So I said: 'We can go with one foot?
Another frontrunner - the US Naval Academy. Always competitive,
the midshipmen bring their spit and polish approach to sub
"We're wet sanding it to make the surface as slick as possible,
get all the little nicks and stuff out. After this, we're
going to wax it."
They're up against a sub of an entirely different color.
"What team are you guys with?"
The Naval Academy, with their small, sleek hull should easily
outpace the bulky, and slower, Santa Barbara boat. Navy takes
an early lead.., but then ... a problem. By accident, they
pull their safety float under the surface. That's illegal..,
and it slows them down. The extra drag of the float nearly
evens the odds. And the Naval Academy crosses the finish only
a nose ahead. The midshipmen are thrilled with their narrow
victory. But it's a short-lived celebration - because of the
safety violation, the judges disqualify their sub. Meanwhile,
the Battelle sub is being prepped for its next run. Their
frog-like device has been working flawlessly, and they're
ready for the water! Their opponent: Texas A&M University.
It's late afternoon and the seas are starting to swell. Visibility
is bad. Conditions are getting marginal. Nevertheless, Battelle
gets a graceful start.
"We have a launch!"
Even the judges are impressed.
# 2 "Great!"
But on the first corner...a miscalculation. Unfamiliar with
heavy ocean currents, Battelle drives too close to the course
markers and they snag their safety line. Stuck on the bottom,
the sub and their safety float. The Battelle boat is somewhere
beneath the chop, but no one knows where.
"There's only one thing that would've stopped them."
"I think they caught a buoy."
"Thatís what I think. I think the currents are bad."
NARRATION The disadvantage of being a landlocked team has
become painfully dear. The engineers from Ohio can only wait
for help to arrive.
"Our designs have to be conformed to whatever currents there
are, whatever conditions. We design on ideal conditions but
you have chaos out there."
Back on shore, the undergraduates from Cairo Zy Tech have
gotten their whale tail up and running. But their uncertain
workmanship has yet for be tested in the ocean. Right out
of the gate, the students realize they overlooked something:,
the boat won't stay level in the water. They peddle furiously,
but the whale-tail sub just dives into the sand. Exhausted,
Cal Poly abandons ship. After four days of racing.., the field
is thinning out. driving.., and bad design. There's back luck..,
"She's coming for the top."
TEAM "What's our number? ONE! Whooee!"
One team has something to cheer about - the Benthos Corporation,
from Cape Cod. Their boat has a slender hull and an efficient
propeller. They build submarines for a living and it shows.
"The prop was designed around our peddler. We found out his
horsepower, used a computer program to find out the right
shape for the prop, and then had a CNC machine it out for
us. Other than that we've just kept it simple, we've got a
lot of practice on it. We go for a good clean race and we're
"Okay, I see the first boat coming. It's the Benthos boat.
Holy Smokes - it is pulling! Whooee! The second boat is coming
by the platform now. It's #25. It's looking good. It's got
a ways to go to catch the other one. The other FLEW by."
Round after round, Benthos has beat all comers. They're demonstrating
why real submarines use propellers - its the most efficient
way to move water. So Benthos heads in to the finals.
"One more race, that's all we got, one more race left. If
they can get us off the starting block dean tomorrow we should
have a damn good chance at it."
In the final showdown, Benthos faces the shark sub - and the
students from Florida Atlantic are planning to pull out all
COULSON "Up until now we've kind of held back a bit, made
sure we got around the course, didn't get tangled in any buoys.
Today we've got to take a few chances."
It's a head-to-head contest between two machines driven by
efficient propellers - and two teams with plenty of ocean
"Uggy, uggy, uggy. Au, Au, Au!"
"Ugly, ugly, ugly. Yes, yes, yes. Is that what you meant?"
Into the starting gates go the high-spirited students, next
to the trained professionals. They'll race twice around the
track: a grueling half-mile sprint. Seven days and 60 races
have come down to this match. And right from the start, it's
too close to call.
GUY "Let's go! Alright, Benthos! Ahh ooo!"
ANNOUNCER "They're going into the second turn now, to the
back straightaway. They're still not more than five feet apart,
is Benthos from Fau. They're really moving now. JUDGE "The
Fau boat is across the finish line! The other boat is right
behind them. They're about eight feet apart."
It's an upset win for Florida Atlantic University! Even in
their moment of glory, the students recognize the biggest
payoff is what they've learned about engineering.
THE BIKE "We took this baby from the basic conception to actually
building it, testing it and getting it to run. And that's
what engineering is all about. So basically they're training
us to be engineers. And this is the result."
"Uggy, Uggy, Uggy!"
RESPONSE "Au! Au! Au!"
RESPONSE "Au!" GUY "Uggy, Uggy, Uggy!"
"Au, Au, Au! YEAHHH!
A KILLER DISEASE
NARRATION Imagine we're here, not to play a game,
but on a mission of life and death. This is a drug, or some
surgical procedure, that we've just created in our laboratory.
Using this new treatment to save lives - that's the goal,
way down there. But in the risky and highly regulated world
of medicine, it can take a decade or more to go the full hundred
yards. The first steps, like testing with lab animals, take
us to about here - the thirty yard line. Now we're ready for
a clinical test --giving this treatment for the first time
to just a few human patients. That's the point at which we
join our next story. Meet Justin Cano. He's ten years old
and just bursting with energy. He and his twin brother Jason
have a lot in common. But Justin can't keep up with his brother
these days. He has a degenerative muscle disease called Duchenne
Muscular Dystrophy. As his twin grows stronger and faster,
Justin grows weaker and slower. There is no cure - the harsh
reality that haunts his mother.
"I just thought it was a crippling disease. I didn't know
it was fatal. So that was the real shocker. I think people
can deal with a handicapped child. But not when it just slowly
makes them whither away."
This form of muscular dystrophy is caused by an abnormal gene,
usually passed on by mother. Justin, on the left here, inherited
the gene, but neither his twin Jason nor his older brother
Joey got it. It's not uncommon among children - one of every
3500 male babies is affected. Once a month Justin is monitored
by physicians at the California Pacific Medical Center in
San Francisco. They are trying to determine how far the disease
has progressed, and to stave off the crippling stiffness that
eventually takes over.
DR. MILLER "Can you lie down for me?"
NARRATION Dr. Robert Miller is testing the strength of Justin's
MILLER "Pick it up and hold it. Now don't let me push it down.
Push. Push. Push. Push. Okay, good effort." Hold it there.
Justin's best efforts are pretty feeble -- he can't resist
the doctor's push.
MILLER Here we go... Pull. Pull. Pull... Let's see these belly
muscles here. Let me just check them now.
As his muscles continue to weaken, Justin will no longer be
able to walk. When his heart and lungs give out, he will die.
,MILLER Can you keep it there? Okay. How about in this leg?
Children with muscular dystrophy rarely reach their twenties.
But now, there may be hope - a new treatment that Justin and
Dr. Miller are helping to pioneer.
DR. MILLER Don't let me push it down. Pul1, pull, pull. Okay.
Confronting families with tremendous sadness and grief and
expectations and not being able to do anything, really, to
fundamentally alter the course of the disease is an extremely
difficult task. And a very discouraging one. And suddenly,
to have something that has light at the end of the tunnel,
where we may be onto a treatment that would actually.., potentially..,
mm this around is extremely exciting." Clinical trials of
the new treatment are just beginning. Justin and a handful
of other volunteers will help determine if it's effective
DR. MILLER I can sit in there? I want to just chat with you
about where things stand in terms of our study and what is
coming up. We've had lots of discussions already about how
we're going to be transferring some muscle from you over to
your brother here. And I wonder if, first of all, if anybody
has any questions this morning?
NARRATION The hope is that Jason's healthy muscle cells will
take root in Justin and slow the disease.
MILLER Are you ready for your big procedure today?
JASON Yeah, kind of.
MILLER I don't blame you.
But I'm kind of scared. It is a little scary.
The key idea of the new treatment is to get Justin's muscle
cells working normally again. Under stress and strain, even
a healthy muscle cell will tear. But other cells, called myoblasts,
come to the rescue. They fuse with the muscle cell and produce
dystrophin, a protein that helps the cell repair itself. In
Justin's kind of muscular dystrophy, cells don't produce dystrophin,
so tears can't heal and muscles eventually fail. By transplanting
healthy myoblasts from Jason, Dr. Miller hopes to coax Justin's
cells to produce dystrophin and start repairing themselves.
So Justin's treatment begins with muscle tissue taken from
Jason - a gift of strength from brother to brother. As Dad
distracts him, the surgeon removes a small bit of muscle from
Everything's numb to where you can't feel nothing.
Even my muscle? Oh, good.
You won't feel it.
JASON I know you're using big scissors. I can see them.
Don't look over because that only makes you think about what
comes next. Look at me, okay?
The tissue is bound for a Stanford University laboratory.
It will take six weeks to grow the millions of myoblasts Dr.
Miller needs to put the new treatment to the test.
MILLER I'd be very happy if we could demonstrate that we could
slow the course of this disease and have any impact on the
declining strength in that muscle.
The six weeks pass slowly for Justin
45, 46, 47, 48!
As he waits, he does what he can to hold off the advancing
stiffness. Mom helps through the endless stretching exercises.
98, 99, 100!
At last, the big day has arrived. Dr. Miller will be performing
the trial treatment.
JUSTIN What is that?
It's soap - just brown soap.
He's chosen a small test patch on Justin's shin muscle. Because
myoblasts don't travel very far, [Dr. Miller will have to
inject cells all over the muscle - every spot on this l00-point
grid will receive a shot of one million cells. This is an
experiment - doctors think the treatment works, but they have
to conduct an objective test. So Justin's other leg also gets
100 injections, of an inactive solution. That way no one who
will evaluate the results, not even Dr. Miller, knows which
leg gets the treatment. 200 injections later...
DOCTOR Justin, guess what? We're done.
EVERYBODY Alright! Yeahhh!
Unbelievable! I'm so proud of you. It went so well. We had
no technical trouble and you would've been so proud of Justin.
He leaves the hospital the same day...a little sore, but still
on his feet. The next day is the beginning of a long wait.
It will be six months before doctors know whether the treatment
works. And if Justin's leg does grow stronger, they will still
have to figure out how to treat all the muscles in his body.
But in this new assault on muscular dystrophy, the courage
of Justin and his family is the crucial first step.
If it extends Justin's life one day, ifs worth it. If it can
keep him on his feet longer, it's worth it. I kind of think
it Will. I know that it probably isn't the answer but it will
answer questions that they need to know. And that is so important,
not just for Justin, but all those other kids out there. They're
NARRATION The latest news from Justin is that he
thinks his right leg might be getting stronger. Letís hope
heís right - and let's hope it's the leg that received the
treatment, too. The full results won't be in for several months
- and even if they're encouraging, we'll only be here, at
the forty yard line. Still an awful long way to the goal of
getting this new treatment out into the world. The next step
will be a full-scale clinical trial, with several hundred
patients, for about two years. After that, three more years,
and another thousand patients, to examine side effects. Then,
after a final government review, the new treatment becomes
OF SCIENCE - PANSPERMIA
NARRATION Science often works with pictures - from
many different kinds of microscopes, all manner of different
cameras, and more and more nowadays, from computers. Now many
of these scientific images are not only useful, they can be
extraordinary, just as pictures. It's "The Art of Science,"
if you like, and as a regular feature on FRONTIERS, we're
going to spend a few minutes each episode showing you some
of our favorites. Our first one, a film by Karl Sims, is actually
an illustration of an entire scientific theory. It's computer-generated
- using one of the latest high-speed parallel processing computers
- and it's a spectacular visualization of "panspermia": the
idea that the seeds of life on Earth traveled here, through
space, from somewhere else in the universe. That's by no means
a crazy idea, although it is just a theory. Right now, let's
not worry if it's right or wrong - I suggest we just sit back
and enjoy the show.
After this rough introduction, the ocean will become home
to these baby sea turtles, they'll swim hundreds of miles
out into the open seas. Then, after 40 years, the females
will crawl ashore to nest. And that's the beginning of one
of nature's most intriguing mysteries. Midnight, on Melbourne
Beach, Florida. Using a surveillance night scope, FRONTIERS
witnesses this female turtle emerging from the sea. Scientists
have long suspected that when it's time for turtles to lay
their eggs, they return to the very beach where they were
MEYLAN AND BRIAN BOWEN The turtle's 40 year migration has
made this theory very difficult to test.
NARRATION But that's exactly what Annie Meylan and
Brian Bowen have set out to do on this stormy summer night.
MEYLAN "We should be able to get something before the rain
BOWEN "Yeah, I hope."
They find a flesh set of turtle tracks. Annie and Brian have
to be extremely careful not to frighten the turtle, who's
just settling in to dig her nest.
MEYLAN "Hold my pack."
BOWEN "I'll follow you."
MEYLAN "She's just starting to dig the egg chambers. Probably,
about ten minutes." So we have a few minutes.
After several minutes, the turtle enters an egg-laying trance,
and now TV lights won't bother her. She's laying her eggs
here - but was she also born here 40 years ago? The traditional
way of finding out would be to use an identification tag.
MEYLAN 'There's a part of a tag broken out here. Can we get
And this turtle does have one.
MEYLAN "Do you see any numbers? I feel them, but I can't read
But itís from just two years ago, when she first nested here.
Tagging can't answer the question of where she was born.
MEYLAN "They grow so tremendously in size, from a few grams
to many hundred pounds. It's hard to put a tag on that lasts
that long that you can prove that that turtle is, in fact,
the one that returned some 30 or 50 years later when it reached
sexual maturity. So really an indirect way to look at that
problem is to look at the genetics of the turtles that are
up on the beach."
To look at the genetics, they need to take a blood sample,
easier said than done in a giant turtle.
BOWEN "Itís alright, Mama."
They will be taking blood from turtles at many nesting beaches
around the world. How will their genetic detective work answer
the question of where these turtles were born?
"I'm going to hand you this one."
BOWEN "Got it."
Say these orange turtles hatch on the beach in Florida and
swim away. And these purple turtles hatch off the coast of
Venezuela. They all get together on feeding grounds in the
Caribbean. But what happens when it's time to nest? One possibility
is that the females return randomly to either beach. In this
case, over many generations, orange and purple turtles would
get all mixed up genetically. The other possibility, the long-standing
theory, is that nesting turtles return to the beaches where
they were born. In this case, even over thousands of generations,
Florida and Venezuela turtles would remain genetically distinct.
Back in his lab at the University of Georgia, Brian processes
the turtle blood samples so unlock the genetic clues. He's
taken blood samples from turtles on both the Florida and Venezuela
beaches. So - do turtles return home to lay their eggs? These
drops of genetic material will answer the question. And here's
the unmistakable evidence, in a side-by-side comparison.
On the left, genetic snapshots of nine Florida turtles. They're
all identical. On the right, Venezuela turtles. Also identical.
But here's the crucial discovery: the Venezuela turtles have
a band here that's missing in the Florida turtles, and the
Florida turtles have a band that's missing in Venezuela. The
two populations are genetically distinct.
The only explanation is that the turtles have been coming
home to nest generation after generation.
BOWEN "This result has astounded us. When we started this
study, we thought it very unlikely that animals three inches
long can leave a beach and come back forty, years later to
the same beach. That for that time span they can remember
where they came from and then find their way back."
Now we know that a turtle does make this remarkable journey
home. The next question is: how does she find her way? One
part of the answer is that turtles might be born navigators.
Biologist Michael Salmon believes that turtles may have an
internal compass that they use for navigation. Maybe they
can navigate the way a ship's compass does - responding to
the magnetic field that runs between the Earth's North and
South poles. Salmon's colleague, Ken Lohman, has designed
an experiment to test this theory. The little jacket will
keep the hatchling in place while it's in the pool.
LOHMAN "Oops! I've actually got it upside down.
They take advantage of the fact that the turtle is attracted
to light to start it swimming in one direction.
SALMON "Okay. We found that if they swim in one particular
direction, in this case east, for about an hour that that
turns on an ability to respond to the earth's magnetic field
when the lights are turned out."
Sure enough, the turtle continues swimming east, even without
the light as a cue. Once it's been oriented, it's able to
maintain a direction. But is it really relying on some sort
of internal compass to do this? The best way to find out is
to change the magnetic field and see what happens. That's
exactly the function of these wires surrounding the pool.
They can generate a magnetic field as strong as the Earth's,
but in the opposite direction.
SALMON "If it's responding to the earth's magnetic field,
we now expect it to go west."
NARRATION Soon, the turtle does head west -- a convincing
demonstration that it is responding to the Earth's magnetic
field. The discovery of this internal-compass is a breakthrough
in the understanding of turtle navigation, but alone it does
not explain the turtle's remarkable journey. There's another
mystery -- at journey's end, how does the turtle know she's
really home? What cues does she use? Well, scientists have
shown that turtles have a very keen sense of smell. So maybe
they form a memory, or imprint, of their home beaches based
on smell. If this imprinting theory is true, it could help
save a highly endangered species of sea turtle. Rancho Nuevo,
on the Mexican Gulf Coast, the only nesting beach in the world
for the Kemp's ridley turtle. 45 years ago, tens of thousands
of Kemp's ridley turtles came ashore to nest here every year,
as this rare home movie footage shows. By the late 1970s the
numbers had plummeted. Only a few hundred nesting females
remained. Scientists believed a second nesting beach might
help save the Kempís ridleys, and they used imprinting theory
to create it. First, they had to catch the new eggs before
they ever even touched the sand. This Way, no memory of their
native beach could be formed. The eggs were flown up the coast
to Padre Island National Seashore in Texas. There, the eggs
hatch. As they scramble towards the sea, they are getting
the smell of their adopted home beach. If the theory is right,
the memory of this sand and water should guide the females
back here when it's time to nest. It takes Kemp's ridleys
only 10 years to mature, so the first females should have
returned already. But 12 years later, no turtles have come
back to the Texas beach. Maybe they died as hatchlings, maybe
they will come back next year, or maybe the imprinting theory
is wrong. Now, under laboratory conditions, the theory will
be tested directly by Texas A & M biologist David Owens. These
Kemps ridleys were briefly exposed to the Texas sand and water
when they hatched, and have been in captivity ever since.
OWENS "The experiment itself is to some degree a long shot
because we don't know that we can duplicate nature well enough
in the laboratory to convince the turtles to demonstrate their
Graduate student Heather Kalb is on a mission - the first
step in the attempt to duplicate nature. They moved .through
this sand as they headed for the ocean. And this is the first
ocean water they swam in. Will a few buckets of sand and water
capture enough of this complex environment for the turtles
to recognize their home beach? In the laboratory version of
nature, this is the ocean. For the experiment, the tank will
be divided into four sections. To allow the turtle to swim
around freely, the divisions are created by the flow of water
four strong currents are pumped from the center. Three quarters
of the tank will be filled with artificial sea water. For
the fourth section, Heather takes the sand from the Padre
Island beach and filters the ocean water through it. Heather
makes her way to the middle of the tank to pump the Padre
Island solution into the 4th section. Now, she's ready for
a subject. For this step, timing is everything. In nature,
a female would only look for a nesting beach when she's ready
to lay eggs. This is also the only time she'd be interested
in the Padre Island solution in the tank. So Heather and David
do an ultrasound to check for the presence of follicles, the
first stage of egg development.
Looks real good.
DAVID Ahhhh, there are lots of follicles.
She's slowly gearing up. Every month we're seeing lots of
follicles. So we're almost to the size where eggs should be
secreted. In which case she should be headed to whatever beach
she'd want to be at.
Time for the experiment. This struggling female weighs almost
100 pounds, so it takes a firm hand to get her into the tank.
After twelve years in captivity, will this turtle recognize
the smell of the beach where she hatched? Right from the start
- she heads directly for the Padre Island section of the tank.
And she spends a considerable amount of time hanging out there.
It looks like good news for the imprinting theory. Throughout
the turtle's swim, Heather watches a monitor and keeps close
track of its position. The turtle is sail in the Padre Island
quadrant. But as the experiment continues, she wanders all
around the tank. With this turtle, as with all the others
they've tested, the only dear pattern is that there is no
pattern at all.
I'm a little discouraged with the results. We all look for
good results right off the bat to at least encourage us and
keep us going a bit. But like I said earlier, there are a
lot of doors that are open and I hope to at least narrow down
our state. You know - alright, this is what's been done. Let's
try it. But change some variables. And that's the only way
to get the question answered.
So the verdict on imprinting theory is still out. We know
that nesting turtles come back to the beach where they were
born, and we know that an internal compass helps them get
there - but how they recognize the right spot remains a mystery.
NARRATION I sure hope Heather and her colleagues
keep at it. But they might be aiming too high, maybe trying
to go too far in one step. Here's how I think about it. My
buddy Duke roams around this neighborhood, but he always comes
back here in time for dinner. How does he know which house
has the food? Well, I've set up a little experiment to test
what Duke can recognize. This snapshot of the house is a little
silly. Now here's a swatch of grass from the front yard, a
sock, and some other things he might be able to identify.
Just think of all the things I would have to put together
to replicate Duke's neighborhood. And that's not even counting
things that might never occur to me because cats and I don't
see the world the same way. Sure enough, Duke doesn't seem
much interested in anything, except maybe getting out of here.
It's really hard to design one experiment, for cats or for
turtles, that includes all the factors that could be important.
Maybe Heather will get it in one incredibly clever experiment.
But it seems more likely that the only way to crack the secret
of turtle navigation is to chip away at our ignorance, bit
NARRATION Want to know what happens when we breathe?
Ok, let's ask a seven year-old. A child could answer that.
Well normally, you breathe in air and you breathe out carbon
NARRATION Wow! It's really amazing how much Anthony
has already learned about the world. But, with a little effort,
I could also teach my computer about breathing. And then it
could do some figuring that goes way beyond a child's ability.
For example, let's ask Anthony how many times he breathes,
in a minute and in a day.
In a minute. I'd say probably 30 or 35 and in a day you'd
probably breathe about 100 times or more.
Now a computer would never make a mistake like that. Let's
see - we actually breathe about 40 times a minute - times
60 minutes in an hour - times 24 hours - that's about 57,600
breaths a day. Anthony was a little off on that one. But this
machine would have a lot more trouble answering questions
that it hasn't been specifically programmed to answer. And
that's where Anthony really shines. For example, here's a
question I'm sure he's never thought about before: Can kids
his age be waiters in restaurants?
NARRATION How come?
Because I really think ...Well, if it was an Italian restaurant
I don't think they could hold the tray up like that. Or, if
it was a normal restaurant, that they just serve food in,
I don't think they'd know how to do it as well as professional
NARRATION My computer doesn't know the first thing
about Italian restaurants or professional grownups. I'm not
even going to ask it. But what if I did want it to be smart
in the same way Anthony's smart, in a commonsense kind of
way? That's what our next story is all about. A computer with
the same commonsense approach to the world as people seems
like science fiction, but in fact that's exactly the goal
of this small band of pioneers in Austin, Texas. They predict
that by the turn of the century, no one will even think of
buying a computer that doesn't have commonsense. Think about
the range of subjects a computer would have to know about
to match the commonsense of the average person - everything
from children's stories, to mathematics, to popular films.
The project is named Cyc, short for encyclopedia. But almost
from the start, project leader Doug Lenat realized that this
was a misleading name.
Encyclopedias are almost exactly the complement of what we
want in Cyc. If you look at a large encyclopedia, it will
tell you pages worth of material that you can be pretty sure
the average person doesn't know about one topic or another.
And what we really have to focus on is the knowledge people
share in common about these topics.
ANDBILL This common knowledge is so much a part of us that
it's hard to shake loose. For example, Bill and Keith are
struggling to teach Cyc about feet.
All feet are parts of legs.
One thing we haven't told it though, is that any leg cannot
have more than one foot. You can't have like two feet on one
leg. So they tell Cyc that every leg has only one foot - at
least that's what they thought they were telling it.
So this would say that a foot is at most a part of one leg.
So no one shares feet.
But if a foot is a part of one leg does that imply all legs
have one foot?
No. We need to put in both. That every leg has exactly one
foot. And every foot is part of exactly one leg. It's a frustratingly
slow process which the staff calls "brain surgery". They have
to go into Cyc's brain and alter it every time they want to
teach it something. It's a very different process from the
way a child learns.
When you were little, you didn't know anything. But you had
a way of moving around and learning things on your own. The
computer doesn't have that, I mean, that's our ultimate goal.
Our pie - in - the - sky goal is to get it to do that somehow.
About 3 billion years of evolutionary creativity have gone
into creating devices which learn things - which are designed
to learn things.
Right. And we've been working on computers for about... I
don't know, it depends who you ask.., but we might say 100
years. So we're doing it faster. And give us another 100 years
and we'll have some pretty neat things happening.
NARRATION Actually, some pretty neat things are
already happening. Cyc is not only learning commonsense knowledge,
it's learning commonsense reasoning as well. Here's a simple
demonstration: Say Karen is a musician, and Karen is married
to John. Knowing these two facts, what will Cyc infer? That
John is also a musician. How did Cyc come up with that? Cyc
knows this commonsense rule: spouses tend to have similar
interests. So applying this rule to the two facts it knows
- that Karen is a musician and that Karen is married to John,
Cyc deduces that John is also interested in music. Cyc's not
100% sure, but it knows this is a reasonable conclusion. This
example is simple, but the ability it demonstrates is powerful:
Cyc can figure things out for itself, and learn on its own,
much the way Anthony learns. Here's the next step. It's 8
P.M., the end of a long day. Doug is ready to go home. But
Cyc's work is just beginning. Cyc spends the night deep in
thought, making its own original connections between unrelated
subjects. The method is called analogy, and it's another way
for Cyc to become its own teacher. Doug has high hopes for
this new independent learning. And in the morning he's eager
to find out what profound connections Cyc has come up with.
The profession you have in some ways is similar to what you
order when you go to restaurants. Which is either extremely
deep or just wrong. I think it's just wrong.
NARRATION Independent learning is risky, and Cyc
sometimes makes silly connections. But it can also be an extremely
creative process. Last night, Cyc also came up with this analogy:
the dad of a family.., is like the dictator of a country.
The logic of this connection is impeccable, as Doug explains.
Anthony knows that if you brush well, you won't get a cavity.
So he reasons that if you brush really wel1, you can make
a cavity go away. It's a general model of how things work,
and even though Anthony's not applying it correctly, ifs an
inspiration to Doug.
If Cyc has those same general models, then it'll be able to
fall back on them. It'll be able to give general answers and
not be brittle as well. And yes, it might give the same wrong
answers that Anthony does if it doesn't have the same knowledge.
More importantly, I'll give the right answers, the right responses
to a large range of more general situations - situations which
we can't anticipate right now. The kind of situations that
Anthony and all of us face as we go through life.
NARRATION If Doug Lenat succeeds, and Cyc acquires
the common sense knowledge and reasoning skill that people
have - will we be elated or will we start worrying? How would
you feel if you couldn't tell whether you were talking to
a human being or an intelligent machine? We may get some answers
to those questions in a later FRONTIERS program, when we cover
the first-ever contest in which computers try to fool human
judges, Meanwhile, next time on FRONTIERS, We'll look at flying
robots, restored prairies, and a tree that cures cancer. Please
come back and watch.