Rodney Brooks

DATE March 4, 2002 ACCOUNT NUMBER N/A
TIME 12:00 Noon-1:00 PM AUDIENCE N/A
NETWORK NPR
PROGRAM Fresh Air

Interview: Rodney Brooks of MIT's Artificial Intelligence Lab
discusses the development of robots
TERRY GROSS, host:

This is FRESH AIR. I'm Terry Gross.

Robots are no longer just the stuff of science fiction. My guest, Rodney
Brooks, has devoted his life to building robots; some of them are already
functioning within the real world. Brooks is the director of the MIT
Artificial Intelligence Lab. He's also the chairman of the iRobot
Corporation. He was one of the people profiled in the Errol Morris
documentary, "Fast, Cheap and Out of Control."

In the new book "Flesh and Machines," Brooks writes about how he thinks robots
will change us. It's his hope that by designing intelligent robots, he will
also learn more about human intelligence and perception. A robot, as
described by Brooks, is something that has senses. It senses things in the
environment. It does some form of computation and acts in the world in
response to what it perceives.

I asked Rodney Brooks to describe the most intelligent robot he's designed.

Mr. RODNEY BROOKS (Artificial Intelligence Lab, MIT): Well, in my lab, we
built a robot called Kismet, which is just a neck and head robot--it has no
torso or body--and it has a face and facial expressions. It has eyes which
move around, looking, concentrating on different things, paying attention to
different things, and it can understand people's faces. It can make eye
contact with people, realize when people are making or breaking eye contact,
as we do when we're in conversation with someone face to face, and understands
the prosody in their voice, what emotional content they have, and can do the
pre-linguistic parts of social interaction that we all do pretty well
cross-culturally; knowing whose turn it is to speak, sort of knowing whether
the other person is following what we're saying or not, whether we should
speed up or slow down.

Mostly that sort of thing to us is very subconscious, and we don't even notice
it unless we're talking to someone who doesn't quite obey the rules, and then
they seem a little strange to us. So our robot Kismet is able to do that in a
pretty convincing way, and we've had about 300 naive subjects come in and talk
to the robot, not about anything in particular, because it doesn't understand
the words they say, but interact with it in the social level.

GROSS: So if I was talking to the robot and I was talking very fast, what
would the robot do in reaction?

Mr. BROOKS: Well, it doesn't actually care about fast talking, but it does
care if you get too close to its face, or if you're, you know, are showing it
something and waving it around too fast in its face so it can't track it and
keep focused on it. And in that case it will get a little annoyed, and then
it will express, through its emotional display system, that annoyance. It may
pull back, it may widen its eyes, it may turn its head to the side a little
bit. And most people respond to that appropriately. They realize that
they're doing not quite the right thing in its face, and they tend to pull
back. This is called social amplification, and is a well understood thing
that happens between two people, so the robot pulls back and then the person
pulls back.

GROSS: Yeah, I know a few people who might want to be trained by this
robot...

Mr. BROOKS: Yes.

GROSS: ...about what distance to stand when they're talking to somebody
without invading their space.

Mr. BROOKS: Exactly. Exactly.

GROSS: So what does this robot's face look like?

Mr. BROOKS: Well, you know, you can see the wires, you can see the metal,
but it's got eyeballs, which are covering cameras, and the eyeballs look like
a human eyeball, but the iris is about five centimeters in diameter, so
they're much larger. And as it looks around, those eyeballs turn with the
cameras, so you can tell what it's paying attention to, which is very
important for interacting with someone else, knowing what they're looking at
and what they're paying attention to.

GROSS: Now what are you trying to prove with this particular robot?

Mr. BROOKS: Well, we're looking at infant development, how it is that
children build up understanding of the world and ultimately get to language
acquisition, and much of that is really scaffolded by their mothers. Their
mothers, or primary caregiver, tend to overestimate what the child can do and
what the child understands at any particular point in time, but just a little
bit, so they're always sort of leading the child without really understanding
that they're leading the child.

So when the child, early on, looks at something, the mother may interpret that
to mean the child is interested in that, and then the mother starts paying
attention to that same thing, picks it up and starts interacting with the
child over that. So we're looking at that same sort of scaffolding action and
using that to figure how out a robot, modeling a child, can develop some deep
understanding of the world, the way physical objects and people work in the
world.

GROSS: So do you think at some point you might teach this robot language,
or, you know, design a robot that can be taught language?

Mr. BROOKS: Yeah. Well, actually, I've got a couple of students who are
working on early parts of language acquisition with this robot right now. We
haven't published anything on that yet. We're still working away at the very
early stages of that.

GROSS: Now what is the other most humanoid-type robot you've built, another
one besides Kismet?

Mr. BROOKS: Yeah. We have another humanoid robot called Cog, which in this
case does have a torso, it does have arms. And although we have some of the
same social software on that robot, we're really concentrating on physical
interaction with the world, with the robot picking up objects, looking at
objects, turning them over, as a child does in exploring the world, and
coordinating arm-reaching with their visual system, learning--you know, if
they're looking at some object, how does their arm get to that object, which
children have to learn, and also looking at how the robot can learn to imitate
what a person is doing.

We've discovered--not we, but other researchers over the last few years have
discovered in macaque monkeys there are neurons which fire when the monkey
grasps an object, or when the monkey sees another monkey grasp an object or
even a person grasping an object. So the perceptual neurons for understanding
an action are the same neurons that are somehow used in performing that
action, and that seems to be the basis for us being able to imitate what
someone else does. We're exploring that area right now.

GROSS: Well, you're obviously working on several levels at the same time. I
mean, you're investigating human intelligence and memory and movement. At the
same time you're working on a very mechanical and electronic, you know,
computer level, like how to get machines to make certain movements and do
certain tasks. What would you like to see robots do in terms of functioning
in the lives of humans? You know, like if you could design anything you
wanted to, what kind of functions would you give robots in your life?

Mr. BROOKS: Well, I think we're sort of in 1978 in computer years, in robot
years. We've had robots off in factories hidden away from the general public
just as computers were hidden away from the general public. And then around
1978, they started to leak out into society, and we're just starting to see
that happen now with various robots, lawn-mowing robots, home-cleaning robots.

So I'm interested in how we can build robots that will be useful in our
everyday lives and that people will want to have around. You can always argue
that any extra technology is frivolous--we don't need it--but that's been an
argument we've had for hundreds of years, and I think people will find uses
for robots in their lives.

GROSS: And what kind of uses?

Mr. BROOKS: Well, the early ones that are starting to happen are exactly the
home-cleaning robots, the lawn-mowing robots. We're also starting to see
remote presence robots, where someone at a remote location can project
themselves into a physical robot somewhere else over the Internet and give
supervisory-level commands. It's not like they're, you know, running the
robot with a joystick, because the time lags on the Internet wouldn't allow
that to work very well, but they give high-level commands, like `Go over
there, pick up that object,' and the robot does that. So it's as if they're
there.

People are starting to use those robots for embodied teleconferencing, where
the person is there as a robot. So, you know, in a teleconference and it's
just a person on the screen, you're not really making eye contact with them,
and if you walk out of the room, you know, to go off and look at some other
piece of equipment, the person on the teleconference can't follow. With these
sorts of robots, they can. And...

GROSS: So in other words, the robot--it would transmit information to you
through, like, video camera eyes and microphone ears?

Mr. BROOKS: Yes, it has a video camera and microphones, and that all gets put
onto the Internet. And so the person who's remote just has to have any Web
browser and can control the robot.

Now the Japanese are very interested in this concept, because they have an
aging population, very low birth rate and they don't import cheap labor. So
they're very worried about health care for the elderly, and they're looking at
this method as a way that people in other countries, without being present in
Japan, can do some of the physical labor that will be needed in health care.

GROSS: In other words, a robot would be in your bedroom and a nurse could
monitor from a remote location how you look, how you sound, what you're doing.

Mr. BROOKS: Yeah. And they command the robot to, you know, help you sit up,
put you into a wheelchair, all sorts of physical activities which otherwise a
nurse that would have to be present would do. So it's taking the sweat out of
the labor for that remote person, but it's using their brain in Japan.

GROSS: Would you look forward to being a patient with a robot companion?

Mr. BROOKS: Well, these companions--they really do have a person at the other
end.

GROSS: Right.

Mr. BROOKS: It's a particular--so it's not quite as cold as you might think.
Sort of like you and me right now where we're talking remotely, cross-country.

GROSS: Right. I should explain you're in a studio in Seattle, I'm in a
studio in Philadelphia. And we're hearing each other through a satellite
connection.

Mr. BROOKS: Right.

GROSS: Now the cleaning robots--and that's a common fantasy, to have a robot
that's going to clean your house. How much cleaning can a robot do right now?

Mr. BROOKS: Well, Electrolux just started selling them in Sweden a few
months ago, and you can basically put the robot in a room, press the `on'
button and it will wander around that room and clean it. It uses a random
search pattern. It doesn't do things methodically, but if you let it run for
20 minutes or 25 minutes, it pretty much covers all the space and cleans the
floor pretty well. There's a few things they can't do yet. They can't
distinguish, you know, a dirty sock from a crumpled tissue lying on the floor,
so they're by no means the ultimate that we want in our cleaning robots.

GROSS: If you're just joining us, my guest is Rodney Brooks and he's the
director of the MIT Artificial Intelligence Lab, and he's the author of the
new book "Flesh and Machines: How Robots Will Change Us." Let's take a short
break here and then we'll talk some more. This is FRESH AIR.

(Soundbite of music)

GROSS: My guest is Rodney Brooks, and he's the director of the MIT Artificial
Intelligence Lab. He designs robots. He's also the author of the book "Flesh
and Machines: How Robots Will Change Us."

In your book, you say that you can't watch "2001: A Space Odyssey" without
your eyes tearing up. What is so emotional to you about watching HAL...

Mr. BROOKS: Murder people, yes. When I was in--I grew up in Adelaide, South
Australia, which was fairly remote, technologically, from the rest of the
world. And I was, for some inexplicable reason, from age about eight years
old, I was always trying to build computers out of the most of meager parts.
And then when I saw in my teens "2001: A Space Odyssey," it was such a
revelation to me that other people were thinking about the future with
intelligent beings which were made from silicon and steel not just people,
that it was a real watershed in my life, and convinced me that I wasn't alone
in these sort of crazy thoughts I was having.

GROSS: What are the odds of actually designing a computer like HAL that would
have conversations with people and perhaps kill them?

Mr. BROOKS: Well, let's start with the conversations. You know, today,
speech-understanding systems are getting pretty good, and you can have
conversations in limited domains. You know, there are systems around where
you can make travel arrangements, you can ask for weather reports from all
over the world. We can't do the general change of context as well as HAL
could do it.

HAL also, you know, had all those cameras, those red, glowing cameras all
around the ship, and could track people, understand what people were doing and
even read their lips. And we've got that sort of technology working in our
labs already, including the reading lips; not reading lips just by themselves,
but using that to increase the signal-to-noise ratio coming from a noisy
microphone, like doing correlation of lip motion to the sound. So those
things that appeared totally fanciful back in the late '60s when the movie was
made are getting much closer to reality now.

And, of course, HAL could also play chess and beat anyone at chess, and we've
now got a program that can even beat the world chess champion. So the
components are there. What was missing in HAL as the ability to interact with
the environment directly, and that's an aspect where we've done much better
than HAL could do.

GROSS: Now what HAL had was fear and resentment and, you know, he wants to
kill the astronauts because he's afraid that they're going to turn him off.

Mr. BROOKS: Well, first, he's afraid that they're going to mess up the
mission, because they don't really understand the mission.

GROSS: Right. Right.

Mr. BROOKS: And we have to remember that, you know, in every movie, you need
to make a story. So I don't want to go too far in saying that as we make
robots more intelligent, they're going to want to kill us, necessarily.

GROSS: Well, I mean, the theory here is that, you know, a robot like HAL has
some sense of self-preservation, which...

Mr. BROOKS: Has some sense of self and has some sorts of emotions. Maybe he
had more emotions than the astronauts in the movie, in fact.

GROSS: Mm-hmm. Now you've been able to design a robot that looks like it has
reactions to people talking to it. What do you think the odds are that a
computer would actually have emotions? And I feel funny even asking that
question, because I don't believe in computers having emotions. I believe
that computers can be simulated to look as if they're having emotions.

Mr. BROOKS: No, that's a good distinction, and I think that's a key
distinction. I don't happen to agree with you, but I think that's the key
distinction to make. I think most people today, especially artificial
intelligence researchers, are willing to say that a computer can reason about
facts, can make deductions. But most people are unwilling to say that a robot
can have fear. They'll say that it can simulate having fear and appear to
have fear.

But you know, with the robots that we've built, people interact with them, and
the robots have an emotional model, if you want to call it a model. And naive
subjects that we bring in off the street and interact with the robot, interact
with the robot as though it has emotions, and that dialogue can go on for
minutes at a time. The students who designed the robots and sit with the
robot and work in the same room as the robot sometimes find they have to
switch the robot off because they keep subconsciously reacting to the robot as
though it's an animate being, even though they know that it's just a mere
machine that they built. And they're busy trying to concentrate on something
else, and this robot keeps peering over at them longingly and saying something
in a whiny sort of voice at them, and they feel compelled to respond. So they
switch the robot off so they can concentrate.

But suppose that sort of interaction, which lasts for a few seconds or a few
minutes now, lasted for 10 years? Well, after 10 years, are you still going
to say, `Well, the robot's been simulating having an emotional relationship
with me for 10 years, but it doesn't really have an emotional relationship
with me'? When do you make that transition?

And this, I think, relates to a thing that we as humans have had to deal with
for well over 500 years. Five hundred years ago with Galileo, we had to give
up the notion that God had created the Earth as a special place in the
universe. We found that the Earth was no longer the center of the universe,
and that was very hard for the church to accept at first. A hundred and 50
years ago with Darwin, we realized--or he told us that we and animals were not
really distinct. So our special place in the universe became less special,
because we and animals were the same things that evolved in the same way.

What we've got left for us now is that we're different from machines. We have
emotions. When Garry Kasparov, the world chess champion, was beaten by IBM's
Deep Blue two, three, four years ago, he said something to the effect after he
was beaten, `Well, at least it didn't enjoy beating me,' you know. He still
had that pride that it didn't have emotions like he did. So when we talk
about machines being able to have emotions, it's really threatening to us.
You know they're mere machines. They're cold, hard machines. If we can build
machines with emotions, what's left for us then?

GROSS: Consciousness.

Mr. BROOKS: Ah! Well, then, what if we build a machine that's
conscious--which, by the way, I don't think anyone's got a clue how to do yet.
But you know, I take the position that we are machines, and I think implicitly
all the modern molecular biology ultimately is taking that position, that we
are the product of our biomolecules interacting down at that level, and
they're interacting according to laws of physics and chemistry. So
ultimately, we are a collection of little tiny machines that are interacting
in some lawful way, and out of that emerges us.

So in principle, I don't see any reason that we couldn't build machines with
emotions, and I think to some degree we have already. And I don't see any in
principle argument against building a silicon and steel machine that has
consciousness, ultimately, even though we don't know how to do that yet.

GROSS: Rodney Brooks is the director of MIT's Artificial Intelligence Lab,
and author of the new book "Flesh and Machines: How Robots Will Change Us."
He'll be back in the second half of the show. I'm Terry Gross, and this is
FRESH AIR.

(Soundbite from "2001: A Space Odyssey"; soundbite of breathing)

HAL: Just what do you think you're doing, Dave?

(Soundbite of breathing)

HAL: Dave, I really think I'm entitled to an answer to that question.

(Soundbite of breathing)

HAL: I know I've made some very poor decisions recently, but I can give you
my complete assurance that my work will be back to normal.

(Soundbite of breathing)

HAL: I've still got the greatest enthusiasm and confidence in the mission,
and I want to help you.

(Soundbite of breathing)

HAL: Dave, stop.

(Soundbite of breathing)

HAL: Stop, will you?

(Soundbite of breathing)

HAL: Stop, Dave. Will you stop, Dave?

(Soundbite of breathing)

HAL: Stop, Dave.

(Soundbite of breathing)

HAL: I'm afraid.

(Soundbite of music; breathing sounds)

HAL: I'm afraid, Dave.

(Soundbite of breathing)

HAL: Dave, my mind is going.

(Soundbite of breathing; music)

HAL: I can feel it.

(Soundbite of breathing)

HAL: I can feel it.

(Soundbite of breathing; music)

HAL: My mind is going.

(Soundbite of music; breathing sounds)

(Credits)

GROSS: Coming up, more about robots with the director of MIT's Artificial
Intelligence Lab, Rodney Brooks. And we meet retired General Carl Stiner,
the former commander in chief of the US Special Operations Command. He's
co-authored a book with Tom Clancy about the Special Forces.

(Soundbite of music)

GROSS: This is FRESH AIR. I'm Terry Gross, back with Rodney Brooks, the
director of MIT's Artificial Intelligence Lab and chairman of the iRobot
Corporation. He's devoted his life to building robots. His new book is
called "Flesh and Machines: How Robots Will Change Us." He takes a
behavioral approach to designing robots. I asked him why.

Mr. BROOKS: Well, it seems to me that that is how simple animals work, and so
the large, large parts of even the human brain evolved with that as their main
mechanism. Obviously we can do more than that. We can reflect. We can step
back from situations. So we can do more than that. But it's built on top of
that substrate. And when I started working on robots in the mid-'80s, that
just wasn't around, that lower-level instinctive reaction. Everything was
being done from the high-level reasoning deduction direction, and that seemed
to be fundamentally flawed 'cause it missed out on 98 percent of what
evolution had done in building us.

GROSS: So while some people were trying to give computers a `mind' and
reasoning, you were trying to get them to just do behavior response, you know,
movement and response type of things.

Mr. BROOKS: And have a basic emotional system as we have in the core of our
brain, down in the amygdala, a very ancient part of our brain.

GROSS: You point out in your book that in some ways people are becoming more
like robots in the sense that we have technology we can use similar to the
technology used in creating robots that help us with our senses. For example,
certain implants that can help deaf or nearly deaf people hear. Can you make
that connection for us?

Mr. BROOKS: Yes. There are now tens of thousands, and it may even be
hundreds of thousands of people now but certainly tens of thousands of people
who have cochlear implants. These are put in patients who have--usually who
used to be able to hear and have gone deaf for some reason. And these
cochlear implants, along the length of the cochlea, connect directly to
neurons in the person's brain. These enable people to hear. They can't hear
music very well, but they can certainly hear speech quite well. I was talking
to a gentleman just two days with one of these implants and he'd only had it
for two months and was able to carry on a conversation with me quite easily.

There are also clinical pushes for other sorts of implants. The biggest one
is retinal implants for people with macular degeneration. And there are a lot
of research groups throughout the world experimenting with them. There are
some patients right now who have implants, but the results of how well they're
working have not yet been published. And I expect it's going to be quite a
few more years before they're entirely practical.

And there may be implants--instead of putting a silicon chip in the retina,
another approach is to have a camera that the person wears on their head and
have the inputs go to the rear of the brain to an area called V1, which is the
primary visual cortex. Technically, that might work out to be a little easier
because of some of the difficult parts of coupling to the optical nerve inside
the eyeball.

And then there are many other sorts of implants that are either in use or
under consideration. There are a number of quadriplegic patients who have
damage very high up in their brain stem who aren't even able to use the blow
and suck technology that other quadriplegics use. And some of those
quadriplegics now have connections, neural connections that lets them control
a computer mouse. So they at least have some autonomy in the world and they
can read their own e-mail, patiently with the mouse type out e-mail, browse
the Web, etc., and have a little bit of control over their lives.

So there are all sorts of clinical pushes for getting implants connected to
our nervous systems. And I think as that technology proceeds, people will
start to want to use it to augment themselves. You know, it's not too far
from having a retinal implant that lets someone see who was previously blind
to having a deluxe version of that implant which lets them decide whether they
want to see for normal day vision or whether they want to switch to infrared
at night, which would enable them to do something a normally sighted person
can't do. So I think before too long, people will want to augment their
bodies with these technologies in much the way plastic surgery, which was
originally a reconstructive surgery, has come to be a way for people to
augment their bodies in all sorts of ways.

GROSS: What's the connection between this implant type technology that you're
talking about and the technology that you use in the design of robots?

Mr. BROOKS: Well, fundamentally the sensory systems are the same sort of
sensory systems. And by processing the information, it enables people to do
things they wouldn't be otherwise able to do. At our lab, for instance, we've
been building legged robots for many years, and recently we started building
intelligent prosthetic legs which people can strap on, the amputees, and it
understands how it's being walked on and adapts so that people can walk up and
down stairs as a normal person would instead of the way a person with a
standard knee prosthesis currently has to walk up and down stairs. So the
robotic technology is replacing parts of the human functionality. And as
people start to want to augment their bodies, they'll also have this robotic
technology.

And I think it's not very farfetched to think that 20 years from now, we'll be
able to have some sort of wireless Internet connection implanted in our brain
which then sort of connects us to a whole other realm of information
technology in a way that we haven't previously been directly connected. And I
think that this is going to cause a lot of concern for a lot of people.
There's a professor in England who's just put an implant in his body connected
to his nervous system, and there's a number of ethicists saying this is
immoral to extend yourself this way. So I think people are going to be
concerned about where the boundary between human and non-human is as we put
these technologies in our bodies.

GROSS: Are you worrying about this?

Mr. BROOKS: I'm not worrying in the sense of staying awake at night thinking
this is horrible, but I certainly think there's going to be some issues that
we're going to have to face and consider.

GROSS: Rodney Brooks is my guest. And he's the director of the Artificial
Intelligence Laboratory at MIT. He's also the author of the new book "Flesh
and Machines: How Robots Will Change Us."

So what are you in the middle of doing now?

Mr. BROOKS: Well, recently I've sort of been pushed into thinking about what
it means to be really alive. You expressed a certain skepticism about whether
robots could have real emotions or not.

GROSS: Yes, I did.

Mr. BROOKS: And that has made me think, well, what would it take for a robot
to have real emotions, and reducing that question to its smallest, you know,
ultimate question is: What is the difference between living matter and
non-living matter? If the robot was living, then it could have real emotions.
Now what does it mean for a robot to be living? And then getting down to the
cellular level, what is it that makes a cell alive? And we're now able to
start manipulating biological material down at those lower levels, so we can
sort of come at this question in new ways that we couldn't come at it before.
My hypothesis is that there's something fundamental that we still don't
understand about the nature of even a single living cell. We understand lots
of local things about molecular interactions, but at the system level, we
don't yet have a good way of understanding it.

If I could use an analogy, maybe 500 years ago, trying to understand how the
solar system worked, and just--well, we just had algebra and geometry. We
could describe, ultimately, how the planets moved around the sun, and Kepler
was the one who figured that out, but he didn't have a really deep
understanding of how it worked because if he perturbed the system, he couldn't
predict how it would change. And it wasn't till Newton came along and
invented calculus that mankind got a real mechanistic understanding of how the
solar system actually works. So I suspect that in the biology domain, we're
still back at that algebra-geometry stage, where there's some critical tool
for understanding that we haven't yet developed. We don't even know what it
is we're lacking. And so I've started a research program that's trying to get
at that question.

GROSS: Boy, that's a difficult one.

Mr. BROOKS: Yeah, it is difficult, and it's not a new question. Lots of
people have been worried about this over time, and you know, probably I won't
succeed, but I think we'll uncover some interesting territory along the way.

GROSS: Well, Rodney Brooks, I want to thank you so much for talking with us.

Mr. BROOKS: Thanks very much for having me here.

GROSS: Rodney Brooks is the director of MIT's Artificial Intelligence Lab and
author of the new book "Flesh and Machines: How Robots Will Change Us."

Coming up, General Carl Stiner, the former commander in chief of the US
Special Forces. This is FRESH AIR.

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Interview: General Carl Stiner discusses the history of the US
Special Forces
TERRY GROSS, host:

Americans are becoming more aware of the work of the Special Forces through
the war against terrorism and the movie and book "Black Hawk Down." My guest,
General Carl Stiner, was the second commander in chief of the US Special
Operations Command, responsible for the readiness of all the Special
Operations Forces of the Army, Navy and Air Force, including the Green Berets,
Navy SEALs, Rangers and psychological operations. General Stiner commanded
the Special Forces from 1990 to '93. Before that, he commanded many of the
military's contingency strike forces. He's retired now. His earliest
missions were in Vietnam. Now he's written a new history of the Special
Forces with best-selling writer Tom Clancy and Tony Kolts called "Shadow
Warriors: Inside the Special Forces." It's part of Clancy's "Commander"
series. General Stiner told me that the training for the Special Forces is
highly specialized.

General CARL STINER (Retired): They are trained for certain, we call it,
mission areas of responsibility, such as long-range reconnaissance, search and
rescue, direct action missions, disaster relief, humanitarian assistance,
which could be demining activities like they've been involved in for
Afghanistan long before they were inserted in the war against terrorism, and
many other mission areas. In addition to that, they are trained in their own
specialties. They are culturally oriented and trained so that they know how
to fit into the environment, how to survive in a respected way, be accepted in
a respected way in the environment, and they're also language trained. So
their training lasts from--some of their military occupational specialties up
to three years before they ever become effective as an operational part of an
aid attachment.

GROSS: Now you mentioned in the book that some of the training includes
sensory deprivation. What's that aspect like?

Gen. STINER: You're looking for a person that has the inner strength and
wherewithal to give 110 percent of himself 24 hours a day, every day, without
having to be stroked. So they are placed in an environment where we call
sensory deprivation where they're given requirements to accomplish, but
they're never told how well they're doing.

GROSS: So that they learn to be independent 'cause they're not gonna be
praised in the middle of a mission, they're not gonna get any positive
feedback?

Gen. STINER: No, that's exactly right. Because you're in situations where
you are expected to make the right kinds of decisions, tough decisions,
quickly in the absence of orders.

GROSS: You joined the Special Operations in 1964. What was the focus then?

Gen. STINER: The focus then, up until that point, had been primarily the
mission area of unconventional warfare. That was their main role in World War
II during the jet bird days jumping behind the lines in France and carrying
out ambushes and sabotage operations to disrupt the German forces as much as
possible. And that had carried over. But in 1961, General Yarboro(ph), who
had taken over the Special Warfare Center, had a vision that the challenges
around the world created by Soviet-inspired insurgencies were increasing, and
Vietnam was looming on the horizon, so our mission areas expanded to one of
deep, long-range reconnaissance, counterinsurgency-type advisory and
assistance operations, search and rescue and many others, and that was the
focus of our training at that particular time.

GROSS: Now you were in Lebanon in 1983 when a suicide truck bomb destroyed
the American Marine barracks. What did you do afterwards? What was your job
to do in response?

Gen. STINER: I was there as the chairman of the Joint Chiefs of Staff and
the secretary of Defense representative on the ground to report back daily as
to what was happening, to interface directly with General Tunus(ph), the
Lebanese armed forces, the American ambassador that was there, and as
necessary with the government of Lebanon because we had a military assistance
program going there. And the Lebanese army was being upgraded to be
ethnically representative of the mix of the population. And it was making
great progress and it was an army that respected each other. And the mission
was to bring peace and stability to Lebanon. And I also had an additional
duty as a military adviser to then Don Rumsfeld, who carried the rank of
ambassador and was designated as the president's special envoy for trying to
find a solution to peace in the Middle East.

When the Marines were blown up and the French about five minutes after that,
that morning I was sitting in the office of General Tunus--this was early
Sunday morning--having my daily meeting with him and finding out what his
plans were for the Lebanese army's employment for that particular day. When
that explosion went up, we saw it out the window of his office and we felt the
shock wave. I said, `That's the Marines.' He said, `Lord forbid. Let's go.'
And we ran down and jumped in his car. We didn't take a military vehicle. He
felt it would be quicker to go through West Beirut, which was predominantly
no-man's-land at that particular time, to get there soonest. And as we went
through town, you could see the people that had been rocked out of their beds
by this explosion out on their porches looking in the direction of the
Marines.

We got there, I would say, in seven to 10 minutes and saw what had happened.
And we stayed there for about 10 to 15 minutes while we looked at that and
found the commander. And General Tunus offered him the support of all the
heavy equipment that was in Beirut at that particular time trying to clean up
the rubble from the war that had taken place and was still ongoing. And he
moved a battalion of the Lebanese army in to secure the place while they tried
to recover the dead and the wounded.

And after about 15 minutes there, we went about a mile away to where the
French compound was to link up with them. The disaster there had not been
nearly as bad, because the timing was about five minutes off. As far as
everybody knew, the two bombs were scheduled to explode at the same time, but
the one that hit the French was delayed by about five minutes. And therefore,
they had that much warning time, having heard the US Marines being blown up.
And they were shooting at the truck when it approached their compound.

GROSS: Had you been prepared for the idea of a truck bomb or a car bomb? How
much of that had happened before?

Gen. STINER: The US Embassy was blown up in 1982 in Beirut, and that was the
first use of a bomb that was in a van. There had been some intelligence to
indicate that something significant would happen. In fact, one of the
intercepts indicated that we will perform this spectacular act and make the
ground shake underneath the feet of the infidels. And based on that, all
possible measures were taken to try to determine what this meant. And there
were several natural caves that ran under Beirut, led in from the sea--the
Mediterranean. They were all searched by the Lebanese navy. Seismic
detection devices brought in to see if someone could possibly be digging under
these compounds or in the vicinity of them. And nobody suspected a truck bomb
at that particular time, and we never were able to determine what the method
of attack would be until it actually happened.

GROSS: The Islamic extremist group Hezbollah took responsibility for the
truck bomb that blew up the Marine barracks. What were the difficulties for
you then trying to penetrate an extremist religious group?

Gen. STINER: It was very difficult because if you recall when the embassy
was blown up in '82, the CIA station chief and all of his staff, except for
two, were killed in that. And that sort of neutralized our human intelligence
collection capability for that part of the world. And you cannot penetrate
these religious-based fanatical organizations by overhead systems, referring
to satellites. You have to have HUMINT, human intelligence that's capable of
doing that. So we were comparatively blind as to the threat as compared to
what we had hoped to have in terms of early warning and intelligence and plans
that this organization was trying to bring to bear against American interests
in that part of the world.

GROSS: My guest is General Carl Stiner, the former commander in chief of the
US Special Operations Command. He's written a new book with Tom Clancy called
"Shadow Warriors: Inside the Special Forces." We'll be back after a break.
This is FRESH AIR.

(Soundbite of music)

GROSS: My guest is retired General Carl Stiner. From 1990 to '93 he was the
commander in chief of the US Special Operations Command. He's written a new
book about the Special Forces with Tom Clancy and Tony Colts.

A lot of people are learning about the Special Forces from the book and movie
"Black Hawk Down." I asked General Stiner about the portrayal of the Special
Forces in the film.

You know, a lot of the story of "Black Hawk Down" is about, you know, the
command to leave no man behind. And so we see men risking their lives to
bring out the bodies of Special Forces men who were killed in battle, and we
see men dying while bringing out the bodies. And when I was watching the
movie, I was really wondering about that `leave no man behind approach.'
What's the reason behind it? And do you think it's always a good thing? I
mean, do you think there's times when it's not wise to risk the lives of the
living to bring out the bodies of the dead?

Gen. STINER: Terry, that's our culture. It is very important to the
soldiers involved to know that their bodies will not be left and desecrated on
the battlefield. And therefore, every reasonable effort is made to recover
everybody wounded or dead. But there are times when the commander has to make
the decision as to whether more would be lost in attempting to recover bodies
than to leave them there and go back another time. And that's a commander's
decision; it's left up to him.

GROSS: I suppose I shouldn't ask you to second-guess anybody else's command,
but would you have made the decision to leave no man behind in that mission in
Somalia?

Gen. STINER: I would have done everything possible without sacrificing a lot
of lives to recover those people because you didn't know whether they were
dead or still alive at that particular time.

GROSS: I imagine the fear of having your body desecrated is a pretty serious
fear when you're a soldier, particularly if you're doing these Special
Operations missions. What impact did it have on you when dead American
soldiers were dragged through the streets of Somalia?

Gen. STINER: You know, personally, it was a not shock to me because that's
the culture over there and the environment in which they are in. But it was a
very traumatic experience to Americans that were watching that sort of thing
happen back here and wondering how on earth it could happen, and how we got
ourselves involved in that predicament to begin with.

GROSS: Let me ask you what you think of the movie "Black Hawk Down." Have
you seen it?

Gen. STINER: Yes, I have. It is very authentic, very realistic and that's
the way it was.

GROSS: What lessons do you think the Special Forces Command learned from that
mission?

Gen. STINER: The main lesson is you absolutely do not operate--have to
operate in daylight unless it is necessary for the accomplishment of the
mission. It's much better to operate during darkness. That was one lesson.
I'd say the second lesson is if you're going to go into a situation like that,
you need to have a rescue force standing by with the firepower that can punch
its way in and rescue those that may be entrapped or react to unforeseen,
unanticipated contingencies or clashes that may occur during the operation.

GROSS: You're retired now, I'm wondering if there are emotions that you can
allow back in your life that you can't allow in your life when you're actually
in the Special Forces because you have to be so tough and so invulnerable and
so not full of doubt and not questioning your capabilities? You just have to
keep performing and keep handling it.

Gen. STINER: Yeah, I don't think--my life has not changed one bit, you know.
I am what I am and that's about all I can--the only way I can answer that
question.

GROSS: OK. But one other question.

Gen. STINER: OK.

GROSS: On the back of your new book there's a picture of you and Tom Clancy.

Gen. STINER: Yes.

GROSS: Tom Clancy is wearing dark glasses and a cap that says `Vipers' on it.
He's wearing a leather jacket. You're standing there just wearing khaki pants
and a blue button-down shirt.

Gen. STINER: That's right.

GROSS: I mean, you look like you're on your way to the office, you know, on
casual Friday.

Gen. STINER: Yeah. Right.

GROSS: So it's interesting to me that you're not striking the Special Forces
pose at all. You're not trying to look like the tough guy or wearing the
tough guy clothing...

Gen. STINER: No.

GROSS: Yeah.

Gen. STINER: No, you know, I hadn't even thought about that in the picture,
but, you know, that tough guy image does not prevail. You know, many times
Special Forces personnel while they're on active duty dress appropriately for
the environment in which they operate. It may be a coat and tie. It may be
black if they're on some kind of a special action type mission. Or it may be
fatigues--camouflaged fatigues if they're on a deep reconnaissance missions.
They just dress so they can blend in and survive and accomplish the mission.
And in civilian life, they do not wear clothes that would distinguish them as
Special Operations Forces.

GROSS: Well, General Stiner, thank you so much for talking with us.

Gen. STINER: Terry, thanks a million for giving me this opportunity to be on
your program. You do a great job and thank you.

GROSS: General Carl Stiner is the former commander in chief of the US
Special Operations Command. His new book written with Tom Clancy and Tony
Kolts is called "Shadow Warriors: Inside the Special Forces."

(Soundbite of music)

(Credits)

GROSS: I'm Terry Gross.
Transcripts are created on a rush deadline, and accuracy and availability may vary. This text may not be in its final form and may be updated or revised in the future. Please be aware that the authoritative record of Fresh Air interviews and reviews are the audio recordings of each segment.