considerable limitations; nevertheless, virtual reality is a useful one. Nature’s
virtual reality is conscious experience—a real-time world-model
that can be viewed as a permanently running online simulation, allowing
organisms to act and interact.
Millions of years ago, nature’s virtual reality achieved what today’s
software engineers still strive for: the phenomenal properties of “presence”
and “full immersion.” From an engineering point of view, the
problems involved in creating successful virtual environments are problems
of advanced interface design. A virtual interface is a system of
transducers, signal processors, hardware, and software. It creates an interactive
medium that conveys information to the user’s senses while
constantly monitoring the user’s behavior and employing it to update
and manipulate the virtual environment.
Conscious experience, too, is an interface, an invisible, perfect internal
medium allowing an organism to interact flexibly with itself. It is a control device. It functions by creating an internal user interface—an “as
if” (that is, virtual) reality. It filters information, has a high bandwidth, is
unambiguous and reliable, and generates a sense of presence. More important,
it also generates a sense of self. The self-model is much like the
mouse pointer on the virtual desktop of your PC—or the little red arrow
on the subway map that advises “You are here.” It places you at the center
of a behavioral space, of your consciously experienced world-model,
your inner virtual reality.
The Ego is a special part of this virtual reality. By generating an internal
image of the organism as a whole, it allows the organism to appropriate
its own hardware. It is evolution’s answer to the need for
explaining one’s inner and outer actions to oneself, predicting one’s behavior,
and monitoring critical system properties. Finally, it allows the
system to depict internally the history of its actions as its own history.
(Autobiographical memory, of course, is one of the most important layers
of the human self-model, enabling us to appropriate our own history,
inside-time and outside-time, the Now and the past.) Consciousness
gives you flexibility, and global control gives you the Ego. On the level of
conscious experience, this process of functionally appropriating one’s
hardware—one’s body—in a holistic fashion is mirrored as the sense of
global ownership, or minimal selfhood.
Nature, it seems, was engaged in advanced interface design long before
we were. It is interesting to note that the best theorists researching virtual
environments today not only employ philosophical notions such as “presence”
or “situatedness” but also talk about the “virtual body.” For them, a
virtual body is part of an extended virtual environment. It is a tool that
functions much like the little red arrow or the mouse pointer. If the virtual
body is employed as an interface, it can even be used to control a robot at
a distance. The related concept of a “slave robot” is particularly interesting.
To achieve such telepresence, there must be a high correlation between
the human operator’s movements and the actions of the slave
robot. (Recall the monkey controlling the robot arm? Now monkeys can
even remote-control the real-time walking patterns of humanoid robots
halfway around the world, from Duke University in America to the Computational
Brain Project of the Japan Science and Technology Agency in Japan, and through a recording of their brain activity only. As Professor
Miguel Nicolelis reports, “The most stunning finding is that when we
stopped the treadmill and the monkey ceased to move its legs, it was able
to sustain the locomotion of the robot for a few minutes—just by thinking—
using only the visual feedback of the robot in Japan.”) Ideally, a human operator would identify his or her own body with
that of the slave robot, achieving this with the help of the virtual body,
which functions as an interface. Again, nature did just that millions of
years ago: Like a virtual body, the phenomenal self-model is an advanced
interface designed to appropriate and control a body. Whereas
in the case of the virtual body, the slave robot may be thousands of miles
away, in the case of the Ego, the target system and the simulating system
are identical: The conscious experience of being a subject arises when a
single organism learns to enslave itself.
The emergence of an Ego Tunnel created a much more efficient way
of controlling one’s body. Controlling one’s body meant controlling
one’s behavior and one’s perceptual machinery. But it also meant directing
one’s thoughts and regulating one’s emotional states. The integrated
conscious self-model is a special region of the high-dimensional user interface
that emerged in our brains. It is a particularly user-friendly interface,
allowing a biological organism to direct its attention to a critical
subset of its own global properties. Having a self-model is like adaptive
user-modeling, except that it is self-directed and taking place internally.
In an important sense, the resulting Ego is a fiction; however, it is also a
wonderfully efficient control device. You could also say that it is an entirely
new window on reality.
I claim that phenomenal first-person experience and the emergence
of a conscious self are complex forms of virtual reality. A virtual reality is
a possible reality. As anyone who has worn a head-mounted display or
played modern video games knows, we can sometimes forget the “as if”
completely—the possible can be experienced as the real. In a way, the
conscious parts of our brains are like the body’s head-mounted display:
They immerse the organisms in a simulated behavioral space.
Together, the embodied brain and the PSM, the phenomenal selfmodel,
work much like a total flight simulator. Before we get to the little word “total,” let’s look at why a flight simulator is a good metaphor for
the way our consciousness works. A flight simulator is, of course, a
training device that helps pilots learn to control an airplane successfully.
To achieve this, the simulation must be as accurate as possible in integrating
two different sources of sensory information: vision and the proprioceptive
sense of balance. During a simulated takeoff, for example,
the pilot not only has to see the runway, but he also has to feel the acceleration
of the “as if” plane—and in relation to his own body.
Advanced flight simulators have replaced the movable cockpit mockup
and computer screen with a head-mounted display; two slightly displaced
monitors create three-dimensional surround graphics. A special
programming technique called infinity optics allows the pilot to look at
remote objects “through the windows” of the cockpit, even though the
computer-generated images are only a few inches from his face. The
mock-up’s movable platform has been replaced with seat shakers that
simulate a range of realistic kinesthetic sensations, such as acceleration
or turbulence. Moreover, so the pilot can learn to use on-board instruments
and get to know how the aircraft will react to different operations,
the simulations of visual and kinesthetic input are constantly
updated at great speed and with maximum accuracy.
The human brain can be compared to a modern flight simulator in
several respects. Like a flight simulator, it constructs and continuously
updates an internal model of external reality by using a continuous
stream of input supplied by the sensory organs and employing past experience
as a filter. It integrates sensory-input channels into a global
model of reality, and it does so in real time. However, there is a difference.
The global model of reality constructed by our brain is updated at
such great speed and with such reliability that we generally do not experience
it as a model. For us, phenomenal reality is not a simulational
space constructed by our brains; in a direct and experientially untranscendable
manner, it is the world we live in. Its virtuality is hidden,
whereas a flight simulator is easily recognized as a flight simulator—its
images always seem artificial. This is so because our brains continuously
supply us with a much better reference model of the world than does the
computer controlling the flight simulator. The images generated by our visual cortex are updated much faster and more accurately than the images
appearing in a head-mounted display. The same is true for our proprioceptive
and kinesthetic perceptions; the movements generated by a
seat shaker can never be as accurate and as rich in detail as our own sensory
perceptions.
Finally, the brain also differs from a flight simulator in that there is no
user, no pilot who controls it. The brain is like a total flight simulator, a
self-modeling airplane that, rather than being flown by a pilot, generates
a complex internal image of itself within its own internal flight simulator.
The image is transparent and thus cannot be recognized as an image
by the system. Operating under the condition of a naive-realistic selfmisunderstanding,
the system interprets the control element in this image
as a nonphysical object: The “pilot” is born into a virtual reality with
no opportunity to discover this fact. The pilot is the Ego. The total flight
simulator generates an Ego Tunnel but is completely lost in it.
If the virtual self functions extremely well, the organism using it is
completely unaware of its “as if” nature. The self-model activated in the
human brain has been optimized over millions of years. The process
that constructs it is fast, reliable, and has a much higher resolution than
any of today’s virtual-reality games. As a result, the virtuality of the phenomenal
self-model tends to be invisible to its user. But strictly speaking,
it is simply the best hypothesis the system has about its own current
state—presented in a new, highly integrated data format. To illustrate
this point, let’s look at a classic experiment in modern neuropsychology.
Nema komentara:
Objavi komentar