(This is the version submitted, it may have been
edited before publication)
Every other year the problem of consciousness
brings together neuroscientists, philosophers, psychologists and
assorted weirdoes for a week in Tucson, Arizona. Surrounded by the
giant cacti of the Sonoran desert, this is the place to agonise about
what has been called "The most important problem in the
biological sciences" (Searle 1997, 3), and "the last surviving mystery."
(Dennett 1991, 21).
We should not underestimate the problem. While
cognitive scientists unravel the details of visual processing, and
scan brains at ever higher resolution, the mystery remains.
How can the firing of objective neurons give rise to a
subjective world? How can colour discrimination in the visual system
produce the ineffable blueness of the Arizona sky, or the firing of
neurons become the singing of birds. This is William James's (1890)
"fathomless abyss", now more often called the Hard Problem
or the explanatory gap. Nothing we learn seems to help us cross the
chasm.
When it comes to explaining subjectivity, perhaps
the most common assumption, though rarely explicitly articulated, is
something like this. Our senses are constantly taking in information
and building up a rich representation of the world; that
representation is what we consciously experience at any given moment.
In various fascinating ways this assumption took a beating.
The conference began with a whole morning on
sensory substitution. From Paul Bach-y-Rita (University of Wisconsin),
we learned that with a few hours of practice a congenitally blind
woman could see and manipulate objects using a stimulating array on
her tongue. With no training at all, one who had lost her vestibular
system could stand and walk using feedback from a helmet, and one of
the speakers was a blind woman who has laboriously learned to see
using "soundscapes"; a scanning video to audio system. But
is it really vision? She says it feels like vision, and can pin-point
the time when it began to feel that way. The soundscapes do not
interfere with her other hearing, and she even dreams in soundscapes.
That there is nothing intrinsically visual about
the visual cortex was suggested by the results of Mriganka Sur's
(MIT), experiments in rewiring the cortex of newborn ferrets. Visual
input was re-routed to what should have become auditory cortex, which
was then found to be organised into maps and columns much like visual
cortex. In other words, the characteristic organisation of different
cortical areas may be driven by the input rather than being prewired
or intrinsic. In discussion Alva Noë (UC Santa Cruz) suggested that
even the chemical and membrane differences between auditory and visual
cortex might be an effect of the type of input not a cause.
Noë challenged the usual assumption that any
given neural activity must correlate with the same experience
regardless of how that activity is produced. On the sensori-motor, or
enactive, theory of vision this is not so (O’Regan & Noë,
2002). What makes an experience visual rather than, say, auditory, is
not which neurons are firing, but how the system masters the
sensorimotor contingencies involved. The basic idea is that vision is
action, not representation, and seeing is a skill. This theory
provides a challenging alternative to the usual representation-based,
cognitive view of visual experience. It does away with the dualism
inherent in representational theories and shifts the mystery to a new
one - how action becomes subjectivity.
By contrast many other speakers assumed the more
accepted view. For example Vittorio Gallese (Università di Parma)
presented fascinating data on goal-related and mirror neurons in F5.
In monkeys these neurons only fire when the object of the goal is
visible, or the monkeys know it is present, whereas humans can respond
to goal-directed actions with no seen object. Gallese discussed the
idea of action perception as action simulation which furthers our
understanding of the role of mirror neurons, and even hints at an
enactive view. However, he characterised consciousness as a Phenomenal
Model of the Intentionality Relation (PMIR); a kind of conscious
mental model. This leaves the old problems of how internal
representations comes to be subjective, or why some mental models are
conscious while others are not.
Representational versus action-based theories
reappeared in the session on machine consciousness. Ray Kurzweil
presented an enjoyable series of speculations on the future of human
immortality. According to his "patternism" each of us is a
unique pattern in the brain that might soon be copied and downloaded
into different machines. By contrast, Rodney Brooks explained how AI
has moved away from centralised control and towards distributed
behaviour-based systems that rely on leaving much of the information
that they need right where it is, in the outside world. He showed an
interactive robot, Kismet, in which apparently sophisticated social
behaviours emerge from simple abilities with no central control and no
central representations. In this case a visual attention system
provides the coherence. Could we be systems like this? Could our sense
of conscious unity really arise from a distributed system with no
integrated mental model?
It doesn't feel that way. Or does it? I urged
participants to look harder into their own experience and try really
believing in the results of change blindness experiments and enactive
theories. Doing so can break up the habitual sense that we have
definite "contents of consciousness" and that some
representations are 'in' consciousness while others are 'out'.
Consciousness is not a container, I claimed.
Throwing out some basic assumptions may be
necessary, but some of the theories explored at this conference are
surely doomed to failure. Large-scale quantum coherence in cellular
microtubules might seem to provide unity across the brain but how
would these new quantum effects interact with ordinary, boring,
neuron-size effects, and why does quantum physics explain subjectivity?
Evidence of "presentiment", or emotional responses occurring
before an arousing stimulus is seen, caused much discussion, but are
we really to reject the assumption that causes precede their effects?
In addition to the plenaries, the week included
six parallel sessions every afternoon, poster sessions in the evenings
and two days of pre-conference workshops. We learned about pure
consciousness and the neural correlates of meditation, decision
making, emotion, art, memes, sleep, drugs and time. Since the Tucson
conferences began in 1994, consciousness has become a thriving
multipdisciplinary field. And it looks set to continue that way since
we are still very far from crossing that chasm.
References
Dennett, D.C. (1991) Consciousness Explained.
Boston and London; Little, Brown and Co.
James,W. (1890) The Principles of Psychology,
London; MacMillan
O'Regan, J.K. and Noë, A. (in press) A
sensorimotor account of vision and visual consciousness. Behavioral
and Brain Sciences, 24
Searle, J. (1997) The Mystery of Consciousness.
N.Y. New York Review of Books
