The “hard problem” of
consciousness in fact constitutes the central question of the philosophy
of mind: the broader relationship between the body and the mind that
philosophers have been trying to understand since the dawn of time.
This
“mind-body problem” of course takes a different form today
than it did in the time of Descartes. In those days, science
was still in its infancy, and it was still possible for a serious thinker to conceive
of this problem in a non-materialist framework—something
that most scientists today regard as problematic, to say the least.
This major difference affects the very way we approach the mind-body problem
today. Descartes said that he doubted everything except his own thinking and his
conscious experience of having qualia.
Today, the definition and the very existence of qualia are questioned by philosophers
such as Daniel Dennett.
Thus the field of the philosophy of mind
may have narrowed since the time of Descartes, but it is still far from exhausted.
Some authors, such as philosopher
Colin McGinn and linguist Steven
Pinker, think that the difficulties we experience in dealing with
the hard problem of consciousness are simply due to the limitations of the human
brain. This brain is the product
of evolution, and, like the brains of other animals, it has its cognitive
limits. For example, a human brain cannot hold 100 digits in its short-term
memory, cannot visualize a 7-dimensional space, and perhaps also may be unable
to understand how neuronal activity that can be observed from the outside can
give rise to our subjective inner experience. But of course, the authors who hold
this position cannot completely discard the possibility that new ideas might emerge
from the head of a future Darwin or Einstein to shed a completely new light on
this question.
WHAT IS CONSCIOUSNESS?
The idea of consciousness
covers
a variety of phenomena; the categories of consciousness as distinguished by
philosopher Ned Block are described in the first box below.
In
1994, at the first “Toward a Science of Consciousness” conference
in Tucson, Arizona, philosopher David Chalmers proposed that
the problems posed by the study of consciousness could be divided into two distinct
types: the “easy problems” of consciousness and the “hard problem”
of consciousness.
When Chalmers talked
about the “easy problems”, he was of course speaking in relative terms.
These problems are “easy” in the same sense that the problems of curing
cancer and sending a person to Mars are “easy”: they are far from
having been solved, but scientists have a good idea of the steps that they must
still complete to solve them.
In the case of consciousness, the “easy
problems” are to explain certain characteristics of consciousness that seem
solvable by the classic methods of scientific observation and experimentation.
For example, a conscious experience of pain may be attributed to an injury suffered
by the body. Scientists can then investigate further and discover that pain reception
in human beings occurs through one system of “A” nerve fibres, another
system of “C” nerve fibres, and so on.
The same classic methods can be used to
investigate the mechanics of all
of the unconscious processes (vision,
memory,
attention, emotions, etc.) that make consciousness
possible. Thus, when it comes to the “easy problems” of consciousness,
investigators can hope to identify the brain processes underlying them and attempt
to understand why they have evolved. Or to paraphrase Chalmers, we can hope to
find adequate functional explanations for these phenomena.
The
“hard problem” of consciousness arises from discoveries
made in the field of physics in the first half of the 20th century. These discoveries
made it hard to find a place in the world for consciousness. Everything was so
much simpler before, when philosophers and scientists had no trouble in assuming
that the
reality of consciousness was at least as “real” as the reality of
the physical world. But once the world in its entirety came to be understood
as the relationships among forces,
atoms, and molecules, that left very little room for the subjective aspect
of consciousness.
And it is precisely
this aspect that constitutes the heart of what Chalmers calls the “hard
problem” of consciousness, or what his fellow philosopher Joseph
Levine calls the “explanatory gap”. For these two thinkers,
no explanations about the causal role of our states of mind and their instantiation
in a given nervous system (the easy problems) will ever tell us anything about
the subjective dimension of consciousness or, to borrow the language of another
philosopher, Thomas Nagel, about “what
it is like” to be oneself and to experience qualia subjectively.
In other words, any solution to the
hard problem of consciousness must do more than explain, for instance, the processes
that let us distinguish the colour red from the colour green, which is an example
of an “easy problem”. A solution to the hard problem must also explain
how this particular subjective impression of the “redness” of some
object can arise from the activity of our neuronal
assemblies. In its simplest form, the hard problem may be couched as follows:
why does the activity of my brain make me feel something rather than nothing?
Thus the hard problem focuses chiefly on the phenomenological aspect of consciousness,
whereas the easy problems are more concerned with the functional aspects.
Access consciousness
refers to any state of consciousness where, when you are in it, a representation
of its content is immediately available to you. You can then use this representation
as a basis for reasoning and apply it in your rational control of your actions
and speech. The concept of access consciousness recalls that of the global
workspace.
Phenomenal consciousness
corresponds to the qualitative aspects, or qualia,
of our mental lives—in other words, “what it is like” to feel
a pain, perceive a colour, etc.
Reflexive
consciousness (or “monitoring” consciousness) is the ability
to deliberately scrutinize the flow of one’s own thoughts, to engage in
introspection, and to track one’s own behaviours.
Self-consciousness
is the representation of self that imparts a certain unity to one’s mental
life.
The expression “altered states
of consciousness” refers to usually temporary experiences in which people
have the impression that the normal functioning of their consciousness has been
disturbed. For example, after experiencing an altered state of consciousness,
someone might say that they had had the impression of living in a different relationship
with the world, with their identity, or with their body.
Altered
states of consciousness may be associated with events connected to sleep,
such as falling
asleep, sleepwalking,
or being sleep-deprived.
Altered states may be induced by another person (as in hypnosis), or by oneself
(through meditation, prayer, or physical exercise, for example), or by a group
dynamic (as in a collective trance state). Altered states may also be associated
with pathological conditions, such as fever or oxygen deprivation, or with the
use of drugs.
One example of a drug-induced altered
state would be the very well documented effects of marijuana
on the various components of consciousness, such as sensory inputs, clarity of
perception, distortion of time, memory, expectations, functional associations,
and attention.
The pursuit, attainment, and consequences
of altered states of consciousness may be socially acceptable and recognized or
socially disapproved and illegal (as, for example, in the case of the use
of certain drugs).
In general, there seems to be
an optimal register of external stimulation that is needed to maintain a normal
waking state of consciousness. Levels of stimulation above or below this register
seem to result in altered states of consciousness. For example, the extreme boredom
that comes with being deprived of social contact for prolonged periods can induce
altered states. But altered states can also be induced by situations that demand,
on the contrary, sustained alertness or a high degree of concentration (being
intensely absorbed in a task, reading, performing in a sport, etc).
Contrary to what McGinn and Pinker
believe (see sidebar), Gerald
Edelman sees no intrinsic problem in the scientific study of
consciousness. He summarizes his position in the following two sentences: “If
the phenomenal part of conscious experience that constitutes its entailed distinctions
is irreducible, so is the fact that physics has not explained why there is something
rather than nothing. Physics is not hindered by this ontological limit nor should
the scientific understanding of consciousness be hindered by the privacy of phenomenal
experience.”
An even more radical stance than
epiphenomenalism, psychophysical parallelism, postulates that
mind and matter have distinct statuses but evolve in parallel, without exerting
causal influences in either direction. In other words, matter only affects matter,
and mind only affects mind.
The German philosopher
Gottfried Leibniz defended this thesis of “pre-established
harmony” in the 17th century. He believed that God had arranged
things in advance so that minds and bodies would remain in constant harmony with
each other.
Because many of today’s dualists
acknowledge the complete nature of the causality of the physical world (see next
sidebar), they have concluded that mental processes might not exert any causal
influence on the physical world. Even though common sense encourages us to think
that our intentions, desires, and feelings directly affect our behaviours, this
may ultimately prove to be only an illusion, and consciousness might be causally
powerless.
That is why, if these dualists are going
to remain dualists, the concept of psychophysical parallelism is so useful to
them. Because if mind and matter can’t really influence each other, then
we can continue to assume that they go along together like two identical trains
on parallel tracks. Thus mind and matter always remain in phase, so that, for
example, sitting on a tack will always provoke a conscious sensation of pain.
Even for
property dualists, the question of how subjective states can influence matter
without violating the laws of physics remains a thorny one, because the physical
world demonstrates “complete causality”. A soccer
goalie stops an incoming ball because his muscles have contracted, because electrical
stimuli have travelled along certain nerves, because his motor cortex has experienced
a certain activity, because his sensory cortex had done so just before that, because
his retinas had registered the shape of the incoming ball, and so on.
Enaction
proposes an especially original approach to this problem.
One widely shared materialist view
is that it is the particular physiology of the human brain that makes us feel
pain, just as it is the particular physiology of the octopus that must make it
experience pain, and not, as functionalism would have it, some hypothetical structural
similarity.
This suggests that other animals that
do not share the physiology of the human brain would have trouble in experiencing
human feelings. In contrast, functionalism offers the attractive proposition that
such sharing of mental states among species is possible.
But
many materialists dispute this proposition. They say that an octopus’s pain
can very readily be distinguished from a human’s and that these two kinds
of pain can very easily not be the same. For example, many neurobiologists consider
it entirely possible that humans can attenuate their pain (or, indeed, amplify
it) with all sorts of conscious thoughts that arise from their highly developed
cerebral cortex, something that octopi obviously cannot do.
A position similar to functionalism
is to admit that there is something very mysterious about subjective consciousness,
or at least the impression that we have of it, but that it must be possible to
explain this something by means of the physical processes of the material world.
Hence the thinkers who espouse this position - many of whom are physicists or
mathematicians - believe that the “hard problem” of consciousness
can be solved only through the application of some elements of the equally mysterious
realm of quantum
physics, or even of some physical principles that have yet to be
discovered.
As regards quantum physics, it is the
probabilistic, indeterminate nature of this branch of science that, some people
believe, opens a breach through which mental states might have an influence on
the physical world.
To demonstrate
the difficulties of the functionalist approach that allows for the possible existence
of conscious states in machines, philosopher John Searle has
devised a famous thought experiment known as the “Chinese room”.
In this experiment, Searle imagines that a man is
in a room where he receives, through a slot in the wall, a piece of paper bearing
a set of symbols in Chinese, which he does not understand. He then looks these
symbols up in a book that matches them with another set of Chinese symbols, which
he also does not understand, but he writes them down on a piece of paper and sends
it back out of the room.
Now suppose that the first
set of symbols was actually a series of questions in Chinese, and the second set
was the right answers.
According to Searle, this man
is like a computer program that is doing nothing but responding to inputs with
appropriate outputs in a systematic, causal way. In other words, all he is doing
is manipulating symbols according to the instructions written by a programmer
(in the book containing the match-ups), without being able to assign any meaning
to them. But Chinese-speaking people receiving the outputs outside the room might
still be convinced that the man inside understands Chinese, because he would have
passed the Turing
test: he would have succeeded in fooling them.
This
experiment thus shows the limitations of the Turing test and how the appearance
of a conscious state may be mistaken for consciousness itself.
But
not all functionalists have capitulated to the Chinese room argument. Some believe
that what is essential is not to conclude that the man inside the room is not
conscious, but rather that the man/room system as a whole may be. Nobody argues
that a single transistor inside a computer can be conscious. It is rather the
combination of all of the computer’s components that might be conscious,
and not just some of them.
For example, to avoid
the
pitfalls of substance dualism, some philosophers have proposed an approach
called “property dualism”. Property dualism recognizes
that everything consists of matter, but holds that matter can have two types of
properties, physical and mental, and that the latter cannot be reduced to the
former. Property dualism is also often referred to as “non-reductive physicalism”.
According to this approach, pain, for
example, would have a physical property (the action potentials transmitted by
the C nerve fibres that sense pain) and, at the same, a conscious mental property
(the feeling of pain).
Gottfried
Wilhelm Leibniz (1646-1716) (Regarding Leibniz,
see also the sidebar on “pre-established harmony”.)
One
of the arguments for property dualism is the “knowledge argument”.
This argument was first presented by the German philosopher Gottfried Leibniz,
in the 17th century. It was updated by Australian philosopher Frank Jackson in
the 1980s, in the form of a little
fable about anticipation.
Suppose, says Jackson,
that a great neurobiologist has learned everything that she knows about colour
perception from reading books about it, but has never seen a single colour in
her entire life. The day that she sees a red rose for the first tine, she will
learn something new about colour: what
it is like to see the colour red. Which proves, according to Jackson, that
matter has two distinct categories of properties.
For
modern property dualists such as David Chalmers, this option does not constitute
a rejection of science, but rather a call to broaden its horizons, by recognizing
consciousness as a full-fledged entity, just as fundamental as space, time, or
the force of gravity.
But the question
of how subjective states can influence matter without violating the laws of physics
remains unresolved. Not to mention that property dualism leads to panpsychism—the
idea that all matter (even a thermostat, even a rock) has some conscious properties,
however limited they might be.
Some
dualists accept that the closed causality of the physical world must be respected
at all times and that no mental state can influence the brain. For these dualists,
mental states do exist, and are produced by the brain, but add nothing to its
physical functioning. This philosophical position is called epiphenomenalism;
it recognizes causal influences of the brain on the mind, but not the reverse.
(Some philosophers have defended a more radical position, psychophysical
parallelism, according to which there are no causal influences in either
of these directions; see sidebar.)
For
epiphenomenalists, the impression that our intentions, desires, and feelings directly
affect our behaviours is therefore only an illusion, an “epiphenomenon”.
Thus we adults would be somewhat like a small child playing with a toy steering
wheel attached to his booster seat while his parent drives the car. The child
gets so wrapped up in his play that he ends up believing that he’s the one
who’s actually driving. Similarly, we give ourselves the illusion that it
is our minds that are actually guiding our behaviours.
Nowadays, among those who accept the idea of a physical world in which conscious
subjectivity exists, many are also prepared to accept that mental processes might
not exert any causal influence on the physical world after all, even though common
sense encourages us to think that our intentions, desires, and feelings do directly
affect our behaviours.
According to these
thinkers, the subjective feeling of thirst that makes us head for the kitchen
faucet has
nothing to do with our taking this action. Even more surprisingly, if consciousness
has no influence on our behaviour, then it follows that our behaviour would remain
exactly the same even if we were zombies, that is, even if our brain activity
were not accompanied by mental states.
But
it is hard to imagine ourselves as zombies, especially when you consider the verbal
behaviours that we usually interpret as reflecting our mental states. David Chalmers
has tried to show the difficulty of this position by caricaturing himself as a
zombie with no mental states who is busy discussing consciousness with other zombie
philosophers.
Other philosophers, whom American philosopher
Paul Churchland has labelled “interactionist emergent property
dualists”, believe that mental phenomena, even though they
are not physical, can nevertheless play a causal role in the physical
world. Once the brain has generated the subjective sensation of
an odour or a colour, for example, this sensation could influence
the functioning of the brain in return.
This stance,
often referred to simply as emergentism for short,
introduces the idea of a hierarchy of complexity in matter. According
to the emergentists, only those systems that, through the process
of evolution,
have attained a sufficiently complex configuration can cause conscious
phenomena to emerge. Emergentists thus see subjective consciousness
as something that is more than the sum of its basic parts (neurons,
for example), because these parts are physical, while subjective
consciousness seems to be different in kind.
Once again,
these emergent properties cannot be reduced to the basic properties from which
they emerge. In other words, they can be neither predicted nor explained by the
conditions underlying them. The classic example is the way that atoms combine
to form a molecule with different properties. For instance, at room temperature,
both hydrogen and oxygen are gases, but the water that forms from their combined
atoms is a liquid.
Some materialist neurobiologists
regard this concept of emergence and its strange nature as an excuse for not buckling
down to the task of actually studying the neural
correlates of consciousness. Other materialists,
however, invoke the concept of emergence when they offer neurobiological models
of consciousness in which consciousness emerges from the complexity of the proposed
neural processes. But, according to their critics, they thereby leave an “explanatory
gap” that tends to reduce their position to a form of mysterianism.
After Valentine, E.R. (1982) and http://en.wikipedia.org/wiki/Philosophy_of_mind
The
materialist (or physicalist) alternative to
dualism gets around this problem by positing that conscious states may not be
distinct from physical states after all . The effect of our mental states on our
behaviour is therefore no longer problematic, because both are part of the physical
world.
Some materialists opt for a “dual-aspect”
theory that considers the brain and the mind to be the same thing, but seen
from two different perspectives: one external and objective, the other internal
and subjective.
Other materialists, known
as reductionist materialists, tend to simply reduce the mental
to the physical. But they offer at least two different theories about the kind
of identity relationship that exists between mental and physical events. Some
reductionist materialists believe in type-to-type
identity, in which a given type of mental event is considered identical to one,
and only one, type of physical event. This is a theory of identity between two
types of things—mental states on the one hand, and states of the brain on
the other. For people who believe in type-to-type identity, if a flame burns your
finger and makes you feel pain, this type of mental state is identical to the
activation of a brain circuit that signals excessive, harmful heat.
But
many objections have been raised to the idea of type-to-type identity.
For example, some people argue that before you can declare a given mental state
identical to a given physical state of the brain, you would have to know precisely
what types of mental states actually exist, so that you would not identify this
particular mental state with a brain state that is actually identical to some
other mental state. But there is nowhere near any unanimity as to what mental
states actually exist.
On the other hand,
if we assume, for instance, that the love that a father feels for his children
is identical to a certain state of his brain, it still seems far-fetched to claim
that anyone who loves their children must have that exact same brain state. And
what about other species? It does not seem impossible that in mice, for example,
some types of pain may correspond to certain types of neurophysiological processes
that are different from those found in the human brain. That is why the philosopher
Ned Block has described the theory of type-to-type identity as “neuronal
chauvinism”.
In contrast, the other
identity theory embraced by some reductionist materialists, token-to-token
identity, holds that no two people can have the exact same mental state. Consequently,
a given mental state is always unique to a given individual. This unique mental
state will therefore be identified with a neurophysiological state that is equally
unique to the brain of that individual. In short, a token-to-token identity is
a specific, individual case of identity between a unique mental state and an equally
unique brain state.
Thus, whereas a “type”
is a general concept, a “token” refers to a particular occurrence.
In this version of identity theory, every individual can love his or her children
in an entirely personal way, producing a mental state that is unique to that individual
and identical to an equally unique neurophysiological state.
But
is it conceivable that two individuals could have exactly the same mental state
at a specific point in time? If it is, as the critics of token-to-token identity
believe, then there could be some cases where the same mental state is physically
manifested in two different ways in the brains of these two individuals. How can
we then still speak of identity when one thing can be identical to two different
things?
The
answer to this question has come from another theory of the relationship between
the body and the mind, known as functionalism. For functionalists,
what is identical in the two different physical occurrences of the same mental
state is the function—the complete set of cause-and-effect relationships
between the internal mental states.
Functionalism
thus accepts the idea that mental states are internal and hidden, yet does not
go so far as to identify them with qualia,
which are purely subjective experiences. Nor does it adopt the behaviourist
stance, from which the mind is regarded as a “black box” that
should be set aside in order to focus on purely behavioural explanations. Also,
functionalists, even though they believe that mental states are hidden, unobservable
phenomena, nevertheless consider them part of the physical world.
The
classic functionalist analogy to describe this position is the analogy
with a computer’s hardware and software. The hardware consists of the
components that the computer is made of—integrated circuits, connectors,
cables, and so on. The software consists of the various programs that the computer
can run, such as your word processor or your e-mail package. These programs can
often be installed on various types of hardware (PC, Mac, etc.), because the programmers
have made sure to adapt the software’s essence—the causal structure
of its components—to each of these types. What is essential is that when
you type a command, it produces a certain internal state in the machine, and this
state in return produces the right response on the screen. And that is exactly
the way that functionalism envisions the human mind—in terms of causal relations
among internal states.
Consequently,
just as various types of computers can run similar types of software, various
types of brains could have similar conscious states if they have structural modules
that perform the same functions. This is the “multiple realization”
argument, that mental states can be realized in multiple ways, on various platforms,
and hence not necessarily in a human brain.
But
functionalism does not necessarily require that the platform be physical, because
there is no necessary logical connection between functionalism and materialism.
The premise that there are mental states that are causally interrelated does not
necessarily imply that these states are of a material nature. But given the current
state of our knowledge, it seems rather implausible that there are causes and
effects that are not physical. And in fact, the main defenders of functionalism
are materialists.
Another
objection to functionalism targets the hardware/software analogy, according to
which the mind is to the brain as a software program is to a computer. American
philosopher John Searle attacked this analogy with what some philosophers regard
as the most difficult argument to refute: the “Chinese room” argument
(see sidebar).
Another problem that materialist
functionalists create for themselves is that by denying the importance of the
particular substrate of neural physiology, they distance themselves from the particular
strength of materialism, which says that the only conscious states of which we
know, those of human beings,come from the activity of their physical substrate,
that is, the form and activity of their networks of neurons.
Functionalism
thus finds itself in the same position as epiphenomenalism, trying to explain
how mental states can have an influence on the physical world.
All of the
positions discussed so far accept the existence of “mental states”,
meaning desires, beliefs, intentions, etc., at the origin of our behaviours. But
the “eliminative” version of materialism states that
these popular concepts are quite simply false, even if they seem to have real
explanatory power.
The way that eliminative
materialists see things, just as the setting of the Sun is an illusion that can
be explained by the
Earth’s rotation on its axis, so conscious mental states are only an
illusion that will eventually be dispelled by progress in the neurosciences. That
is why this form of materialism is called “eliminative”: it quite
simply eliminates the concept that is causing the problem, i.e., mental states.
For philosophers such as Paul and Patricia
Churchland, two of the main proponents of eliminative materialism, no one is conscious
in the phenomenal sense—the sense of the “hard
problem” of consciousness formulated by Chalmers. Instead, all problems
can be reduced to the “easy” problems that may eventually be solved
without recourse to physical properties other than those that we already know.
In short, philosophers like the Churchlands believe that psychological explanations
of our mental states are only temporary stopgaps that will one day be replaced
by new neurobiological
models.
Faced with arguments such
as those of Jackson and his vision
specialist who has never seen a colour, the eliminative materialists get around
them by saying that one can speak of conscious states in two way: as being conscious,
and as being physical. It is not a matter of two different properties, but rather
of a single property that can be discussed in two different ways. It’s somewhat
like discussing an actor’s role in a particular movie: you can talk about
him using the name of the character he plays, or using his name in real life,
but either way, you are talking about the same person, the same reality.
Another
example is temperature. You learn to think of it in degrees first, but then you
learn that it is the average kinetic energy of the particles concerned—two
ways of conceptualizing the same reality.
Thus
the eliminative materialists would say that Jackson’s vision specialist
who is experiencing the colour red for the first time is basically just experiencing
a new way of talking about the same reality.
Many people may think that the importance
that the theory of enaction accords to behaviour is a step backward to a form
of traditional behaviourism.
But enaction differs from behaviourism in the way that it conceives
of behaviour. Unlike behaviourists, enaction theorists see behaviour not only
as individuals’ accessing their mental processes and knowledge, but also
as knowledge itself, which in turn is defined as any action that is adequate in
the world.
Enaction theory also differs from behaviourism
in the idea of agents that enaction theorists have developed. In enaction theory,
every agent is an entity who learns, in a personalized way. Hence the regularity
of the response to a given stimulus, which is the foundation of behaviourism,
does not apply in this case. The same stimulus may trigger various behaviours,
not only in different individuals, but also in the same individual at different
times, because all agents are constantly learning and changing as the result of
their experiences.
In the early 1970s, Humberto
Maturana and Francisco Varela formulated the concept
of autopoiesis to try to capture the essence of living beings.
Until then, living beings had been defined by various properties or functions
observed in living systems. Maturana and Varela saw this reasoning as somewhat
circular or teleological.
The word “autopoiesis”
comes from the Greek words autos (“self”) and poiein
(“to produce”). It refers to the ability of any living system to maintain
its structure and renew its own components. Hence an autopoietic system is a complex
network of elements that, through their interactions and transformations, constantly
regenerate the network that produced them. In short, a living system continuously
engenders and specifies its own organization.
Source: John Stewart
In
this central idea of autonomy, Maturana and Varela distinguish the system’s
“structure”, which is formed by all of its physical
components, from its “organization”, which consists
of the relationships among these components.
At just
about the same time, Henri
Laborithad made a similar distinction with his concepts of “structural
information” and “circulating information”.
The latter enables each level of organization of a living being to be open to
the next level up, which thus becomes the servomechanism for the underlying regulated
system (follow the Tool module link below).
Henri
Atlan’s writings on self-organization also have been influenced
by cybernetics and systems theory and mesh nicely with the concept of autopoiesis,
because of the emphasis that they place on the idea of autonomy.
The
concept of autopoiesis has many other implications. One is that it relegates definitions
of life based on evolution and reproduction to a secondary role, because as the
advocates of autopoeisis point out, for reproduction to occur, there must first
be self-preservation of the living organism that makes reproduction possible.
Only then could phylogenetic evolution come into play, at the time of reproduction.
Source:
John Stewart
Another implication of autopoiesis
is the impossibility of distinguishing between what comes from the environment
and what comes from the system itself. Communication between a system and its
environment is established by a reciprocal process called “structural coupling”
(see next sidebar) that takes place between their respective elements. And it
is the presence of possible couplings that enables the organism to preserve its
identity. The main implication of this idea for the concept of the “self”
is that the self is necessarily autopoietic—constantly changing, never fixed,
and not resting on any stable base or foundation (see the first box to the right,
on Buddhism).
From the enactive
perspective, people interact with their environment through all of the personal
experiences that they have stored in their memories. This interaction gives rise
to what Varela and Maturana call structural coupling. This concept
places the emphasis not on the optimal adaptation of an organism to various regularities
in the world, but rather on the “viability”of a certain number of
couplings between the organism and its environment.
It
was on the basis of the concept of structural coupling that Varela and Maturana
developed their vision of evolution,
centered on the concept of natural drift.
Natural
drift is distinct from Darwinian natural selection (see the Tool Module link below)
in that natural drift sees the environment chiefly as something that prevents
or prohibits certain couplings between an organism and its environment. If an
organism does not have the structure needed to couple with its environment in
one way or another, then that organism will disappear.
Consequently,
the environment is not regarded as something that dictates organisms’ optimal
structure to them, and in fact, a variety of structures may be able to accommodate
themselves to the constraints imposed by the same environment.
Situated robotics (see the last
box to the right) places most of its emphasis on perception, action, and learning
in a complex environment. In contrast, a related field, artificial life,
is more concerned with using computer simulations to attempt to recreate the conditions
of evolution and reproduction of life itself.
In other
words, if biology is interested in the material bases of life, artificial-life
research deals with the dynamic form of life, without reference to its material
substance. Researchers in this field apply the concept of self-organization, and
they conceive of life as a process
emerging from the interactions of a large number of non-living elements.
For
these researchers, understanding the essence of life thus comes down to understanding
its abstract, dynamic organization—its logic. This makes artificial-life
research a functionalist
approach. It is open to the study of any form of “thought” and thus
differs from the neurobiological
theories of consciousness which focus on understanding the functioning of
thought and of the human brain.
THEORIES OF CONSCIOUSNESS IN THE COGNITIVE
SCIENCES
Around the 1950s and
1960s, behaviourism,
the paradigm that had dominated the experimental study of the human mind since
the start of the 20th century, gradually gave way to the cognitive sciences, which
developed two major theories of consciousness: cognitivism
and connectionism.
Cognitivism equates
thinking with the manipulation of symbols and regards it as an algebra that operates
on representations of the world, just as a digital computer does. Connectionism,
in contrast, equates thinking with the operation of a network of neurons and argues
that every cognitive operation is the result of countless interconnected units
interacting among themselves, with no central control. Connectionism does, however,
retain the idea of representation, which it defines as the correspondence between
an emergent global state and some properties of the world.
Some
critics believe that this idea of representation smacks strongly of dualism
and maintains a separation between the mind and the body, as well as between the
self and the external world. This conception of the human mind has also been criticized
as too passive, reducing it to an input/output device for processing information.
Some authors, such as Ryle, Freeman,
and Núñez, have even argued that the concept of internal representation
was an error of category, or simply a fiction.
These
authors subscribe to a theory of human thought known as embodied cognition.
They are influenced by pragmatists such as John Dewey and phenomenologists
such as Merleau-Ponty, both of whom could conceive of actions
and intentions without representations. The embodied-cognition critique of connectionism
and cognitivism (and hence also of functionalism)
revolves around the idea that the embodied experience of the individual in his
or her environment plays a fundamental role in human thought. In addition to shedding
new light on some of the astonishing cognitive abilities that we display every
day (making
unconscious inferences, coordinating
speech and gestures, understanding
language, etc.), the embodied cognition movement argues
that many of the abilities that let us think about the world and interact with
other people actually originated in the bodily experience of the individuals of
our species.
This way of defining cognition
by intimately linking the body with thought has produced some interesting alternatives
to the epistemological difficulties of the representational models (see first
three sidebars), and in particular to the restrictions imposed by linear causality,
the “subject/object” dichotomy, and “body/mind” dualism.
Among other things, it allows for a rehabilitation
of the role of the emotions in cognition, a reconsideration of our everyday
unconscious thought mechanisms as probably being of capital importance in our
cognitive apparatus, and a recognition of the evolutionary data showing that the
human
brain has evolved in large part to allow the organization of social life.
Embodied cognition is also consistent with the work of thinkers such as Vygotsky
who have emphasized the social and cultural environment as the main driver of
human cognitive development.
In the 1980s, a movement
therefore developed that rejected the cognitivist
and connectionist orthodoxy based on the idea of representation. To understand
the essence of this new movement’s argument, consider the example of a baby
who is learning to walk. No one teaches her any rules for doing so, as the cognitivist
approach would postulate. Yet by trial and error, she learns how to maintain her
balance on her articulated limbs, while avoiding obstacles that would upset that
balance.
The connectionist approach is probably more
appropriate for describing what is happening in this case, but another approach
would be to seek a more complete understanding of this phenomenon by considering
this baby’s particular legs and the environment in which she needs to move
about.
This is the approach that
was proposed by Francisco Varela, the author of many works that
inspired the embodied cognition movement, notably The Embodied Mind: Cognitive
Science and Human Experience (1991). This movement does not deny all the
contributions of cognitivism and connectionism, but does deem them insufficient.
For example, it does not discard the idea of symbol manipulation, but sees it
as a higher-level description of properties that in practical terms are embodied
in an underlying distributed system—the network of neurons. For Varela,
this network can therefore be used to describe cognition adequately, but for such
a network to be able to produce meaning, it must necessarily have a history, and
it must be able to act on its environment and be sensitive to variations in that
environment.
Indeed, in everyday life, what we observe
in practical terms are embodied agents who are placed in situations where they
can act and hence are completely immersed in their particular perspectives. For
Varela, this is the subject on which cognitivism and the emergent properties of
connectionism remain silent: our everyday human experience. This critique of the
two major successive currents in the cognitive sciences places Varela in an epistemological
position known as the theory of autopoiesis (see sidebar). This
theory is accompanied by a particular methodology rooted in Buddhism
(see box below). Together they form what is conventionally called the paradigm
of enaction.
Conceptual
chart of the state of the cognitive sciences in 1991, with the contributing disciplines
and the various approaches (the term “Emergentism” here is equivalent
to connectionism). Source: The Embodied Mind: Cognitive Science and Human
Experience, by Francisco Varela, Evan Thompson, and Eleanor Rosch, Cambridge,
MA: MIT Press, 1991.
From
the perspective of enaction, perception has nothing to do with a static, contemplative
attitude. Instead it consists in a form of action that is guided perceptually,
in the way that the perceiving subject manages to guide his actions in his local
situation of the moment. From this perspective, the surrounding world is shaped
by the organism just as much as the organism is shaped by the surrounding world.
In the language of enaction, the senses of smell and vision
thus become not mere sensory receptors, but ways of enacting meanings.
Thus
the reference point is no longer a world predetermined independently of the perceiving
subject, but rather the subject’s own sensorimotor structure. Consequently,
categorization emerges from our structural coupling with the
environment (see sidebar), and our conceptual understanding of the world is necessarily
shaped by our experience. Cognition therefore is not representation, but intrinsically
depends on the capabilities of our own bodies.
Enaction
also invites researchers in the cognitive sciences to place the greatest stock
in first-person accounts and the irreducibility of experience, while refusing
even the smallest concessions to dualism.
This is also why enaction claims a kinship with phenomenology,
not in its transcendental or highly theoretical sense, but rather in its etymological
sense, meaning that which is manifested in the “first person”, in
embodied thought. This thought must be “mindful and aware”, and hence
non-abstract, and open to the body that makes it possible. This practice of mindfulness/awareness,
Varela tells us, can be found in the Buddhist tradition. That
is why he takes an interest in certain Buddhist practices calling for the gradual
development of the ability to be present in both the mind and the body, both in
meditation and in the experiences of ordinary life (see box below).
The main idea of enaction is that the cognitive faculties develop when a body
interacts in real time with an environment that is just as real. This idea has
had repercussions in several fields of research, in particular situated robotics
(see box below).
Enaction also stimulates debates in
still broader fields, such as consciousness, the nature of the self, and the very
foundations of the world. Because if it is in the very nature of a cognitive system
to function in and thanks to an embodied subjectivity, then the body, far from
being cumbersome excess baggage, becomes both the limitation on all cognition
and the condition that makes all cognition possible. And the body/brain system,
far from being a mere cognitive machine that provides itself with representations
of the world and finds solutions to problems, instead contributes to the joint
evolution of the world and of the individual’s way of thinking about that
world. In addition, that way of thinking is conditioned by the history of the
various actions that this body has performed in this world.
To
paraphrase the 19th century French philosopher Jules Lequier, it is a process
of doing, and by so doing, of creating oneself. This concept of the psyche is
also in line with Swiss psychologist Jean
Piaget’s research on child
development, which tended to show that the child’s psyche is constructed
through its contacts with the environment, which leads the child to become conscious
of itself and of the outside world simultaneously.
Both
for Piaget and for Varela, the outside world is no longer the framework for our
experience, against which the “I” stands out as a distinct entity.
In other words, the relationship between the I and the world is no longer one
of differentiation, but rather one of reciprocal engendering.
The world, though it seems to have been there before thought commenced, actually
is not separate from us: it is our body that enables us to discover a
part of it. It is produced by the history of a structural coupling between
a body and an environment, a coupling that is different for every living system.
This co-determination between cognitive system
and environment thus calls into question the entire representational
aspect of cognitivism and connectionism, which implies a world that
is pre-formed and gradually represented. From the enactive perspective,
it is instead the historical sequence of actions in context that
causes the emergence of a world of meanings—an “enacted
world”, to use Varela’s expression.
Francisco Varela wanted cognitive
science, and in particular the theories of embodied cognition, to have an impact
on people’s daily lives. But the daily practice that he proposed for exploring
the “embodied mind” comes not from Western tradition but from Buddhism:
a practice called “mindfulness/awareness meditation”. In
this practice, to be mindful/aware means that one’s mind is present to daily
experience, that one experiences what one’s mind is doing as it does it—in
short, that one co-ordinates body and mind.
In this
form of meditation, the individual also gradually makes peace with the idea that
there is no specific, stable refuge in experience. This feeling is called “no-self”
or“groundlessness” (sunyata). And in fact, the Buddhist path
deals almost entirely with ways of overcoming emotional attachment to the self.
By relying on a philosophical tradition that is not
abstract but is instead anchored in human experience, Varela seeks to go beyond
the Western demand that we find an objective ground for our existence, at any
expense. He believes that practicing mindfulness/awareness meditation would lead
us to what Buddhist tradition calls the “Middle Way”, avoiding the
two extremes of objectivism and nihilism. The two are intimately linked, according
to Varela, inasmuch as nihilism may be a reaction to the loss of confidence in
objectivism.
Objectivism is the idea that things can
be understood as such, independently of the subject who perceives them. But this
idea has had the ground pulled out out from under it by quantum physicists’
discoveries regarding the undecidable nature of reality. And because it is hard
for Western thought to exist on less than solid ground, the temptation is strong
to look for new ground that cannot be shaken. In this sense, the form of reality-denial
that characterizes nihilism can be seen as a subtle form of objectivism, insofar
as nihilism itself continues to provide a form of grounding.
For Varela, Buddhist mindfulness/awareness meditation, which takes our first-person
experiences in our environment into account, provides the ideal way of returning
to ourselves without having to believe that either subjectivity alone or the “objective
world” constitutes an absolute ground for our existence.
This
tradition also affords the opportunity for a reformulation
of ethics in the absence of such foundations. By confronting our own tendencies
to search for foundations, Varela believes, we end up developing a feeling of
friendship toward ourselves that we can gradually extend to those around us.
This
empathy
would then be based not on any pragmatic moral injunctions, nor any system of
ethical axioms, but rather on our ability to respond to ourselves and to others
as sentient beings who suffer, because they cling to a self that does not actually
exist. Groundlessness would thus prove to be a form of all-encompassing, non-egotistical
concern
for all sentient beings.
The embodied-cognition approach
has also had a significant impact on linguistics.
Starting
in the 1950s, Noam
Chomsky sparked the “cognitive revolution” by showing that the
study of language could help us to understand human cognition as a whole. But
whereas Chomsky
focused on syntax, other linguists, such as George Lakoff,
gradually came to regard metaphor, and hence semantics, as more central to human
language faculties. In 1980, he and co-author Mark Johnson published a book called
Metaphors We Live By in which they set out this theory of conceptual
metaphor in detail and thus founded the field of research now known as
cognitivesemantics.
Metaphors
were long regarded as purely linguistic constructs, but in cognitive semantics,
they are instead regarded as conceptual constructs that have some fundamental
effects on the way we think. For cognitive semanticists such as Lakoff, any conceptual
system that we use for thinking every day is metaphorical, and non-metaphorical
thought is possible only when we are speaking of purely physical entities. And
the more abstract something is, the more levels of metaphor we need to express
it.
Moreover, according to Lakoff, these metaphors
are largely unconscious and hard to detect, because they are often too distant
from their origins or too embedded in our language for us to even notice them.
For example, when you think about it, the most common metaphor for an intellectual
debate is a war: if you were a debater, you might characterize your opponent’s
position as indefensible, then attack it violently until you
had destroyed it, at which point you would hope that the judges would
declare that you had won.
Another good example
of the effect of metaphor on cognition is the widespread use of the expression
“the tax burden” in the mass
media. Now that this metaphor has become so firmly established, regardless
of the intentions of the writer who uses it, it tends to reinforce the idea that
taxes are something that weighs heavily on the backs of taxpayers. The mass media
may thus influence public opinion in favour of privatization—a result that
probably does not displease the financial interests with which these media are
often so closely linked.
Lakoff believes that the
development of thought within societies
has always been influenced by the way that metaphors have been used. He also believes
that the application of one field of knowledge to another has often provided new
perspectives and new understanding precisely because it has generated new metaphors.
Indeed,
the way that metaphors work is by enabling us to understand one domain in the
terms of another. For example, when we say “time flies”, we are using
a concept related to space (the source domain from which we draw the metaphor)
in order to understand time (the target domain that we wish to understand).
The
other major intuition of cognitive semantics is that all human cognition,
including even the most abstract reasoning, is embodied. In other words,
cognition uses and depends on basic bodily phenomena such as the sensorimotor
system and emotions (follow the Tool Module link below).
In short, in Lakoff’s
view, because the human brain is so intimately linked with the human body, its
form, and the way it functions, the metaphors that emerge from this brain must
necessarily be based on this body and its relationship to the world. And for Lakoff,
it is precisely from these metaphors that we form the concepts that let us think
about this world.
Lakoff’s ideas have had significant
impacts on epistemology, including the very principle of disproving hypotheses
that is so central to the scientific method. For Lakoff, the hypotheses that we
construct with the help of complex metaphors cannot be directly disproved. They
can only be rejected following empirical observations that are themselves guided
by other complex metaphors.
George Lakoff is also
known for his progressive political analyses and for the theses that he and Rafael
Núñez have developed regarding the embodied origin of mathematics.
According to Lakoff and Núñez, mathematics, far from being a tool
that predates human nature, in fact constitutes a human conceptual system that
is bound by biological constraints, such as the facts that we have frontal binocular
vision, that we walk in the direction of our field of vision, and that we have
certain specific mechanisms for proprioception.
From the enactive perspective, if
machines are ever to become “intelligent”, they will have to be designed
to operate as autonomous agents in an actual physical environment —in other
words, designed from the bottom up and not from the top down, as in the cognitivistparadigm.
This approach is called situated robotics, a research area first
developed in the 1980s by Rodney Brooks in a major departure
from traditional robotics research based on a cognitivist artificial intelligence
(AI) model.
Brooks applied two basic principles to
his robots. First, rather than satisfying abstract criteria for action, they had
to be immersed in and interact with the real world. Second, they had to be “embodied”,
meaning that they literally had to have physical bodies that let them perceive
the world and act on it, but without ever forming complete representations of
this world or of their actions.
At first, the academic
community had much trouble in accepting Brooks’s approach, because it contradicted
years of effort in traditional AI research. For example, Brooks said that to enable
a robot to move along a wall, there was no need to write a complicated program.
Instead, one could simply equip the robot with a device that detected the wall,
and design the robot so that it always had a slight tendency to advance toward
the wall, but also tended to veer slightly away from the wall once it detected
that it was close. If these two tendencies were balanced correctly, then the behaviour
of moving along the wall would emerge
from the robot naturally.
In fact, proponents of embodied
cognition even argue that all of our human cognitive faculties may have their
source in mechanisms of this same kind.
The metaphor often used to discuss
attention, whether initiated from the bottom up or from the top down, is that
of a searchlight sweeping across the landscape with a beam of
light that may be relatively wide or relatively narrow. But some authors, such
as Christof
Koch, say that it would be more accurate to compare attention to the various
stage lights that do not sweep across everything in their path
but instead illuminate one point, then go out, then light up another point somewhere
else.
Our perception of the world around
us depends first of all on what our senses are capable of detecting (for example,
bees can see ultraviolet light, but human beings cannot). Next, our perception
depends on the attention circuits in our brains, which decide what stimuli to
give priority to, applying either bottom-up or top-down mechanisms (see main text
opposite). And it is only what passes through these two filters that
we can consciously perceive.
We learn a whole lot of things without
realizing it. This is called implicit learning. The following
experiment demonstrates this phenomenon. Suppose that there are six locations
that can light up on a display screen, and six keys on a keyboard that match these
locations spatially. When one of the locations lights up, the subjects’
task is to press the corresponding key. What the subjects don’t know is
that these locations are changing according to a specific set of rules. Without
being consciously aware of these rules, the subjects learn them implicitly, because
their response times improve more than they do when the locations that light up
change randomly.
As this experiment shows, our brains
are constantly trying to structure the world around us, but in doing so they must
work with perceptions that are influenced by all sorts of experiences that we
have had unconsciously or implicitly in the past. That is why, for example, we
tend to see what we have seen before and to act in the same way that we have always
acted. Collectively, this mechanism probably contributes to a people’s “national
character”, their collective
memory, and their common cultural
behaviours.
Munakata’s experiment on
how 3-year-old children learn to follow new rules clearly shows that the quality
of their mental representations may not be completely solid.
In
this experiment, children were given a deck of cards, some bearing pictures of
trucks, others pictures of flowers. In both cases, some of the pictures were red,
while others were blue.
In the first task, the children
were asked to sort the cards by colour, the blues to the left and the reds to
the right. The children performed this task successfully.
Next,
the children were told that they were now going to play the shape game instead
of the colour game, and that the rule had now changed so that their task was to
sort the trucks to the left and the flowers to the right. When the children came
to a card with a red truck on it, they mistakenly sorted it to the right, as they
would have in the colour game.
The crucial point is
that when the children were asked explicitly where they were supposed to place
the trucks, they answered correctly and pointed to the left. But if they were
given a card with a red truck again, they again placed it to the right. In this
experiment, the quality of the representations of the task that the children
had developed was still fragile. They could respond properly to questions
that called on only one dimension at a time, but not to stimuli where colour and
shape conflicted.
Because human beings can express
the contents of their consciousness through language,
there is the temptation to equate consciousness with language.
But in fact, consciousness does not seem to owe its existence to language. For
example, there are many human beings who cannot speak (such as babies, or people
who are mute, or people who suffer from aphasia),
but nobody questions whether these people experience conscious states. Scientists
have also studied the neural correlates of consciousness in many non-human primates.
However, some authors believe that language, and more
particularly “interior language” (or the phonological
loop), may refine or augment a consciousness of self that is already present.
This question is complex, however, and remains the subject of much debate.
Consciousness is constantly passing
from one state of organization to another. Much of its activity can thus be described
as chaotic, from both a phenomenological and an electrical standpoint.
Scientists must therefore use tools of
non-linear mathematics to describe these stable states that alternate
with destabilizations. The most obvious example is the
alternation
between sleep and wakefulness. Another example is the alternation among moments
when our attention is focused, moments when it is shifting, and moments when we
are not paying any attention at all. Because if consciousness is still a big mystery,
so is the absence of consciousness when we are awake but have our heads “in
the clouds”!
The idea that we are
fully conscious of the world around us has been discredited
by a multitude of experimental data. These data indicate that our environment
is far too rich and complex for our nervous systems to process all of the information
from it continuously in real time. Whereas the
classical
model of consciousness proposes that we are conscious of the entire world
around us, in reality we pay attention to only a tiny proportion of our environment,
while ignoring the rest.
Because of our
limited cognitive abilities, evolution appears to have favoured the emergence
of two complementary types of mental processes: attentional processes and unconscious
processes. The phenomena of consciousness, attentional processes, and unconscious
processes have all arisen from the same need: to facilitate
effective action in a complex environment.
First,
let us consider the attentional processes that are so intimately connected with
consciousness. We know that attentional processes existed far back in the evolutionary
past, because they have been detected even in flies. Some scientists even believe
that consciousness itself may be nothing more than an extension of the
attention mechanism associated with working
memory. Michael Posner and Mary Rothbart,
for example, find it reasonable to hypothesize that phylogenetically speaking,
our conscious functions developed from the attention mechanism.
Indeed,
it seems quite plausible to say that when some ancient predator first noticed
a prey animal, then focused all of its attention on it, that predator had just
taken its first step toward conscious thought. And that first step was most likely
followed by a second, in which the predator engaged in some rudimentary reasoning
involving various internal and external stimuli and then either triggered or inhibited
a movement to capture the prey.
One thing
is certain: there are many theories about attention, but all
of them say that we are attentive to something when we select it.
Attention occurs when we focus on a certain stimulus by being more sensitive to
it than we are to others. Our attention may thus be spatial, or based on a property
of an object, or a type of object, etc.
For
example, suppose you are watching TV, and you are also vaguely aware of the sound
of a conversation in the next room, the noise of a fan, and the smell of bread
in the toaster. But then suppose the toast gets stuck in the toaster and starts
to burn. You will unconsciously attribute a meaning
of danger to that smell, then shift your attention to it abruptly.
The
corollary to this selective aspect of attention is that when you pay attention
to one thing, you automatically ignore a lot of others. Our human
attention resources are limited, and it is hard for us to allocate them to more
than one object at a time. If you try to pay attention to two complex tasks simultaneously—for
example, driving in rush hour while negotiating a contract over your cell phone—you
will necessarily neglect one or the other, very likely with negative consequences
in either case.
If you ever read the
written transcript of a conversation, evidence of this same phenomenon will leap
off the page: you’ll be surprised at the number of pauses and “ums”,
because when you’re involved in a conversation yourself, you simply ignore
these things. The same process explains why it’s so hard to find errors
in a text that you’ve written yourself: your attention is so focused on
what you’re trying to convey that you overlook minor misspellings.
Thus,
the only stimuli that reach our consciousness are those that we have selected
as worthy of attention. If anyone is still skeptical that this is the case, a
phenomenon such as inattentional blindness (see box below) will
generally manage to convince them.
Attention
also comes into play when a task or a stimulus requires special processing.
In such cases, attentional processes may be required in order to hold certain
data in memory, to discriminate between two similar stimuli, to anticipate certain
events, to plan a course of action, or to co-ordinate various behavioural responses
so as to achieve an objective.
Neuroscientists
also frequently distinguish two different kinds of attention mechanisms.
The first is initiated from the bottom up, that is, by neuronal
signals from specialized processing modules whose job is to detect and process
stimuli. These signals then trigger a global activation of the neural control
networks.
The archetypical bottom-up
attention behaviour is the orientation
reaction: sensory organs such as the eyes
or the ears detect a new stimulus in the environment, and the entire body then
turns toward this stimulus to learn more about it. This reaction explains why
it is so hard to look away from the TV screens in sports bars or airport waiting
areas: the changing images on the screen are constantly drawing your attention.
When the initial stimulus is visual,
our bottom-up attention responses depend on a network of interconnected areas
of the brain that include the parietal
lobe, the pulvinar
and the superior colliculus.
The
second kind of attention mechanism operates from the topdown,
such as when the neural control networks execute an action that is motivated by
a goal. Our thoughts, motivations, perceptions, and emotions then become available
to consciousness when we pay attention to them.
Top-down
attention mechanisms are what let you read in noisy surroundings.
Top-down
attention mechanisms come into play, for example, when you consciously decide
not to pay attention to something that is very attractive (such as the TV screen
in the airport or the people around you in a club) so that you can concentrate
on something else (such as the book you’re reading). Top-down attention
mechanisms are also what enables you to continue pretending to read while actually
concentrating your attention on an interesting conversation going on nearby.
This ability for attention to be an active
phenomenon has also been well demonstrated in experiments where the subjects were
required to press a button as soon as a visual target lit up. These experiments
showed that the subjects’ reaction time was faster when they were shown
an indication of where the target was going to light up shortly before it did
so. And when they were shown an intentionally misleading indication, their response
time was slower than when they were given no indication at all. These results
clearly show that our expectations influence our perceptions, or, in other words,
that attention can be a mental process directed from the top down.
In
short, we live in a complex environment where stimuli from all directions make
claims on our attention, but our attention focuses only on those stimuli that
are new or already have some meaning for us. We may think that we are conscious
of the entire scene around us, but that is indeed an illusion, as phenomena like
change
blindness clearly demonstrate.
It
is therefore a trap to accept the standard, naïve, realist view of the world
around us as a place where “objective reality” is composed of objects
whose characteristics are independent of our senses. The projective nature
of perception means that the content of our consciousness—each
individual’s psychic reality—is greatly influenced both by biological
predispositions and by learning,
both of which are projections of the distant or recent past onto the present.
A visual scene that
you have not yet explored is equivalent to chaos, or, in more modern language,
to “latent information”. It is your brain that immediately
projects a meaning onto the scene, establishing order within it. And
the processes by which your brain decides which lines and surfaces go together
to form particular objects, such as the cubes of cheese in the photo to the right,
are entirely unconscious. In other words, the attentional processes
that lead to consciousness are applied not to a reality that exists completely
outside of us, but rather on the basis of our preconceptions and remembered forms
that we project onto our environment. That is why all of our perceptions can be
regarded as constructed—a fact that is the basis for many optical
illusions .
When
you look at a complex image like this one, you don’t see an undifferentiated
mass of disjointed lines and surfaces, but rather a pile of separate, three-dimensional
cubes of cheese.
We
therefore cannot speak of consciousness and attentional processes without also
considering the extensive unconscious
activity that is going on in our brains at all times. On
the one hand, attention plays a fundamental role in learning by amplifying the
important representations that enable us to take appropriate actions at any given
time. But on the other hand, by thus increasing the quality of
certain mental contents, learning increases the likelihood that a given piece
of mental content will be at the centre of the attentional processes of our subjective
consciousness sometime in future. That is why many people believe that the true
relationship between the self and the world is one of reciprocal
engendering.
Many experimental findings
support this description of consciousness in terms of quality of representation
(see sidebar concerning Munakata’s experiment). From this standpoint, the
quality of representations is therefore a continuous variable, a continuum
that allows a gradual transition from the unconscious to the conscious. This quality
gradient between unconscious and conscious, which is shaped by
learning, also suggests that the corresponding representations depend on the same
underlying processes, rather than on two distinct neural systems.
But
although learning, by increasing the quality of a stimulus, can facilitate its
access to consciousness, recurrent consciousness of a stimulus can paradoxically
return it to the realm of the unconscious, through habit or automaticity.
We can therefore describe consciousness (or the realm of the explicit) as
a peak between two domains of lesser consciousness in which the two extremes
would be two very different types of unconscious.
In the first type of unconscious, the quality of the representations is too
low to enter consciousness. Like fish that are too small and
get thrown back into the water, these stimuli do not reach consciousness.
They remain unconscious, but this does not mean
that they cannot affect our behaviour.
In the second type of unconscious,
the quality is so high that it enables the representations to
be expressed all on their own, unconsciously, because they are so well
engrained
in our memory.
This transfer
from the conscious to the unconscious clearly comes into play in the learning
of motor skills, such as how to ride a bicycle, or ice skate, or tie your shoelaces.
In the beginning, everything is conscious and laborious, but as you practice,
it all becomes automatic and unconscious. In a process of automatization
such as this, our conscious experience is diminished as we gain in expertise.
But the opposite can also happen, when
we try to master the fine distinctions that characterize a particular field of
knowledge, such as oenology or philosophy. In this case we are no longer engaged
in procedural learning, but rather in explicitly expanding our knowledge in a
particular field. And the more we learn about the complexity of this field,
the more conscious we become of new distinctions that make it even more rewarding.
We can therefore distinguish three different
possible situations involving two different characteristics of knowledge: its
availability to our consciousness and our ability to
control our conscious representation of it. When both availability and
controllability are low, this knowledge is implicit (unconscious). When both availability
and controllability are high, this knowledge is explicit (conscious). And when
availability to consciousness remains high but controllability drops back to a
very low level (as the result of habit or automaticity), then the knowledge is
said to be automatic.
If
you aren’t expecting to see something at a given time, you may not see it
at all, even if it’s something big.
This curious
phenomenon is called inattentional blindness, and there’s
no better way to discover it than to experiment on yourself. To do so, click on
the link at the end of this paragraph. It will take you to a video that shows
two basketball teams. The three players on one team are wearing black, and those
on the second team are wearing white. You must count the number of passes made
by the players in white. This task is made harder by the fact that the players
in black are passing a second ball around in the same space. Go ahead, try the
experiment before you read on.
Now
that you’ve watched the video, here are not one but two questions: how many
passes did the players in white make, and what was the gorilla doing?
What
was the gorilla doing? In actual experiments, this second question leaves more
than half of the subjects at a loss, because they never saw any gorilla at all
while they were watching the video. And yet, when they watch it again, this time
without their attention being directed toward the white team, they see something
really strange: someone dressed in a gorilla suit walks into the space where the
players are passing basketballs around, stops, turns toward the camera, thumps
his chest, then calmly walks out the other side of the frame! (If you didn’t
see this the first time, run the video again, and see whether you do now.)
How
come only about half of the subjects in the actual experiments noted the incongruous
presence of the fake ape? Though the explanation is relatively simple (the subjects’
attention was focused elsewhere), the phenomenon of inattentional blindness is
still surprising, to say the least, and raises serious questions about the
idea that we have a total consciousness of the world around us. On the contrary,
this experiment shows that the world is too complex for us always to have a detailed
awareness of it, and that numerous attentional and unconscious
processes are at work in our brains.
There is not “a” consciousness,
but rather many levels of consciousness, a continuum of intermediate states. But
this continuum also contains a dichotomy: because of
the sequential nature of consciousness, a given representation at a given
time either is conscious or is not. Just as the gradual and continuous changes
in the temperature of a mass of water are accompanied by a sudden change in state
at 0°C (when the water becomes solid) and at 100°C (when the water becomes
steam), the degrees of unconsciousness may thus vary up to an activation
threshold past which the representation suddenly enters into the realm
of consciousness.
Consciousness may thus be similar
to what physicists call a “phase transition”: a sudden
transformation that occurs on a large scale following numerous microscopic changes.
One example of a phase transition is the emergence of superconductivity in certain
metals when they are cooled to a certain critical temperature. It is not surprising,
then, that the concept of emergence
is frequently invoked to describe the appearance of consciousness.
Ludwig Wittgenstein
believed that philosophical problems
needed to be addressed not so much with solutions as with therapies that revealed
the underlying confusion from which they sprang. As he described the purpose of
philosophy, “We must show the fly the way out of the fly-bottle.”
That’s why it’s so important to keep alert to other points of view,
just in case the problem lies with your formulation, and not with the actual reality
that you are trying to understand— probably a good idea to keep in mind
if you are studying consciousness!
SOME PROMISING CONCEPTS AND MODELS FROM
THE NEUROSCIENCES
The development of the
cognitive
neurosciences in the late 20th century enabled the first empirical models
of human consciousness to be developed. Subsequently, many researchers have refused
to accept the existence of a “hard problem” of consciousness that
would preordain the failure of any attempt to model consciousness on the basis
of neurobiological data. For these researchers, this “hard problem”,
which refers to the subjective
aspect of consciousness, comprises a multitude of more concrete problems,
such as how the brain accesses information, integrates sensorimotor functions,
or controls executive functions (attention,
anticipation, planning, learning of rules, abstract thinking, etc.).
Today,
thanks to new technologies such as brain imaging (follow the Tool Module link
to the left), each of these phenomena can be subjected to experiments that the
neurobiologists of the early 1980s could never have imagined. The findings of
such experiments are gradually shedding light on these problems and tending to
invalidate
certain models of consciousness that are no longer consistent with the experimental
data.
This cartoon by Saul Steinberg, published on the cover
of The New Yorker on October 18, 1969, vividly depicts the human
“stream of consciousness”: an endless succession of items that come
to our attention, and then are sometimes encoded in memory, but other times forgotten
forever.
As a result,
attempts have been made to develop new models that incorporate these data on the
brain’s anatomy and function and integrate them into a coherent whole that
explains the various aspects of our conscious processes. The following paragraphs
briefly describe some of these neurobiological theories of consciousness. Many
of these theories have some concepts
in common, which the most optimistic of their authors believe points to the
beginnings of an overall explanation of how the brain goes about “producing”
consciousness.
As phenomena such as change
blindness demonstrate, we are not conscious of so many things as we think
we are. Philosopher Daniel Dennett had already predicted the existence of this
phenomenon back in the early 1990s, long before experiments had provided the spectacular
examples of it that we know about today.
That is why
Dennett considers introspection such a bad starting point for any attempt to understand
consciousness. Our immediate subjective access to our own consciousness makes
us think that we can know its workings directly but masks its true nature, which
is proving more and more counterintuitive.
Dennett
does not categorically deny the insights that
the first-person perspective can offer, through the practice of meditation,
for example. But he would like it if everything that was discovered from this
first-person viewpoint could then be validated by neutral observers— in
other words, from the third-person viewpoint—by the scientific method. From
this stance comes Dennett’s concept of hetero-phenomenology,
which comprises not only what a subject reports about his or her conscious experience,
but also the data that can be gathered concerning the subject’s brain and
the subject’s immediate environment.
We generally sense that the winning
interpretation of a conscious state arises from the Darwinian
competition that plays out among its many unconscious versions. But sometimes
we get a glimpse of this subliminal struggle - for example, if you are walking
through a snowstorm, and see a person’s silhouette, but as you draw closer,
you suddenly realize that a rival interpretation has taken its place, and that
what you are really seeing is an oddly pruned apple tree.
Daniel Dennett has been greatly
influenced by evolutionist thinkers such as Richard
Dawkins, who introduced the concept of “memes”
and who, like Dennett, staunchly criticizes creationism.
Daniel
Dennett is a philosopher, but probably one of the philosophers who are
trying the hardest to integrate the findings of neuroscience into their conception
of consciousness.
Dennett wants to put
an end to what he calls Cartesian materialism, a position that rejects Cartesian
dualism but accepts the idea of a central
(though material) theatre from which consciousness springs. To escape from
this Cartesian materialism, Dennett proposes two metaphors that he considers more
consistent with the neuroscientific data: “multiple drafts” and the
“virtual machine”.
For Dennett,
there is no little homunculus inside the human brain, no little “self”
sitting in front of the theatre of consciousness and observing or even directing
the spectacle before it. In his multiple drafts model, consciousness
is not a unitary process, but a distributed one. At any given time, several concurrent
neuronal assemblies are activated in parallel and competing with one another to
be the centre of attention—to be “famous”, as Dennett puts it.
The conscious self is therefore nothing more than “fame in the brain”,
fragile and changing like any other ongoing reconstruction.
Thus, the result of this competition
is not an “average” of the various drafts, but rather the success
of the one that is the most effective in and best adapted to a given situation.
What Dennett is positing here is a process of selection similar to what Edelman
has called “neural Darwinism” (follow the Tool module link).
According
to Dennett’s multiple-drafts model, a given piece of conscious content arises
from a rapid succession of events in the brain whose order in time cannot be determined.
Various neuronal assemblies distributed throughout the brain respond to the various
properties of an object (during an interval of about one-fifth of a second). Consequently,
the answer to the question “When did I become conscious of such-and-such
an event?” can only be vague, never precise. The following diagram attempts
to illustrate this point.
The first thing that the brain detects is simply
that something has happened. Next, the brain determines that this something is
located toward the left, and that it is a circle. The brain then detects that
the circle is blue, and finally
binds all these elements together and determines that the object is a blue
circle to the left. There is no way of determining precisely when the brain became
conscious of this blue circle, because each
of its characteristics has been detected a few hundredths of a second before
or after the others.
According
to Dennett’s model, with these multiple drafts in constant competition,
the single, conscious narrative flow so familiar to all of us and depicted in
the Steinberg cartoon above can only be an illusion, just like the Cartesian theatre.
To explain this illusion, Dennett posits the existence of a virtual machine
that runs in serial fashion on a “hardware platform” that
operates in massively parallel fashion—i.e., the human brain. This virtual
machine would be a sort of mental operating system (like Linux or MS-DOS), capable
of transforming the inner cacophony of the brain’s parallel activity into
a conscious, serial narrative flow. It is this virtual machine that would enable
us to think about our own thoughts and engage in deliberations with ourselves.
Unlike the behaviourists,
Dennett thus does not dismiss what each of us may subjectively report about our
emotions, feelings, or mental states. But he does not grant these reports any
special status either. He regards them as real data, but as data concerning the
way that people feel things, and not about how these things actually are (see
sidebar on heterophenomenology). For Dennett, the task clearly becomes to understand
how this illusion is generated.
Dennett thinks that many people
do not want to consider explanations about consciousness, just as many people
do not want to hear explanations of magic tricks. To such people, the aura pf
mystery surrounding the “hard problem”of
consciousness seems impenetrable. But Dennett says their error begins
when they conceptualize “the hard problem” as a single problem and
consciousness as a single thing.
To explain what he
means, Dennett uses the example of a famous magician who showed his colleagues
a card trick that he called “The Tuned Deck”. A volunteer
chose a card from the deck, noted which card it was without telling the magician,
then returned it to the deck. The magician then held the deck close to his ear,
riffled the cards, and successfully identified the volunteer’s card, claiming
that he had done so by listening to the tiny variations in the sounds of the individual
cards. His colleagues thought that they had figured the trick out, so they asked
him to repeat it, and he obliged. Fooled once more by the trick, but still doubting
how he had really done it, they had him to do it again, and once again, they experienced
the same perplexity together with the same vague feeling that they had figured
out the subterfuge.
At the end of his life, the magician
revealed his secret to his colleagues. The trick was simply the name of the trick:
“The Tuned Deck”, and more specifically the word “The”.
By using this word, the magician implied that he was performing a single trick,
but in fact, every time he found the volunteer’s card, he was using a different
method. And his colleagues were familiar with all of these methods, because identifying
a card chosen by a volunteer was a classic trick that could be performed in many
different ways. But the other magicians never identified their colleague’s
methods, because they were too busy looking for a single trick.
And
that, says Dennett, is exactly what happens with “the” hard problem
of consciousness. We remain in the dark, because consciousness is a whole set
of “tricks”, and no one of them constitutes the solution all on its
own. The solution consists of all these tricks taken together . The error with
consciousness is the same as with the cards: imagining that there is one specific
problem beyond the numerous “tricks” that our consciousness performs.
And the same thing goes for qualia,
about which Dennett is no more sparing. He describes them as the residue that
is left after one has explained whatever there was to explain about perception,
for example. In other words, the error here lies in continuing the analysis without
realizing that one has already completed it and that there is nothing more to
analyze. Or stated yet another way, the error consists in seeing insurmountable
problems in a fairly ordinary fact: that any subject is inexhaustible. Even for
a single grain of sand, once everything has been said that there is to say about
it, there will still always be something that has not been said—about its
history, for example.
Playing a variation on the theme
of the hard problem versus the easy problems of consciousness, philosopher Ned
Block developed what has become a classic distinction between “phenomenal
consciousness” and "access consciousness".
Phenomenal
consciousness is the kind addressed by question such as “what
it is like to be a bat”, or a human being. It is this qualitative aspect
of our mental states that poses a problem for Block (“Why are the neurobiological
bases of certain subjective experiences precisely the ones they are and not something
else?”) but not for Dennett (see the sidebar above).
On
the other hand, a representation is conscious, in the sense of access
consciousness, if and only if it is sufficiently distributed
within a cognitive system so as to be freely used in reasoning. We are thus dealing
with something fairly close to the global workspace model proposed by Baars.
The global workspace model is compatible
with two phenomena that are much discussed by students of the neurobiological
bases of consciousness: the probable
“binding function” (the subject of the “binding problem”)
and the concept
of an “interpreter” , which comes from Michael Gazzaniga’s
research on “split brain” patients.
With
regard to the binding problem, the temporal
synchronization of neuronal oscillations is often cited to explain how numerous
specialized subsystems pool the results of their work. With regard to the interpreter
concept, the interpreter would be the function that makes “public”
the unconscious work done by a multitude of independent agents.
The theories
that invoke the concept of a global workspace date back to the
work done by Alan Newell and Herbert Simon in the cognitive
sciences in the 1960s and 1970s. Newell and his colleagues were the first
to show the usefulness of a global workspace in a complex system composed of specialized
circuits. By providing a place to pool the information that each of these circuits
had processed, this workspace would allow problems to be solve that no one of
them could have solved on its own.
That
in short is the great principle of the global workspace, which scientists have
continued to apply and enhance ever since. In fact, most neurobiological models
of consciousness incorporate some aspects of the global workspace concept. Examples
include Gerald
Edelman’s “global cartography”, Rodolfo
Llinas’s mechanism of global synchronization from the thalamus, Antonio
Damasio’s cortical convergence zones, Daniel Schacter’s “conscious
attention system”, Francisco Varela’s
“brainweb”, and the model of Jean-Pierre
Changeux and Stanislas Dehaene, who make the neuronal workspace their primary
hypothesis.
But ever since the 1980s,
it is Bernard Baars who has been the strongest proponent of this
model, which attempts to answer the famous question: how can a phenomenon such
as consciousness, in which everything happens in series, with only one conscious
object at a time, emerge from a nervous system that basically consists of countless
specialized circuits that operate in parallel and unconsciously? Baars’s
answer: by having a workspace where the information processed by these specialized
circuits is made accessible to the entire population of neurons in the brain.
Source: Dehaene and Naccache (2001)
Baars
thus suggests that there is a close connection between global availability of
information on the one hand and consciousness on the other. In his view, this
global accessibility of information, made possible by a global workspace, is precisely
what we subjectively experience as a conscious state. The
“easy problems” and the “hard problem” of consciousness
are thus here regarded as two
different sides of the same coin.
The
global workspace is a process that involves first convergence and then divergence
of information. Baars thinks that this process can best be understood through
the metaphor of a theatre stage where attention acts as a spotlight cast on certain
actors. These actors represent the content of consciousness selected by competition
among specialized circuits. What we have here is a Darwinian
selection process by which certain actors (pieces of conscious content) become
“famous” for a fleeting moment. During that moment, this conscious
information is disseminated or made accessible to the vast audience of unconscious
circuits that fills the theatre.
Baars
argues that this way of formulating the theatre metaphor does not make it what
Dennett calls the “Cartesian
theatre” and criticizes so harshly. Indeed, the Cartesian theatre has
always been predicated on the existence of a single point, such
as the pineal gland identified by Descartes himself, where everything is brought
together within a “self” that receives the conscious thought or perception—almost
as if there were only one spectator in the audience watching the entire stage.
In contrast, in Baars’s theatre
metaphor, a multitude of entities, all of which remain unconscious, have access
to a particular piece of information at the same time. Baars thereby not only
avoids the problem of regression to infinity,
but also offers a straightforward definition of consciousness as this in-depth
exchange of information among brain functions, each of which is otherwise independent
of the others and unaware of what they are doing.
Many of these capabilities
show up as integral parts of Baars’ theatre metaphor, when you get into
the details. The spotlight of attention determines what place on the stage will
be illuminated and therefore conscious. The stage as a whole represents the content
of working memory that is immediately accessible to be picked out by this spotlight.
The part of the stage that is in the “glare of the spotlight” will
be seen, or, if you prefer, distributed to the unconscious audience seated in
the semi-darkness, while other unconscious craftspersons, working behind the scenes,
influence the course of the performance by associating a particular context with
it.
Baars reminds his readers that the
purpose of these details of the theatre metaphor is not to “explain”
consciousness, but to provide tools that can be used to organize the existing
data, to clarify certain concepts, and to formulate testable hypotheses, in particular
the involvement of certain
structures in the brain, so as to better understand this complex phenomenon.
Simulations
have shown that a global workspace based on the thalamocortical
circuits and on long corticocortical connections behaves like a non-linear,
self-amplified dynamic system.
One of the
characteristics of such a system is that it has a threshold above which an activation
propagates explosively across the entire network. In contrast, an activation slightly
below this threshold (such as a subliminal activation) damps itself out very rapidly.
Another characteristic of such a system is that it
can inhibit other activations within this same global workspace, thus preventing
rival stimuli from being processed consciously. Such inhibition is seen in the
phenomenon of the “attentional blink”, in which, for a short time
after we are presented with one stimulus, we cannot detect a new one.
Lastly,
the dynamics of access to consciousness are an abrupt, “all-or-nothing”
phenomenon: when the visibility of a stimulus is increased gradually, it “appears”
in consciousness suddenly. The global workspace model may account for this threshold
effect by means of the connections from the neurons in the workspace
back to the primary and secondary sensory areas.
Some authors
agree that talking about a conscious workspace probably represents some progress
over the idea of consciousness that prevailed in the mid-20th century. But others
think that this idea is some like the earlier notion of the “dormitive virtue”
of opium, which provided no information at all about the underlying
mechanisms by which opiates affect the brain.
What
then does this global workspace consist of, in concrete terms? If it is a set
of neural pathways that interconnect the various parallel processors in the brain,
then which pathways are they? And what determines whether the activation of any
given set of neurons will be propagated into the global workspace?
Many
neurobiologists, including Jean-Pierre Changeux and Stanislas
Dehaene, are conducting research programs to attempt to answer these
questions. Changeux and Dehaene start from the premise that the brain does in
fact contain a conscious global workspace that combines all of the information
quietly processed in the background by the numerous independent, unconscious modules
to which it is connected.
Noting that
layers
II and III of the prefrontal, parieto-temporal, and cingulate cortexes contain
pyramidal
neurons with long axons that could reciprocally interconnect distinct cortical
areas, Changeux and Dehaene identify these circuits as possibly constituting the
neuronal substrate of this global workspace. These
circuits become activated only during conscious processing and are very strongly
inhibited in individuals who are in a vegetative state, or under general
anaesthesia, or in a coma.
Both of these facts lend support to Changeux and Dehaene’s hypothesis. In
contrast, the brain’s unconscious processors are more localized, generally
in the sensory cortical areas.
In 2006,
Changeux and Dehaene’s research using brain-imaging techniques led them
to distinguish three major forms of mental processing in the human brain: subliminal,
preconscious, and conscious.
For
conscious processing to occur, there are two conditions that that must be met:
a sufficient level
of alertness (being awake rather than asleep, for example) and sufficient
bottom-up activation, in other words, a response
must be occurring in the primary and secondary sensory areas.
But
these conditions do not suffice to let a stimulus enter consciousness. For example,
some individuals may be awake and display an activation of the extrastriate visual
areas of the cortex, yet deny having seen any stimulus whatever. Hence,
according to Changeux and Dehaene, for a piece of information to access consciousness,
a third condition also must be met: the activation of the associative cortexes
by these neurons with long axons that create a reverberation between distant neuronal
assemblies.
But why do certain pieces
of information become conscious while others do not? To explain this phenomenon,
Changeux and Dehaene distinguish four situations.
In
the first situation, the information remains unconscious, because it is processed
by circuits that are not anatomically connected to the conscious neuronal workspace
(for instance, the circuits that regulate the body’s digestive functions).
But in the three other cases, the access door to the conscious workspace does
exist.
•
If the signal is too weak, the processing of the information remains localized
in the unconscious processor and dissipates rapidly; this corresponds to the subliminal
state, which remains unconscious because the degree of activation stays below
the threshold needed to enter the self-amplifying loops of the conscious workspace
(see sidebar).
• If the activation is
strong enough to spread into several specialized sensorimotor areas, but the individual’s
attention is not focused on these stimuli, then the information remains inaccessible
to consciousness. It is therefore in a preconscious state, which
can, however, become conscious if it undergoes sufficient amplification from
the top down.
•
Lastly, a stimulus can become conscious in two ways: if it causes
us to deliberately turn our attention to a preconscious state, or, if this signal
is so strong and unexpected that it forces its way into the conscious workspace.
An example of the first case (“top-down”
activation) would be when you decide to eavesdrop on a nearby conversation
at a cocktail party. An example of the second (“bottom-up”
activation) would be what happens when someone calls out your name or yells
“Stop, thief!” at that same party.
In
these three illustrations, the darkness of the coloured dots is proportional to
the intensity of activation, the small arrows represent the interactions among
the specialized circuits, and the large arrows represent the direction of attention
from the top down toward the stimulus or toward something else. Note that there
is a continuum of states between the first illustration and the second, but a
sudden transition between the second and the third. After Dehaene et al.,
2006.
But in
both of these last two cases, the transition from preconscious to conscious is
abrupt, as is the case for any non-linear, self-amplified dynamic system (see
sidebar). This transition is also always accompanied by an activation
in the parieto-frontal and anterior cingulate cortex.
When
this activation reaches several richly interconnected associative areas, two things
can happen. The activation can undergo a reverberation in the workspace and thus
remain available for a much longer time than the initial stimulus lasts. The activation
can also rapidly propagate into several specialized systems of the brain that
can then make use of it.
In contrast
to the classic binary distinction between not conscious and conscious, this model
divides the non-conscious state into subliminal and preconscious ones, thus providing
a classification into three distinct states (or four, if you count vegetative
regulatory processes, which by definition are always unconscious).
This
distinction makes sense, in light not only of the results of brain-imaging experiments,
but also of phenomena such as inattentional
blindness, which show that even when there is substantial preconscious activation,
a stimulus can remain unconscious if the individual is not paying attention to
it.
“The drama of the human condition
comes solely from consciousness. Of course, consciousness and its revelations
allow us to create a better life… but the price we pay for that better
life is high. It is not just the price of risk and danger and pain. It is the
price of knowing risk, danger, and pain. Worse even: it is the price
of knowing what pleasure
is and knowing when it is missing or unattainable.”
-
Antonio R. Damasio, The Feeling of What Happens: Body and
Emotion in the Making of Consciousness (1999)
To the
various levels of accessibility of the contents of consciousness described by
Changeux and Dehaene, another continuum must be added: that of the brain’s
ability to form its own representation of the “self”. How does this
representation of self contribute to conscious experience? This question has been
central to the concerns of researchers such as Edelman,
Tononi, Llinás,
and, especially, Antonio Damasio.
In
his book Descartes’ Error: Emotion, Reason, and the Human Brain,
published in 1994, Damasio argues that conscious thought depends substantially
on the visceral perception that we have of our bodies. Our conscious decisions
arise from abstract reasoning processes, but Damasio shows that these processes
are rooted in our bodily perceptions, and that it is this constant monitoring
of the communications between the body and the brain that lets us make informed
decisions.
This is what Damasio means
by his concept of “somatic markers”, which also clarifies the role
and the nature of the emotions from an evolutionary
standpoint. The somatic manifestations of the emotions are processed in working
memory so as to “mark” perceptual information from the external
environment with an affective value and thus assess the importance of this information
to the organism. This process is essential for any decisionmaking that involves
the survival of the organism in question.
In
a later book, The Feeling of What Happens:
Body and Emotion in the Making of Consciousness(1999),
Damasio develops his model further to account for the various possible levels
of consciousness of self. The visceral monitoring described above becomes the
proto-self, which is a moment-to-moment perception of the body’s
internal emotional state and is made possible by the insula,
among other structures in the brain.
Subsequently,
a perception of the outside world becomes conscious when it is placed in relationship
to this proto-self. This higher-order relationship, which Damasio calls core
consciousness, answers questions such as “What am I feeling about
the scene I am looking at (or the sentence I am reading )?” Many species
of animals may be equipped with this feeling of the “here and now”.
A
third level, extended consciousness, becomes
possible when the individual can form a mental representation of his or her own
conscious experiences in the past or the future by means of memory and the higher
functions that make abstract conceptualization possible.
Thus, according
to Damasio, this autobiographical consciousness—the sense that humans have
of being themselves and not someone else—is grounded in all those moments
in life when our core consciousness attributes affective value to what we are
experiencing. As a result, this autobiographical self is in a constant process
of
reconstruction that is both informed by the past and influenced by our expectations
for the future.
Damasio also
distinguishes extended consciousness from what is referred to by the general term
“intelligence”. According to Damasio, extended consciousness provides
access to the widest possible body of knowledge, whereas intelligence is related
more to the ability to manipulate this knowledge in order to devise new behavioural
responses.
Damasio also regards this
uniquely human extended consciousness as the source of such faculties as creativity,
systematic consideration of others, and moral
consciousness .
Damasio’s
proposed three-level hierarchy of consciousness of self can be found with many
variants in the works of other authors
from our own time as well as the past.
For example,
Damasio’s core consciousness corresponds to William James’s agent-self
or to Endel Tulving’s noetic consciousness (though these terms do not overlap
completely). The general expression “awareness” is
often used to describe this core consciousness or “attention-consciousness”—our
state of alertness to our environment.
Human beings
would thus share primary consciousness with most other animals that are equipped
with sophisticated sensory organs and a complex brain. The term “creature
consciousness” is also used to describe this elementary form of consciousness
which, for example, enables flies to navigate through the air and presupposes
that they know how to distinguish between their own movements and what is happening
in the world around them.
Newborn human babies, on
the other hand, are still incapable of distinguishing between themselves and the
rest of the world, including their mothers. The consciousness of self develops
between
the ages of six months and one year.
The consciousness
of existing as a person in time—what Damasio calls extended consciousness—is
often classified into two types.
The first is reflexive
consciousness (or introspective consciousness), which
mainly stresses that it is I who am perceiving, I who am directing my attention
toward such-and-such an object or such-and-such a thought, I who am controlling
my reasoning or my behaviour. It is this consciousness, of being aware of something,
that appears to be common to humans and the great apes.
This
reflexive consciousness would be the necessary condition for consciousness
of self—knowing my own personal history, knowing why I am where
I am at the present moment or why I will be somewhere else tomorrow night. It
is Tulving’s autonoetic consciousness, James’s object-self—in
other words, the construction of self in time, which does not begin before age
2 in human beings. It is this consciousness that eventually enables us to tell
our life stories, to act our role in life’s drama, and to modify
our memories as this drama unfurls.
This idea
that there are various degrees of consciousness and that the emotions constitute
a fundamental elementary form of it is defended by many scientists, including
neuropharmacologist Susan Greenfield. Greenfield stresses that
the activity of the neuronal
assemblies distributed in the various parts of the brain is constantly modulated
by the neuromodulatory
neurons associated with an individual’s emotional state.
There
are also some molecules, in particular some peptides,
that may influence not only the forming of assemblies of neurons in the brain,
but also the
body’s hormonal and immune system. The feedback loops between the brain
and the body are therefore due not only to the vegetative
nervous system, but also to all of these chemical molecules that can act on
the brain and the body simultaneously.
From
this perspective, consciousness, whose most primary levels appear to be related
to the emotions, is not simply a matter of brain activity, but rather an experience
involving the body as a whole. This description comes closer to the conception
of the human mind advanced by Freeman and Varela, who attribute a central
role to the individual’s body situated in its environment. These authors
thus take exception to the traditional cognitivist
approach, in which the human brain is regarded as a system that applies rules
to manipulate internal representations of the world.
Until the mid-20th century, scientists
classified natural phenomena into two categories: phenomena that were random (and
therefore unpredictable), and phenomena that followed deterministic rules (and
hence were predictable—in other words, if you knew their initial conditions,
you could predict their future behaviour).
But then
scientists noticed that there were certain systems that could be described by
deterministic rules but wherein a slight modification in their initial conditions
could suffice to make their behaviour unpredictable. These systems that are sensitive
to initial conditions are said to be “chaotic”.
The magnitudes that define these chaotic systems do
not vary in an absolutely random and infinite manner over time. Instead, they
appear to be contained by an element of order called a “strange attractor”.
The presence of strange attractors is one of the main characteristics
of “deterministic chaos” (to learn about the mathematical
tools used to model deterministic chaos, see the box to the right).
We will
now briefly discuss neurophysiologist Walter J. Freeman’s
approach to the problem of consciousness. But to do so, we must first go back
down to the cellular (neuronal) level, because Freeman focuses less on the anatomy
of brain structures than on the
way that neurons communicate with one another and the patterns of activity
to which this communication gives rise in the brain as a whole.
Freeman
observed that the neuronal connectivity of the human brain engenders chaotic activity
that, like weather phenomena, follows the laws of non-linear dynamics (or “deterministic
chaos”—see sidebar). He therefore applied the mathematical tools of
non-linear dynamics to interpret the electrical states observed in the brain.
By analyzing the electroencephalograms
(EEGs) of human brains while they were performing many tasks, Freeman showed that
the
various rhythms of the human brain do it fact follow the “laws”
of spatial/temporal chaos.
Behind what
seems to be nothing but noise, these chaotic fluctuations actually display regularities
and properties consistent with those of human thought. One example is the capacity
for rapid, extensive changes. Another is the brain’s ability to almost instantly
transform sensory inputs into conscious perceptions.
Vast
assemblies
of neurons can change their activity pattern abruptly and simultaneously in
response to a stimulus that may be very weak. This destabilization of a primary
sensory cortex travels to other areas of the brain, where it is “digested”
in a way that is specific to each individual, depending on the content of that
particular individual’s memory. This is why, according to Freeman, far from
being harmful, the chaotic activity of millions of neurons is what makes all perception
and all new thoughts possible.
This is
also why Freeman thinks that phenomena such as consciousness cannot be understood
solely by examining the properties of individual neurons. Just as a hurricane
that develops from the collisions of billions of air molecules affects the physical
environment, so an overall pattern that emerges from the activity of billions
of neurons throughout the cortex affects, for example, the motor
areas of the brain so as to produce a given set of body movements. In the
latter case, the loop is then closed when the environmental changes caused by
these movements in return produce a perception that transforms the brain’s
general activity once again.
Freeman
therefore does not regard perception and action as two independent phenomena,
one input, the other output. Instead, he sees them as the
same process enabling the individual to act upon the world. Thus he goes further
than the idea of perception as an active phenomenon that is currently accepted
today.
The rhythmic
synchronization of the neurons occupies a central place in Freeman’s
model. Its role, according to Freeman, is to co-ordinate activity among the various
areas of the brain so that physically distant sets of neurons are combined into
a single functional activity pattern. Every brain, by virtue of its history, generates
a unique context (a “chaotic attractor”, in the terminology of chaos
physics) in which meanings develop from such activity patterns. Consciousness
would then be the highest-level pattern that connects these meanings to one another.
It is not in itself the cause of any neuronal effects; it is rather a way of harmoniously
and globally interconnecting the fluctuations in the brain so as to facilitate
their interaction.
This model offers
a way out of the impasse of trying to determine the origin of conscious will through
linear causal reasoning: instead it proposes a process of circular causality—an
epistemological breakthrough made possible by the mathematics of chaos. The origin
of a conscious action therefore should be sought not solely in the brain of the
individual who takes that action, but also in that individual’s continuing
relationship with other individuals and with the rest of the world. Because according
to Freeman, what we call our decisions are constructed in real time by the behaviour
of our entire bodies, and we
are informed of these decisions on the conscious level only after a slight delay.
Thus consciousness would intervene only to smooth out the various aspects of an
individual’s behaviours, to modulate them, and probably to legitimize
them in relation to the entire set of meanings that constitute that individual’s
personality.
In the terminology of
dynamics, consciousness would correspond here to an “operator”, because
it modulates the brain dynamics from which past actions have arisen. This conception
is consistent with a hypothesis proposed by William James in 1878, that consciousness
interacts with the brain’s processes and is neither an epiphenomenon
nor a first cause.
Chaotic systems (see
sidebar) are both random and determined (which is no small contradiction). They
may comprise an infinite number of unstable cyclical movements of various frequencies,
which makes them very useful for describing the cognitive processes that arise
from various
neuronal oscillations.
The ideal tool for analyzing
such systems is the language of non-linear mathematics, which
makes it possible to distinguish between the mere “noise” in disorderly
systems and the hidden order in chaotic systems. Researchers are constantly discovering
new phenomena that cannot be understood through an analysis of their components
alone. Some of these phenomena are physical (such as aerodynamic turbulence),
others are chemical (such as oscillating chemical reactions), and still others
ecological, meteorological, or even economic. In order to understand such phenomena,
it is absolutely essential to analyze their behaviour as a whole, which can be
done by means of non-linear mathematics.
“In the course of evolution,
when did the neuronal system appear? Not in plants, not in fungi, not in bacteria.
It appeared in animals. To feed themselves, animals found the solution of eating
prey. They therefore had to move, and locomotion is the constituent logic of the
animal. So that was when the neuronal system appeared, because to hunt, to move,
you need a perception-action loop.”
- translated
from an interview with Francisco Varela
in La Recherche, No. 308, April 1998.
The dynamic
approach just described in terms of Walter J. Freeman’s work actually is
part of a broader theoretical approach called “embodied
cognition”. Contrary to the computational
approach, the dynamic approach deals with neuronal activities rather than
with symbols, and with global states of the brain observed by means of functional
brain imaging, rather than with calculations and rules.
Embodied
cognition theory, of which Francisco
Varela was one of the greatest proponents, challenges the separation
between human cognition and its embodiment. For Varela, and for the many researchers
who have come to embrace this approach, we cannot understand cognition, and hence
consciousness, if we abstract it from the organism embedded in a particular environment
with a particular configuration.
In these
“ecologically situated” conditions, in which “situated cognition”
occurs, any perception results in an action, and any action results in a perception,
as we have just seen with Freeman. Hence a perception-action loop constitutes
the foundational logic of the neuronal system. Cognition, consciousness—in
short, the individual’s entire inner world—emerges from that individual’s
actions; it is an “enacted” world
. Cognition is continuously enhanced by movement,
because the human brain has been constructed in this way throughout
the history of phylogenesis.
Non-linear
mathematics have also contributed greatly to our understanding of the phenomena
of self-organization
and emergence that are inherent
in the embodied approach, through the concept of an “attractor”, borrowed
from the theory of dynamic systems.
Current
knowledge thus irretrievably distances us from the traditional, causal model of
consciousness as a simple matter of input, process, and output, patterned on the
operation of computers. Instead, the ideas of circular causality, and the primacy
of action, the emotions, and a living body in a given environment provide us with
a richer matrix in which to attempt to understand human consciousness.
The image that the nervous system
provides of the world may vary hugely from one species to another. For instance,
our perception of colours is no doubt very different from a pigeon’s, because
we
have only three types of cones, whereas pigeons have five. And there are
other species that have sensory modalities that we lack, such as a bee’s
ability to sense ultraviolet radiation, or a bat’s
ability to itself by echolocation.
How then can
we argue that the human brain provides us with a consciousness of our environment
that corresponds to the one true reality? What about all the other aspects of
this environment that we don’t perceive at all? We are forced to admit that
the activity of the human brain, like that of other animal brains, instead seems
to create a virtual image of what we regard as reality—an evolutionarily
useful sampling of the world. To paraphrase British psychologist Max Velmans,
the universe has differing views of itself through my eyes, your eyes, the eyes
of a pigeon, and the eyes of a bat.