The
Problem of Qualia
By BRENT SILBY
Department of Philosophy
University of Canterbury
Copyright (c) Brent Silby
1998
www.def-logic.com/articles
One of the major concerns for the contemporary philosophy
of mind involves the problem of qualia. The word qualia refers to
our subjective experience of the world and includes the properties
of our experience that cannot be located in the world external to
our minds. For example, the ineffable feel of a blue experience
when one looks at the sky, or the pain one feels when one
is stuck with a pin. These sensations are the essence of our experience
and yet cannot be pointed to in the external world.
This paper is based upon a talk I gave to the 1998 cognitive
science class and will follow the same structure of that talk. To
start, I will describe the problem of qualia and will show that it
poses a real problem for physicalist and functionalist theories of
the mind. Further highlighting the problem, I will talk about androids.
These entities of the future are functionally identical to humans,
however it is claimed that if they existed, they would have no qualia
or any conscious experience whatsoever. The final section of this
paper will be concerned with Frank Jackson's Knowledge argument. The
Knowledge argument is a very powerful thought experiment, which is
supposed to show that physicalism about the mind is false.
The Problem of Qualia
The existence of qualia is a problem for cognitive science
and philosophy of mind because it seems that our subjective sensations
of experience only exist within the mind. According to the traditional
view, qualia are characterised as being:
(1) ineffable
(2) intrinsic
(3) private
(4) directly apprehensible in consciousness.1
Sensations of colours, pains and sounds seem very real
to us and yet they cannot be found in the world external to our minds.
If we wish to ask what science can tell us about the colour blue,
all we learn is that there is a certain wavelength of electromagnetic
radiation that is being reflected off various objects. There is no
blue, just radiation. The question we are left with is: where is the
blueness if it is not in the outside world? One possible
answer would be to describe the visual process like this:
When light enters our eyes, various frequencies
of radiation cause the retina's cones to activate. This initiates
an electrochemical cascade by which signals are sent to the visual
cortex and a neural state is set up. A connected `box' is then filled
with a representation of the colour, which can then be examined by
other parts of the brain in order to produce behavioural output.
Now, it certainly feels like this is going
on in our brains but there are problems with this picture. Daniel
Dennett suggests that it would be a waste of resources for the brain
to work in this fashion. Once a sensory discrimination has been made,
there is no need for it to be made again by other parts of the brain.
In other words, the visual cortex has discriminated the colour so
the colour representation box is unnecessary. The brain does not need
to contain a box within which a colour is represented in order to
be examined by other parts of the brain. Nature has given the brain
a visual system that can make the colour discriminations by itself.
But this brings us back to the question: Where is the blue
that we experience? There is no blue in the outside world, and there
is no box filled with blue colour inside our heads so why do we have
the blue sensation? It is this question that has led many philosophers
to the opinion that qualia does not supervene on the functional organisations
of the brain. This is to say that even if we had a complete understanding
of the brain's functional organisation, we would not be able to account
for qualia because they cannot be captured by a functionalist account
of the mind. So, we could imagine a situation in which qualia do not
exist and yet the brain's functional states remain the same. Many
thought experiments have been devised to show this point. They usually
involve imagining an entity that is functionally (and sometimes physically)
identical to us, but has no conscious experience whatsoever.
What would it be like to be an Android?
Computer technology is progressing at an astonishing
rate and is likely to continue to grow. Humans have designed computer
programs that can distinguish colours with great accuracy but we do
not suppose that these computer programs experience any colour qualia.
But what about computers of the future? A great deal of effort is
being put into researching artificial intelligence and we can imagine
that at some point in the future, humans will successfully design
intelligent androids.2 Now, if we imagine that these androids
are behaviourally identical to humans we can ask the question: do
they experience qualia? Would the designers of these androids need
to build qualia into their programming? Presumably not. If we consider
pain, we can imagine that the android's software would contain a pain
handling routine similar to this:
The pain handling routine this flow chart depicts is
simplified but the point is clear. Once instantiated in an android,
it will monitor sensory input and will jump to other sub-routines
depending on the strength of the input. There are no qualia included
in this system and it would appear that including feeling
would not be necessary. The system is effective as it is. An android's
brain could be full of similar routines that handle all of its sensory
input and it would survive in the world perfectly all right but without
the feel of experience. Of course the android has been designed
to imitate human behaviour, so if we asked it a question about its
experience, it would respond in much the same way as we would. The
point is that we could not tell from its outward behaviour that there
is nothing it is like for it to be an android. Now, what could we
say about qualia if we could show that our brains work in a similar
fashion to the android brain? Qualia would seem to be an unnecessary
bonus that escapes our functionalist picture of the mind. Of course
we could question the idea that our brains work in that way but the
point should be clear. The android scenario shows us that beings without
qualia could be behaviourally indistinguishable from beings with qualia.
Let's alter the thought experiment by imagining that
neuroscientists have a complete understanding of how the brain works.
We can imagine that they supply cybernetics researchers with a complete
wiring diagram of the human brain. The cybernetisists could then construct
an android whose brain is functionally identical to the human brain.
Its silicon brain would work in exactly the same way as the human
brain and its behaviour would be identical to human behaviour. Such
an entity is known as a `Functional Isomorph'. The only difference
between its brain and a human brain is what materials it happens to
be constructed of. Now, here's the important question: Would the android
be conscious and experience qualia? Intuitively, many people would
say no. They might claim that the android brain is not the right type
of system to give rise to conscious experience. On the other hand,
many people (myself included) may claim that if the android's brain
was functionally identical to our human brains, it would be conscious
(in so far as we are conscious). But that might imply that any
system would be conscious if it was functionally identical to the
brain. Ned Block points out that it would be possible to get every
person in China to simulate a neuron and allow them to simulate neural
connections by equipping each of them with cell phone links. We can
imagine that if this was carried out effectively, China would be functionally
the same as the human brain. But, says Block, this system would not
be conscious.3 Intuitively it is tempting to agree with
Block. It certainly seems that the population of China (like an android
brain) is the wrong type of system to exhibit a collective consciousness.
But surely we could make the same claim about the human brain. We
could say that intuitively it seems implausible that this spongy organic
brain would be the right type of thing to give rise to conscious experience.
But it does.4 This does not prove that an android brain
or the population of China would be conscious, but it is an effective
counter to the claim that these systems would not.
The android and Chinese nation examples of `absent qualia'
are supposed to convince us that the functional organisation of the
brain cannot account for qualia. These thought experiments appeal
to intuition and are easily answered: If the functional organisation
of the brain can give rise to qualia, then why not the same structure
instantiated in a silicon system? The absent qualia supporters have
two possible replies to this question:
(1) Qualia and states of consciousness are special properties
that cannot be explained by the physical sciences. (This leads to
a Dualist approach to consciousness.)
(2) Qualia require that the functional system upon which they supervene,
be constructed from a certain type of physical material. (Biological
neurons as in the brain.)
Frank Jackson's famous knowledge argument was put forward
in the early 1980's and suggests that reply 1 may be the option to
take. The knowledge argument suggests that qualia do not supervene
on the physical structure of the brain and thus, cannot be explained
in physical terms. Jackson's argument carries a strong intuitive force
and has many supporters. I will now outline the argument and address
some of the standard criticisms of it.
The Knowledge Argument
Jackson asks us to imagine a woman named Mary who is
brought up (from birth) in a black and white room. Throughout her
life in this colourless environment, Mary reads many black and white
books and learns all the laws of physics. As time goes by, Mary becomes
an expert in neurophysiology and of the functional roles that brain
states play in the process of colour vision. Mary's knowledge of the
physical and functional organisation of the brain becomes complete
to the point that there is nothing that she does not know. But, says
Jackson, even with her complete physical knowledge of the brain, Mary
does not know everything there is to know about the brain
because she does not know what it is like to see a colour. Jackson
believes that when Mary leaves her room for the first time and experiences
her first colour, she would learn something new about the world. For
this reason, Jackson claims that the physicalist picture of the brain
does not capture everything there is to know about the mind. The qualia
are left out. The natural conclusion to draw is that the physicalist
story of the brain (and more importantly, the mind) is false. Mary
knew every physical fact about the world, yet she did not
know everything about the world.
Jackson's knowledge argument is certainly intuitive.
It simply seems to be the case that upon leaving her room, Mary would
learn something new about the way the world is. Furthermore, this
knowledge could not be predicted by her complete physical knowledge
of the brain.
Dennett's Reply:
Daniel Dennett asks us to imagine Mary leaving her black
and white room and looking at the sky for the first time. "Ah yes",
she says, "that's exactly what I thought blue would look like." Dennett
suggests that this is exactly what would happen.5 According
to Jackson's story, Mary knows everything about the physical
world and this includes everything there is to know about the neurophysiology
of the brain. Dennett believes that if Mary knew absolutely everything
physical about the world and the brain, she would know exactly
what to expect when she had her first colour experience. It follows
that Mary would not be surprised when she saw the colour red for the
first time. She would have predicted all the neural events that would
occur upon encountering the colour red and she would be able to make
an inference as to what sort of experience she would have. It does
not matter that Mary has not learned about colours in the usual way
because "Mary is not your usual person."6
Dennett's reply to the knowledge argument has attracted
a lot of support. Dennett is trying to convince us that if it were
possible to know every physical fact about the world, then we would
know what colours looked like before we happened to experience them.
This line of thought is counter-intuitive and has not convinced everyone,
but this is to be expected because it is impossible to imagine what
a complete knowledge of physics would be like. It would be difficult
for us to say what Mary would or would not be able to anticipate upon
leaving her room. Robert Van Gulick suggests that it is at least possible
that Mary would know what to expect when she leaves her room.7
There is, therefore, a way of answering the knowledge argument before
discussion needs to take place.
Know-That / Know-How Reply:
Lawrence Nemirow suggested that the knowledge argument
draws on a distinction between propositional knowledge (knowing
that) and procedural knowledge (knowing how).
According to this reply, Mary gains no new knowledge of facts or propositions
about the world. What she gains upon leaving her room are new skills
and practical abilities.8 This is to say she learns how
to recognise the phenomenal properties that her complete physical
knowledge of the world had predicted the existence of. The difference
between these two types of knowledge can be understood by considering
ones knowledge of riding a bicycle. We can construct a complete list
of propositions that tell us the rules involved in riding a bicycle,
but reading this list will not tell us everything about how
to ride a bicycle because procedural knowledge is an ability. Now
Mary, before leaving her room, knows all the facts and propositions
about the world and has a complete understanding of what is involved
in colour vision but she lacks some procedural knowledge. She does
not know how to recognise colours.
This reply to the knowledge argument suggests that upon
leaving her room, Mary gains no new facts or propositions about the
world. I do not agree with this claim. It may be true that Mary gains
new abilities, but surely that is not all she gains when she looks
at the coloured world for the first time. It simply seems to be the
case that when she leaves her room, Mary learns something new about
what the world is like, and this involves propositional knowledge.9
Now, because Mary's complete knowledge of the physics involved in
colour vision falls under the label of propositional knowledge, I
do not believe that making a distinction between these two types of
knowledge offers a refutation of Jackson's knowledge argument.
Churchland's Reply:
One of the most interesting replies to the knowledge
argument was offered by Paul Churchland in his paper "Knowing Qualia:
A Reply to Jackson".10 In this reply, Churchland shows
that the Knowledge Argument is invalid. For clarity, I will state
the Knowledge Argument here:
(1) Mary (before her release) knows everything physical
about other people.
(2) Mary (before her release) does not know everything about
other people.
Therefore
(C) There are truths about other people that escape the physicalist
story.
Churchland describes a possible system by which the
human visual system comes to recognise colours.11
In the human visual cortex, there is a region known
as V4. When an infant human first sees a colour, a neural pattern
is set up within V4 and is reactivated when the infant is exposed
to that colour again. Similar patterns are set up when the infant
is exposed to other wavelengths of light. In abstract terminology,
the V4 centre has partitioned itself into several labels. Each one
of these labels is a neural pattern which represents a different colour.
Now, during the life of the individual, whenever a colour is experienced,
the synaptic weight of the sensory input reactivates the colour's
label in V4. Thus the visual cortex has identified and `labelled'
the wavelength of light that is present in the visual field. The important
point to note here is that these colour recognition labels must be
set up at a very early age and require exposure to colour. Effectively,
people have to learn how to see colours. Of course in Mary's case
these labels will only be suitable for representations of black and
white, and shades of grey because that is all she was exposed to when
she was young. Unfortunately she is now too old for new patterns to
form in the V4 centre of her visual cortex, so when she leaves her
room, she will be unable to see any colours. Her brain will process
the environmental input in terms of what it already knows - black
and white.
In this story Mary is missing much more than an ability,
she is missing a neural processing structure. So, Jackson's premise
(2) seems to be true. There is something that Mary does not, and indeed
cannot know about other people. This story gives an entirely physical
account of visual processing, and as such Mary knows all about this
process. In fact, given her complete knowledge, Mary should also be
aware of her own representational defect as described above. Thus,
it seems that Jackson's premise (1) is true.
Paul Churchland offered the above story as a logical
possibility for the process of colour vision. Whether or not it is
true remains an empirical question. The important point for Churchland
is that it shows a possible situation in which Jackson's conclusion
does not follow from his premises and as such, Jackson's argument
is invalid.
Conclusion
The possibility of absent qualia in the case of android
brains and the Chinese nation thought experiments suggest that it
is possible for systems to be functionally identical to the human
brain and yet have no qualia. I think that the problem with such arguments
lies in the fact that they offer no empirical support for the claims
they make. They rely mainly on intuition and can therefore be answered
by people who have different intuitions. It is not at all clear to
me that androids or other functional isomorphs would not have qualia
as we do.
Frank Jackson's knowledge argument also relies largely
on intuition but I think it does so in a more appealing way. It is
easier for us to imagine Mary's situation than it is for us to imagine
the mind of an android and so the argument carries a stronger intuitive
force. According to Dennett, the problem with the Knowledge argument
is that if we could truly imagine what Mary's knowledge of the world
would be like, we would realise that she would know what
to expect when seeing colours for the first time. This would support
physicalism and provide an answer to the mystery of qualia. To further
refute the knowledge argument, Churchland has given us reason to believe
that the argument is invalid. This was done by describing a physically
possible way in which the visual system works. This would mean that
because of her upbringing, Mary would be neurologically incapable
of experiencing colours. It is for this reason that I believe that
the knowledge argument fails to refute physicalism about the mind
and qualia.
References
Block. N, "Inverted Earth" in The Nature of Consciousness,
edited by Ned Block, Owen Flanagan, and Guven Guzeldere, Massachusetts
Institute of Technology, 1997.
Braddon-Mitchell. D and Jackson. F, Philosophy of Mind
and Cognition, Blackwell Publishers, 1996.
Chalmers. D, The Conscious Mind, Oxford University Press,
1996.
Churchland. Paul, "Knowing Qualia: A Reply to Jackson"
in The Nature of Consciousness, edited by Ned Block, Owen
Flanagan, and Guven Guzeldere, Massachusetts Institute
of Technology, 1997
Dennett. D, Consciousness Explained, Penguin Books,
1993.
Dennett. D, "Quining Qualia" in The Nature of Consciousness,
edited by Ned Block, Owen Flanagan and Guven Guzeldere, Massachusetts
Institute of Technology, 1997.
Van Gulick. R, "Understanding the Phenomenal Mind: Are
We All Just Armadillos? Part 1: Phenomenal Knowledge and Explanatory
Gaps" in The Nature of Consciousness, edited by Ned Block, Owen Flanagan,
and Guven Guzeldere, Massachusetts Institute of Technology, 1997.
1 Daniel Dennett, "Quining
Qualia" in The Nature of Consciousness, edited by Ned Block, Owen
Flanagan and Guven Guzeldere, Massachusetts Institute of Technology,
1997, Page 622.
2 Literature on this subject often refers to robots
but I prefer to use the term android. This is because androids, unlike
robots, are defined as being similar to humans in appearance and abilities.
3 Ned Block, "Inverted Earth" in The Nature of Consciousness,
edited by Ned Block, Owen Flanagan, and Guven Guzeldere, Massachusetts
Institute of Technology, 1997.
4 David Chalmers, The Conscious Mind, Oxford University
Press, 1996, Page 251.
5 Paul Churchland also offers this reply against the
knowledge argument, but I will look at Dennett's version, which is
contained in - Daniel Dennett, Consciousness Explained, Penguin Books,
1993, page 399.
6 Daniel Dennett, Consciousness Explained, Penguin Books,
1993, page 401.
7 Robert Van Gulick, "Understanding the Phenomenal Mind:
Are We All Just Armadillos? Part 1: Phenomenal Knowledge and Explanatory
Gaps" in The Nature of Consciousness, edited by Ned Block, Owen Flanagan,
and Guven Guzeldere, Massachusetts Institute of Technology, 1997,
page 560.
8 Robert Van Gulick, "Understanding the
Phenomenal Mind: Are We All Just Armadillos? Part 1: Phenomenal Knowledge
and Explanatory Gaps" in The Nature of Consciousness, edited by Ned
Block, Owen Flanagan, and Guven Guzeldere, Massachusetts Institute
of Technology, 1997, page 560.
9 David Braddon-Mitchell and Frank Jackson, Philosophy
of Mind and Cognition, Blackwell Publishers, 1996, page 131.
10 This paper appears in The Nature of Consciousness,
edited by Ned Block, Owen Flanagan, and Guven Guzeldere, Massachusetts
Institute of Technology, 1997, page 571.
11 This is a somewhat shortened version of the story.
The full version is contained in Churchland's paper "Knowing Qualia:
A Reply to Jackson" in The Nature of Consciousness, edited by Ned
Block, Owen Flanagan, and Guven Guzeldere, Massachusetts Institute
of Technology, 1997, page 572.
|