tiistai 30. joulukuuta 2008

Simo Hemilä

This text considers conscious experiences from an informational
point of
1. Based on sensory information and stored neural information my brain
produces an on-line model of my surroundings and my body. My conscious
bodily activities are based on this model.
2. The model-constructing brain activity also calls for conscious
experiences. From an informational standpoint the brain processes and the
corresponding conscious experiences are two representations of the same
3. Qualia of conscious experiences are coding tools needed to construct
the representation of the brain process.
4. As the conscious experiences are not entities of the physical world
and do not affect the materia, experiences like pain do not cause
behaviour, and animal evolution cannot affect CEs directly.
5. Mind may be defined as all the brain information contributing to
conscious experiences.


The material world consists of entities like particles, waves,
forces, and fields. Biological materials, including our brains, belong to
the material world. The "mental world" consists of conscious experiences
(CEs): sensations, feelings, thoughts, will, and self-consciousness. These
entities do not belong to the material world, they are something else. On
the basis of physics, chemistry or physiology we cannot describe
experiences of pain or colour. As mental entities do not belong to the
material world, they do not situate anywhere in the material world,
especially not in the brain.
According to a basic principle in physics all events in the
material world are caused by the four known interactions (gravitational,
electromagnetic, strong, and weak). This statement has very important
implications. For example, all bodily activities are caused by neural
stimulations or other material interactions. This leads to my first basic
assumption: Mental events do not cause muscle movements or other
bodily activities. However, it is generally assumed that a specific brain
activity somehow leads to CEs. I use the expression of Von Wright
(1998): "The brain calls for CE" . This leads to my second basic
assumption: In addition to physical interactions brains are able to
for conscious experiences. This "action" cannot be studied by
physiological methods and has no causal effects on the material world.
For many physicists and other scientists these two assumptions are
self-evident, but their implication are not. I consider the consequences
of these basic assumptions from the point of view of information. This
approach is quite fruitful. For example, the concept mind is quite
ambiguous. Webster's new world dictionary presents more than ten
meanings of this word, e.g. "the thinking and perceiving part of
consciousness". However, in order to explain the fact that your mind
today is nearly the same as your mind yesterday, you have to define
the mind as specific information, stored in brain overnight (see below).
I combine facts and hypotheses of many neurophysiologists and
other brain scientists, cognitive scientists, and psychologists in order to
construct a unified picture. This analysis is restricted to simple CEs like
sensory experiences and feelings of pleasure and fear, which we possibly
share with animals. The self-consciousness, introspection, and the central
role of speech in human conscience are not discussed.


The neuronal networks of the brain act like a computer, analysing
sensory inputs and combining them with relevant information stored in
brain "memory". Thereby, the brain continuously constructs models or
internal representations of the material world. The neurophysiologist
Yadin Dudai (1989) describes in some detail those internal
representations. According to him, most of the memory traces in the
neuronal networks are organised as such internal representations. The
evolutionary value of the models constructed by the brain is clear. Based
on such models humans and animals can predict events and act
The somatosensory map, and the maps of the visual field in the
cerebral cortex, are well known models of the material world. Complicated
information processing is necessary e.g. in the analysis of visual
information (see e. g. Marr, 1982, and Crick, 1990). The neural network
maps appear after a small delay, because it takes time to send the
information from the sense organs to the brain and to carry out necessary
information processing.
The tactile sense, vision, and hearing are the dominant sensory
modalities when the model of the human external world is constructed. My
brain also constructs a body image, a model of my own body, and here the
proprioseptive sense is most important, although vision and balance
organs of the inner ear contribute. Proprioseption is an elusive sensory
system based on signals from muscles, joints etc. Based on this sense I
know the positions of my limbs even when I do not see them.
When the motor neurones of the brain activate muscle movements or
other physiological activities, all information based on different sensory
modalities and previous experiences has to be taken into account.
Neurophysiologists are eager to know how different sensory maps and
other on-line information are connected in the brain. This is the binding
problem. Various hypotheses have been suggested
(see e. g. Crick, 1990), but the question is still open.
Conscious actions are based on a specific part of the currently
active information in the brain. That information is a special on-line
model of the external world including the body image. Because the
conscious actions are based on this world model in the neural network
I will call it the model for actions.
Our CE is also a model of the material world. The consciously
experienced external world model represents the material world so well
that in everyday life we generally identify the experienced world with the
external world, and the same applies for our body image. However, if the
sensory information is ambiguous or if the brain does not work properly,
the experienced world may be grossly distorted. Hallucinations, visual
illusions, phantom limbs, and many peculiarities in the experiences of
brain-damaged people are examples of such discrepancies.
According to psychophysical studies a conscious sensation often
appears a few hundred milliseconds after the actual event. For example,
conscious hearing appears up to 500 ms after the stimulus sound (Hari,
1995). Thus both the brain model of sensory information and the
corresponding conscious sensation have comparable delays. Moreover,
when a simple sensory stimulus is applied, the information content of the
response in the neural network is related to the information in the
corresponding conscious sensation (see e. g. Barlow, 1981). Generalising
these observations, I assume that the brain activity which leads to the
model for actions also somehow leads to a simultaneous CE. In such a
case, all information of that CE apparently comes from the simultaneous
model for actions in the neural network, as suggested by Chalmers (1995a,
b). The neural model for actions and the corresponding CE are two
representations of the same information, the same event in "an
information world". However, at each moment only a fraction of the
brain activities contributes to the model for actions and only the final
result of this information processing calls for CE. The brain information
associated with the model for actions and with CE is denoted apparent
Because a CE always contains information, a material system which
can call for such an experience must have data processing capabilities.
However, not all data processing systems call for CEs. In human brains
many advanced cognitive processes contribute to regulation of physio-
logical activities, although they are not apparent: they do not call for
CEs. For example, proprioseptive signals are essential in walking, but we
are not aware of them. Our visual system also includes blind sight and
other subsystems whose operation is normally not apparent to us (Wolfe,
1983). The same applies for hearing. Even a sleeping man may hear sound
signals and react to them. Some sounds of specific meaning are able to
wake him up (mentioning his name, suspicious sound from outside, slight
sounds of a sick baby). In psychology unconscious perception has been
studied extensively (for a review, see Merikle, 1998). These examples
demonstrate that a brain activity is apparent (calls for CEs) only if a
specific "process" is operating. I call this process a consciousness
evoking process. We do not know what such a consciousness evoking
process is. It could be a specific property or state of the neural network.
The information which is included in the model for actions is determined
by physiological processes, but because of this specific process the model
automatically calls for the CE. The question "Why conscious experiences
exist?" is as fruitless as the question "Why materia exists?".


Each CE has two aspects. First, it includes information. The infor-
mation in a visual or acoustic sensation is apparent, although the
quantitative amount of such information may be hard to measure. Also the
information of smell or taste sensations can be characterised. E. g., a
specific smell sensation has three types of information. Firstly, the
quality of that smell, specifying the source type. Secondly, the strength
of the smell. Thirdly, the property of the smell describing how unpleasant
or pleasant the experience is. Such information can be transmitted to other
Second, CEs include qualia, strictly private qualities of sensations
like the blueness of the sky, a specific pain, or a specific smell of rotten
eggs. Qualia are coding tools, just like the letters are coding tools of
spoken sounds. Such coding tools are needed to represent the neural
information of the model for actions in a perspicuous CE. Qualia do not
have information, as they are left over when all information is abstracted
from the CE; thus they cannot be described to other persons. Moreover, as
the only connection between the model for actions in the neuronal network
and the corresponding CE is information, I conclude that even my brain
cannot know the qualia of my CE. There is nothing to be known!
Here I have to point out that the quale defined this way is very
restricted. Let us consider a group of people hearing a symphony or
inspecting a painting. Such arts arouse different feelings in different
people, but most of the information in those feelings is qualitative and
hard to describe. However, because it is information, it can be transmitted,
possibly by poetic metaphors. So according to my definition the
informative aspects of those feelings are not qualia.
The term "coding tool" may be clarified. Let us assume that each
one in a group of students uses self-invented letters for writing diaries.
Any of them may read his or her diary to other students, but all of the
codes used by students are strictly private, just like the qualia of CEs. In
our communication the word "blue" is a synonym of subjective
experiences evoked by a range of short-wavelength lights without any
reference to the actual quale, which in principle could be different in
different people.
Animals and human beings are intentional, and thus in their models
for actions the value of a thing, an event or an effect is essential
information. Their neural networks classify the things and the events in the
axis very important - useful - neutral - harmful - dangerous - life
threatening. Specific chemical signals and tissue damages are easy to
detect and effective in revealing dangers. Therefore, the chemical senses
(smell and taste) and the mechanical senses (tactile sense) appeared early
in evolution.
Because in the model for actions the values carry important informa-
tion, they are also coded in the CEs of animals having a consciousness
evoking process. Important and useful objects are coded as good, pleasant,
enjoyable, while the harmful or dangerous objects are coded as bad,
unpleasant, disgusting. A very hungry animal could eat putrefying meat,
even when that could be poisonous. Thus a strongly negative classification
of putrefying meat has been evolved in the neural model for actions, and
the corresponding CE, the smell, is nauseous. Similarly, tissue damage
signals call for pains.
The intentional activities in the brain are also coded in the CE
e.g. as craving, hoping, wanting, deciding actions and executing decisions.
Pleasure seeking and craving are different reasons for conscious acts,
elicited by different neurotransmitter systems in the brain. However, many
goal-directed brain activities are not apparent and do not call for CE.
Current sensory information is not always needed for CE. When I
imagine or plan something or when I am day-dreaming, my brain activity
may be based solely on information stored in my brain.


The consciousness evoking process is a selective channel through
which information of the model for actions is transmitted to CEs. As a
useful metaphor, the consciousness evoking process is like a window in a
room through which you can look out. The apparent information
corresponds to the information visible through that window at each
moment. The accessible information corresponds to all information seen
through that window from different directions. The non-accessible
information corresponds to information of objects which cannot be seen
through that window (as the scenery on the other side of the house).
According to this classification current apparent brain information
and the memory traces of previous apparent brain information constitute
the accessible information. As time passes, memory traces deteriorate and
the corresponding accessible information becomes less accurate and may
be modified. The non-accessible information contains information which
never came to consciousness, and also the memory traces which cannot
call for CEs any more. For example, memory traces from the early
childhood or memories from extremely disconcerting or life-threatening
situations may become non-accessible.
The concept mind seems to be useful, as it exists in practically
all human cultures. Mind does not belong to the things of the physical
world, but neither is it just conscious experiences. Clearly your mind
today is nearly the same as it was yesterday, so somehow your mind is
preserved overnight. Based on information your mind may be defined as all
the accessible information in your brain. When activated, that information
becomes apparent and will call for CEs similar to those experienced when
the information was originally stored in the brain. The mind defined this
way exists during whole life. As it is the information of all the previous
and present conscious experiences, it "grows" with experiences from a
primitive mind of childhood to full maturity, and deteriorates during
Subconscious processes are usually assumed to be mental activities.
However, they refer to non-apparent, mainly non-accessible brain
information. Therefore, subconscious is not mental; it is that portion of
brain activity which affects the behaviour of an individual even when he
or she has little or no conscious perception of his motives.
Because the consciousness evoking process is not caused by physical
interactions, it cannot be studied physiologically. However, the physio-
logical correlate of this process is the brain mechanism which chooses
which information is included in the model for action, i.e., which
information in the brain becomes apparent. Finding this brain mechanism
is an important goal in neurophysiology, the great challenge of
neurophysiology. It could be a neuronal process, as suggested by Crick
(1990): ‘Where in brain are the "awareness neurones" and do they behave
in any special way?’
Vast amounts of accessible information are active and available in
brain at any particular moment but a brain mechanism called directed
attention chooses which sensory information and which information from
the brain "memory" are included in the model for actions and becomes
apparent. Because directed attention seems to be essential in CEs, the
electrophysiological experiments which seek to reveal which neurones are
involved in arousal are promising (see e.g. Livingstone & Hubel, 1981).
In our CE we generally feel that we have a full command on our
actions. However, in specific situations such a full command fails:
1. The delay between an event in the material world and the model
for action in the brain is not negligible. Thus in the fastest actions a
reflex reaction is needed. The reflex may be unconditional, as when pulling
the hand from a hot object, or it may be learned, as when returning an
approaching table tennis ball. These reactions do not prevent a feeling of
full command, as they are rational and to be expected in those situations.
2. Even in very complicated cognitive processes only the final
result usually appears in the CE. Why should I be conscious of all the
complicated processes leading to the synthesis of a visual sensation. On
the contrary, in everyday life we can treat the experienced external world
as being the true world. Similarly, when an engineer is planning
complicated equipments or when a physicist is trying to solve a difficult
problem, most of his brain activity is not apparent. Part of that activity
is accessible by introspection, but the most creative associations and
inferences are non-accessible. This is especially clear when the solution
appears after a delay of a night or after several days
3. Non-accessible information often contribute to our emotions,
decisions, and speeches. Usually we are not aware of that contribution; we
rationalise our behaviours. Sometimes the role of non-accessible
information is apparent, leading to confusion and annoyance, as the
reasons of behaviour cannot be understood. Typical examples are phobias,
addictions, compulsory disorders, outbursts of rage, automatic acts of
senile people, and forgotten childhood experiences.


Let us assume that while I am walking in the forest I come across a
strawberry. I see the berry, I remember its delicious taste, and I decide to
take it and eat it. These are mental events, and indeed I eat the berry
because I decided to do so. On the other hand, my optic nerve transmits
visual information from my eyes to my brain, where that information is
combined to the memories and analysed in my neural networks.
Especially, the memory circuits tell to the so called "reward centre" of my
brain that the berry has high nutritional value, that it is valuable food.
Thus my brain sends signals to my hand muscles to take the berry and bring
it to my mouth.
It seems that this event has two causes. Von Wright (1998) has
presented a philosophical analysis of this problem. In the material world a
sequence of events is caused by material interactions. Thus the brain
activity is the physical cause of my berry picking. However, this
brain activity also calls for CEs, and each CE is another aspect of the
simultaneous material event. The visual sensation of the berry and the
conscious memories of that berry lead to my decision to eat that berry.
This decision is the reason to the action of eating the berry. So the
sequences of events in my brain and in my CE are two representations or
view points of the same event.
As CEs cannot affect the events in the material world, the pains
and aches seem to be unnecessary nuisances! Indeed, strong tissue damage
signals may lead to fast unconscious avoidance reflexes, and the pain
comes slightly later. Moreover, in a "fight or flight" situation tissue
damages may occur without experienced pains, because of physiological
inhibition of the pain signals. However, in normal circumstances the
information of a tissue damage or a danger of tissue damage should be
integrated with other apparent information in order to achieve optimal
behaviour. After all, we accept some tissue damage if it is important! Thus
in normal situations the tissue damage information is an essential part of
the model for actions and automatically calls for pain sensation. When
aboy burns his finger in the candle light and learns to avoid the flame, he
learns because of the information of tissue damage in his brain. However,
as the model in the neuronal network and the CE are the same event in
"the information world", we cannot say that in normal situations pain is
unnecessary. Using the terminology of Von Wright, the tissue damage
signal in the brain is the cause of the avoidance activities of the muscles,
and pain is the reason for the conscious act of avoidance.
The pains of the phantom limbs (R. Melzack,1992), the pains of
migraine and other headaches, and many chronic pains are unnecessary in
the sense that in those cases the information of tissue damage is erroneous.
Even when the tissue damage is real, the tissue damage signals are useless
if the damage cannot be avoided (childbirths, tortures, a chronic pain after
many types of serious tissue damages). In some cases such useless pains
can be eliminated by preventing the peripheral tissue damage signals to
reach the cerebral cortex. However, if e.g. the "model of damaged limb" is
created by cortical activities and this model calls for an experience of a
painful phantom limb, such CE may be hard to be eliminated. These are
examples of imperfect activity of brain circuits. Similar erroneous
sensation in the modality of hearing is the tinnitus.
The brain saves both apparent and non-accessible information, but
only apparent information calls for CEs. When we recall an old event,
parts of the old model for actions are recreated. Thus it is to be expected
that remembering an event which originally called for a CE again calls
for the same CE. If the original event left non-accessible information into
the brain, the brain may retrieve memories of that event but they do not
call for CE. Thus we can recollect only the things and events which we
have earlier experienced consciously.
Because CEs do not affect the events in material world, evolution
does not affect CEs directly. However, the consciousness evoking
process has a structural basis in the brain, and an effective model for
actions is a very important goal in evolution. When the brain structures
which create this brain model evolve, the models for actions improve
leading indirectly to the evolution of CEs.
According to Antonio Damasio (2001) it is important to distinguish
feelings and emotions: according to him feelings are conscious
experiences and emotions are physiological correlates of feelings.
However, if the brain information and the corresponding information of
the experience are identical, such distinction is usually useless. Let me
use an analogy from computers. I have a file "SUMMER03" both on my
computer desk top and in my hard disk. Because the information in those
files is identical, I consider the files identical and call them both by
the same name. Only in operations like copying or deleting files I have to
specify e.g. "SUMMER03 in the hard disk". When I talk about my
feelings, thoughts, decisions etc., my speech always applies both to the
information in the model for actions and to the identical information in
the CE. Only in the fields of neurophysiology, cognitive science, or psycho-
logy it is sometimes necessary to specify which information is considered.
On the basis of reactions and facial expressions I may also talk
about the feelings, thoughts and decisions of an animal. If the animal does
not have conscious experiences, my talk applies to the brain information
only. If the animal has conscious experiences, my talk again applies both
to the brain information and the information of the experience. However,
because the experiences are strictly private, even in those animals my talk
is based on the behaviour of that animal.


Barlow, H. B. (1981). Critical limiting factors in the design of the eye
and visual cortex. Proc. R. Soc. Lond. B 212, 1-34.
Chalmers, David J. (1995a). Facing up to the problem of consciousness.
Journal of Consciousness Studies 2, 200-19.
Chalmers, David J. (1995b). The puzzle of conscious experience.
Sci. Am. 273, 62-8.
Crick, Francis (1990). The astonishing hypothesis.
The scientific search for the soul. London: Simon & Schuster.
Dudai, Yadin (1989). The neurobiology of memory: concepts, findings,
trends. Oxford: Oxford University Press.
Damasio, Antonio (2001). Fundamental feelings. Nature 413, 761.
Hari, Riitta (1995). Illusory directional hearing in humans.
Neurosci. Letters. 189, 29-30.
Livingstone, Margaret S. & Hubel, David H. (1981). Effects of sleep and
arousal on the processing of visual information in the cat. Nature 291, 554-61.
Marr, David (1982). Vision. San Francisco: Freeman and Co.
Melzack, Ronald (1992). Phantom limbs. Sci. Am. 266, 90-96.
Merikle, Philip M. (1998). Psychological investigations of unconscious
perception. Journal of Consciousness Studies 5, 5-18.
Von Wright, Georg Henrik (1998). In the shadow of Descartes.
Dordrecht, Boston,and London: Kluwer Academic Publishers.
Von Wright, Georg Henrik (1974). Causality and determinism. New York
and London: Columbia University Press.
Wolfe, Jeremy M. (1983). Hidden visual processes. Sci. Am. 248, 72-85.


Tietoja minusta

D. Techn. Docent in biophysics, retired