FSMN of NZ Article - Experiencing Greenness

Friends of the Scientific and Medical Network
Auckland, New Zealand

FSMN Articles

Experiencing Greenness - an Investigation into Life & Robots

Record of points raised at FSMN meetings in March and April 1999

Friends of the Scientific and Medical Network - March 1999

EXPERIENCING GREENNESS

A context for sensation and perception. Attention / Awareness / Memory <-- 3 integrated subsystems. From awareness comes cognition / emotion / volition <-- 3 capabilities. [note: picture of person] Conscious self (mind?) Csikszentmihalyi (c1991) from "Music Matters: A new Philosophy of Music Education", David J Elliott, OUP 1995 p 52.

Cognition means knowing in the widest sense of the term and includes all processes involved in the verbal and non-verbal organisation, retrieval, use and application of our apprehensions. Cognition refers to the various processes by which we recognise, relate, and deploy information from inside and outside ourselves. Information includes all the differentiated sights and sounds, all recognized thoughts and emotions, all the situations and events, that we encounter.

Human experience of greenness is an evolving dynamic.

Continuum of attention-awareness-memory

Humans receive sensations --> interpret them --> have emotional response. I agree.

ROBOT/COMPUTER AWARENESS

We underestimate the degree of complexity of a human.

The Robot is possible only since the physics of colour has become known since Newton published his book Optics 1703. Prior to this demonstration that colours are derived from white light and that colours are part of the spectrum colours that can be measured in terms of electric phenomena.

DIFFERENCES BETWEEN HUMAN AND ROBOT/COMPUTER

The human is capable of experiencing a whole that is greater than the sum of the parts of the data whereas a computerised robot only has the sum of the data. A computer can only deduce, that is retrieve information that is implicit in the data.

I do not believe the robot has an awareness of self.

I believe human experience is more complex than the circuitry of robots because there is an emotional response (like/dislike, safety/terror) based on prior experience. Therefore baby's experience is like the robot's, growing, evolving. Yes. Robots can respond only from prior "experience" but humans can create new responses. I agree.

Recognising greenness. The robot receives signals that enables its circuitry to acknowledge "there is greenness in this field". In the human machine, a part of the brain receives the signals which enable greenness to be identified, as in the robot, but then at a higher level of neural circuitry is the recognition within the brain that it knows it is experiencing greenness as a category. A human brain knows that it knows. A computer just knows.

Colour - qualitatitive difference between human perception vs computers.

Once you step beyond the literal simplicity of the hypothesis the human has consciousness that the machine cannot aspire to.

Qualitative difference ... Robot more reliable. Not deceptive.

The robot cannot ever mimic the human perception of green, because along with the biophysical analysis of the green signal, a sense that green involves a sense of security and safety. Why? Because green is associated with having plants, which through photosynthesis created the oxygen which supports our life, plus the food which supports the entire food chain. THe robot cannot ever be expected, as an inanimate creation, to possess understanding. In the same way we perceive red, the colour of blood, with danger. Could a machine be expected to have this sense.

There may be some source of input to the mind of man that may offer info that could not be deduced from any knowledge within the experience of the robot, however large.

Cognition, Volition and Emotion. There is a huge qualitative difference between a robot's response to green and a human's response. A very complex robot may be able to report many previous connections in its experience of green based on its previous experience (or programming). But the human can not only do this but also through poetic and artistic creativity actually create new awareness which were not in existence before, e.g. a sonnet about greenness, a painting about greenness.

[Note: original Einstein statement appears to be missing.]

Einstein's picture: If I was an extremely complex robot programmed with information about Einstein I could do all sorts of clever analysis and description of his life and his work. But as a creative human being I can make a feeling response to Einstein's picture - I could write a poem about his contribution and my response. In this way I can evaluate and express the significance of Einstein's life and work that a computer cannot do.

Einstein: interpretation is added by humans. More than the sum of deductive data. A 2-year old would read the data as "Grandad!" - a human response based on interpretation, based on less experience than necessary for accuracy, but still human! Actually a holistic response.

Doesn't there need to be a human observer to confirm that the robot has registered green..... then I read Edward O Wilson's Consilience which said every mental scenario we create in our heads is a virtual reality...... then I thought about the paper Einstein and realised he was dead and no longer "real" in any sense.

SIMILARITIES BETWEEN HUMAN AND ROBOT/COMPUTER

I am happy to accept that a robot perceives green in the same way that 'another' person does.

There is no evidence that another person (as distinct from robot) has any actual awareness of green - other than that they say they do. What if the robot says "I did actually experience green"? It may be argued that it was just programmed to do this to which I reply that for all we know say may a person.

As a simple hypothesis the result is the same. But we are already answering the result from a human standpoint. Hence the machine is at a disadvantage as it is being judged by humans.

Who are we to judge the consciousness of a complex robot, who might be more intelligent or capable than us.

To the comment that a robot might say to a human critic "but I also know what greenness is" I commented that a robot that could make such a reply to me (a human critic) would be a very much more complicated device than the simple robot first described, which simply observed a green something and registered its colour as "green".

MIND READING AS HYPOTHETICAL SOLUTION

A hypothetical possibility is that some human can read human minds. If so they could be aware that another human is aware of green. The same mind reader might also be able to read the "mind" of a robot and know whether it was aware of green. If the mind reader cannot read the mind of the robot then this may or may not indicate the robot is unaware of green.

In response to the hypothetical possibility of a person mind-reading another person and a robot to see whether they experience green I suggest that to be fair a hypothetical mind-reading robot should also be used.

Mind reading ... read person ... simple robot and complex robot both have the same ability.

I would have raised the PEAR anomalous experiments at Princeton where there is measurable interaction between humans and random digital machines. Intention in humans --> action in machines. Doesn't this raise the issue of machine consciousness i.e. a receiver of human consciousness.

SOURCE OF CREATIVITY IN ROBOT/COMPUTER BEHAVIOUR

Creator (of a computer) places his creativity in his creation.

Robot experiencing colour green. The robot's apparent "consciousness" is totally subjugated to human consciousness that created it.

Robots the creation of the modern technological society. No single person has the ability to create a robot. Robots extend human capacities. e.g. can receive radio signals which are not received by humans. Bridge requires a lot of learning but a computer can play all 4 hands in a game of bridge.

There exists software that has created a new, previously unknown, mathematical theory. [Note "proof" rather than "theory"]

In response to the suggestion that a computer or robot only has the responses of its programmer, I mention that learning programs have been around for over 20 years. The program "animal" learns to ask questions and guess what animal a person is thinking of. Initially the programmer tells it almost nothing but it gradually learns about every animal that people think of, whether or not the programmer knew of it.

OTHER

Complexity doesn't necessarily mean consciousness.

"There is no fundamental difference between a simple and a complex robot."

Construction of reality.

Meaning of greenness.

Friends of the Scientific and Medical Network - April 1999

Recorders summary: The prime question is whether a robot or computer might be able to tell us it sees green or recognises something as a flower, and that being so, whether it is actually conscious.

Many feel that human behaviour has too many aspects which are beyond a computer, mainly in the areas of creativity, affection or feeling and self awareness but also in the degree of complexity which has been achieved over vast aeons of evolution. A very simple world model was advanced which considered the mind as a process of matching the observed world to the desired state and attempted to make changes and this would lead to possible "emotions". In "judging" machines we must avoid a too human perspective.

The question then must be asked as to whether there are special laws of physics for humans as distinct from machines? No-one ventured the opinion that there were but the possibility of present physics being incomplete was advanced. Generally any such incompleteness might be expected to be related to the mystery of life and so explain the "extra ingredient" that was often considered to be present in people. One possibility in this regard is the whole question of parapsychological phenomena which, if they exist, are an area where accepted science is certainly incomplete or incompletely understood.

If the universe is considered to be based on physical laws then it is not unreasonable to apply Godel's theorem to it and to conclude that there will be true theorems that cannot be proved, and propositions whose truth is undecidable and so any attempt at a complete understanding of the universe is impossible. In general we must consider any part of the universe that we study as an open system and so Godel's theorem does not strictly apply although this does not help with the complete understanding.

It was argued that a computer or robot was a tool or an extension of the programmer just like a crane. Counter examples were given of programs going beyond the programmers capability or knowledge in, for example, playing chess or developing new maths theorems. This touches on the area of the spark of creativity argument that there is something special in people.

Whereas May's meeting was generally of the opinion that computers were fundamentally different to people, but April's meeting softened a bit with more acceptance of the idea that the difference might be a large one of degree rather than of kind.

EXPERIENCING GREENNESS

ROBOT/COMPUTER AWARENESS

DEGREE OF COMPLEXITY OF PROBLEM

We underestimate the degree of complexity of a human.

Agreed that aeons of evolution have produced incredible complexity and fitness, and that Bach or Cezanne cannot at present be produced by computers/robots, but considerable results do exist and are being extended. Thinking about the Turing test, even some of the early "play" psychoanalyst programs were thought by people to be very compassionate, understanding and interested in them.

Computers are at present much less complex and intelligent than a person, but this does not mean "in principle" that a computer cannot do something such as think or whatever. Unless the laws of physics are especially different for humans then it must "in principle" be possible for machines to achieve thinking and consciousness.

In principle, we might build a robot which incorporated the aeons of evolutionary experience to date, but only if we understood ourselves completely.

I believe the same physics underpins both computers and humans, but we cannot explain life in terms of our present physics because the latter is incomplete. (Later: I do not believe consciousness is an emergent property of matter but an intrinsic property which our present physics does not contain. We might say present physics overlooks, amongst other things, the element of 'spirit' which I perceive as intrinsic to the cosmos. I am not in this last statement wishing to imply anything about what might lie beyond our cosmos.)

I can accept that, in principle, the difference between a robot and a human may be one of complexity. We might be able to build a robot which appeared to be human but only if we understood ourselves a great deal better than we do now. And we would still not know if it possessed self awareness. (Later: To repeat, the essential problem is our inability to say what consciousness actually is. While correlates between brain processes and the activities of mind are known, (by MRI and PET scans for example), this does not mean we can equate brain and mind. We do not know the seat or dimensions of mind. Perhaps our robot would have to incorporate some biological material!)

DIFFERENCES BETWEEN HUMAN AND ROBOT/COMPUTER

I do not believe the robot has an awareness of self.

Colour - qualitative difference between human perception vs computers.

Once you step beyond the literal simplicity of the hypothesis the human has consciousness that the machine cannot aspire to.

Qualitative difference ... Robot more reliable. Not deceptive.

There may be some source of input to the mind of man that may offer info that could not be deduced from any knowledge within the experience of the robot, however large.

[Note: original Einstein statement appears to be missing.]

Doesn't there need to be a human observer to confirm that the robot has registered green ... then I read Edward O Wilson's "Consilience" which said every mental scenario we create in our heads is a virtual reality ... then I thought about the paper Einstein and realised he was dead and no longer "real" in any sense.

SIMILARITIES OF HUMAN AND ROBOT/COMPUTER

I am happy to accept that a robot perceives green in the same way that 'another' person does.

As a simple hypothesis the result is the same. But we are already answering the result from a human standpoint. Hence the machine is at a disadvantage as it is being judged by humans.

Who are we to judge the consciousness of a complex robot, who might be more intelligent or capable than us.

The reason that a person thinks that a person is conscious and a computer is not is that they themselves are a person.

GODEL'S THEOREM The statements "There is no difference in principle between a simple robot that sees a green scene and responds 'green', and a much more complex robot with bumpers front and back etc" and "A robot has devised a theorem which a group of experts called a creative act", make an implicit claim that logic and mathematics can reproduce lifelike behaviour.

The assumption that mathematics can provide an unambiguous logical framework for a robot (a mechanism) received a shattering setback in 1931 with the demonstration by Kurt Godel that within any formal logical system there are true theorems that cannot be proved, and propositions whose truth is undecidable.

Godel's method of proving his theorems involved the same idea of operating on two levels which arose at the beginning of our March meeting in response to the opening proposition. The validity of his work was accepted without question by his fellow mathematicians but provoked controversy among philosophers. The upper layers of meta-mathematics had been introduced by the grand old man of mathematics, David Hilbert, as part of his protracted efforts to establish for mathematics a rigorously self-consistent logical structure. Hilbert's motivation was a strong philosophical scepticism but his framework landed him in a similar problem to that which plagued logical positivists; as Poincare put it "Are all the theorem filling volumes of textbooks only indirect ways of saying A is A?" Thus Hilbert's dilemma: how to avoid trivialising mathematics. Working within Hilbert's framework, Godel showed that Hilbert's goal is unattainable. The objections some philosophers raised against Godel concern this philosophical framework. They are wrapped up with the question "Are mathematical theorems invented or discovered?" For instance, was it true that the square on the hypotenuse equals the sum of the squares on the other two sides before Pythagoras' theorem was enunciated? To believe so is to believe in metaphysics and as one of the objectors said "Mathematics has always skirted dangerously close to the shores of metaphysics ... for all the vaunted positivism of the modern scientific age, atavistic metaphysical instincts are never far beneath the surface."

Godel is defeated by having the computer develop approximations to the world and to note that the computer becomes an open system when external input is added. Emotion = f(needs, current context) Current context = f(sensory input, past experience) Affect = f(emotion) --> could imagine a robot expressing emotionn and affect.

MIND READING AND THE SOLUTION

Mind reading ... read person ... simple robot and complex robot both have the same ability.

And a postscript of a different kind: We did not consider at all the implications of the many parapsychological phenomena for which there is good evidence but which are discounted as nonsense by mainstream science. In general such phenomena seem to imply that minds ( and maybe mind/matter) can interact by means other than the normal senses. Should we expect to resolve the AI question while these phenomena are excluded?

SOURCE OF CREATIVITY IN ROBOT/COMPUTER BEHAVIOUR

Creator (of a computer) places his creativity in his creation.

A chess playing computer is really an absent chess playing computer programmer. We are fooled into thinking that it is the computer that is playing the human chess champion.

Robot experiencing colour green. The robot's apparent "consciousness" is totally subjugated to human consciousness that created it.

There exists software that has created a new, previously unknown, mathematical theory. [Note "proof" rather than "theory"]

I disagree that a computer is, like a crane, an extension of the person who created or programmed it. Are Adam and Eve in control of everything that happened since? The reason is that open systems mean that new interactions produce an entity that goes beyond the original creation. As a matter of fact, chess playing computers are far better chess players than the programmers.

POSSIBLE WORLD MODEL PERSON OR ROBOT

world model ----->
sensory input ----> content -----> emotions - - - - if frustration > 90% do a crazy random act
needs/desire---->
e.g. vote alliance

Responding to the model which was described diagrammatically by circles labeled world model, sensory, context, needs(desires), and able to generate an output labeled emotions, I said: This is a model whose operation we can examine only from the outside. We are not inside it. We don't know how to get inside. (Later reflection: To get inside we have somehow to share the device's consciousness, if any.)

POSSIBLE EXTRA INGREDIENT IN PEOPLE

Human sensory perception is more than the sum of all the physical senses and processing power. The synergy comes from genetic development over aeons, the notion of feelings, et al.

Is the fact that a human experiencing a flower has an accumulated memory of experience over 15 billion years of interaction with the world, a factor in considering the difference between a human observer and a computer observer?

What is the difference between a computer recognising a flower and a human experiencing a flower? What is the qualitative difference between perception by a robot and perception by a person? The camera equipped computer can recognise "flower" as a stereotypical category. However I doubt it could recognise an insect shape of an orchid's flower which is a clever enough mimicry to fool male insects to attempt to mate with it. A botanist can use other conceptual and observational evidence to create the category "flower" as the reproductive structure of an angiosperm plant and could even use these abstract categories if looking at "plants" on another planet. Can the robot perceive these categories? Could the robot ever be built that would have the manual dexterity to investigate the environment as well as the brainpower to think about it?

Psychotherapists think of the mind in terms of cognition and affect (emotion). Machines can develop cognition but emotional responses are complex, non-rational, even spiritual.

Comparison of computer and human in relation to flower illustration: An artist can see a flower, look at it selectively, feel emotional responses and use this experience for creative potential in such a way that a work of art ( eg a Cezanne painting) can be produced. A computer can capture the image of a flower, identify it with the symbol "flower" and combine elements of the image in patterns but until it can produce a significant art work with the emotive and creative properties of satisfactory human art, it is qualitatively different. Note: To identify a feeling is different in quality from actually feeling a feeling.

Machines don't produce artistic creation like Beethoven or Cezanne. There is a spark of creativity, stirring emotion.

Billions of years of development built into our memory and genes.

Human children raised by wolves develop wolf thinking not the full human mind. Therefore we need human interaction to develop human awareness. Babies die if they don't get adequate stimulation and loving.

Training of counselors shows the difference between a computer response " see you're upset" and a human response which genuinely connects with the feeling of the other person.

Can machines laugh? Humour is a spontaneous reaction to something seen as unexpectedly absurd or highly pleasurable. Even babies laugh.

Why do I feel horror at the thought of living in a society of robots? Excluding psychopathic personalties, there is some kind of soul-spark that connects between human beings.

OTHER

Complexity doesn't necessarily mean consciousness.
"There is no fundamental difference between a simple and a complex robot."
Construction of reality.
Meaning of greenness.