Can Machines Dream? (Part Five)

John Avery continues the discussion

"I would like to add my own thoughts about dreaming: I think that one of the things that happens during dreaming is that data is transfered from a temporary mode of storage in the brain to a permanent mode. In this process, connections are made to previously stored associated data. Some of the associations that are tried and rejected are bizarre, and this, I believe, is why dreams often have a bizarre quality, while simultaneously including events stored in the temporary memory before sleep. Probably computers of the future will be able to reorganize their memories and to form associations in much the same way that the brain does. It is possible that computers of the future will also be capable of emotions."

Ron Burnett responds

I am not sure that we fully understand how the brain and the mind "store' memories. This is the difficulty with your comment. Gerald Edelman refers to the brain as part of a 'rich nervous system'. He does see memory as a process of categorization (something that computers could replicate) but then goes on to talk about the dynamic nature of neural populations and the continual process of change that characterizes the operations of the mind. This is important because if change is the main feature of thinking and remembering then it is likely that there is no fixed code for memory. The key question is whether computers could store ambiguous and continually evolving pathways of memory, where even the pathways themselves are open to change and transformation.

The other important question is whether it would be possible for computers to generate new properties, new ways of thinking and remembering. Our minds classify the world around us and then upset the applecart by dynamically combining complex properties gained from experience and the autonomous operations of the brain itself. Autonomy in this case means that the brain maps experiences and then recreates, transforms and recategorizes and then redistributes the entire flow in a way that we don't fully understand.

Further comment from Jan Visser

John's assumption falls in line with my own suggestion in my previous comment that dreaming has something to do with going beyond the immediacy of our perceptions and the processing thereof. This assumption seems to be supported by certain animal studies. It was already known that animals display REM sleep but it had not yet been established if this corresponds to actual dreaming. Recently, though, a 2001 MIT report of the Center for Learning and Memory states that “Animals have complex dreams and are able to retain and recall long sequences of events while they are asleep.? The same report places these findings against the scientific belief that “memories are formed in at least two stages: an initial process that occurs during the experience itself, followed by a consolidation period in which the experience is transformed into long-term memory. The hippocampus is believed to be involved in both stages.? As to the reason why animals dream, the findings in the report seem to suggest that “animals are capable of re-evaluating their experiences when they are not in the midst of them? and that such ability is not too far removed from the reason why humans dream.

According to Latrup’s remarks referred to by John, the emergence of dreaming would be a discontinuity in the evolution. If so, the quality of dreaming - the actual meaning of it from the point of view of conscious contemplation of the world - may equally be a discontinuity that distinguishes human beings from the other animals. However, evolutionary discontinuities do not represent changes of state from pitch black to brilliant white. There is a lot of gray in between and arguments may be construed (e.g. along the line of Michael Shermer’s “The science of good and evil?) to extend the concept of rights accorded to humans progressively to the animal kingdom and, why not, as machines become ever more lifelike, to machines. There would most certainly be an incredible benefit to the ecology of biological and physical resources if we became more conscious of such rights.

I am using the word ‘conscious’ in the above paragraph, thereby suggesting that a discussion of the capacity to dream in humans, animals and machines cannot be separated from questions about the meaning of consciousness in each of these categories, particularly moral consciousness. When I said earlier that “Perhaps we should [first] make machines enjoy the natural rhythm of their being? I was deliberately using language that pertains to the realm of consciousness. I am conscious and thus can enjoy (as well as suffer). The more I am conscious, the more I am able to enjoy and suffer, whence the traumatic experience of death in humans. Again, it’s not black and white and I am pervious to considerations about the suffering we inflict on animals other than ourselves as well as find it difficult to throw an outdated piece of computer equipment in the dustbin. I am also impressed by the display of humanlike emotions in primates when confronted with death, as for instance reported in the work of Frans de Waal of the Living Links Center.




Can machines dream? (Part 4)

Ronny Siebes of the Free University of Amsterdam continues the debate
* I think that we are both convinced about the limitations of the current way of doing science and especially the reductionistic approach.
* Also I agree with you that the mind is more than the physical brain itself. Now I remember again an insight that I had some time ago. In my viewpoint (I read it somewhere but do not remember the reference) the individual brain without communication is only a bunch of meat and blood. The brain *needs* input, and that input is culture (and nature).
Also culture *needs* humans and can be seen (metaphorically) as a collective mind. Therefore I would like to see our individual minds as the individual brains fed by collective input, and the collective mind is the collective input plus all the individual brains. Therefore only looking at the biological brain does not allow us to understand the way our individual 'minds'work.
If you want to do that, one needs to combine the insights from not only neuro-physiologists, but also sociologists, psychologists, antropologists and artists (and probably many more).
To make my point and to come back to your original question "Can machines dream", I can now, given the insight during our discussions, say the following:
* Dreams are events that happen in the physical brain, but can only occur when the brain is also a mind (meaning that it had input from outside itself). Therefore to understand dreams, it is not enough to understand the brain, but one also needs to understand culture.

* Currently machines are brains without (or with very limited) input, so therefore at the moment a machine cannot dream because its culture is not rich enough (or it is still not able to see/hear/feel human culture). The Internet (and especially the Semantic Web) will be the collective mind of the individual machines and also provide input to them. So, when the Internet becomes culturally rich enough, machines will be able to dream too.

Ron Burnett responds

Networks are representations of collective engagement and of community in all of its variations. Whether they are a collective mind is an intriguing question. Is a family with six members a collective mind? How would that collective mind be represented? Perhaps this discussion needs to move to questions of networks and what they mean.

Part Five… 


Can machines dream? (Part 3)

Quoting Ron Burnett
One of the metaphors we have been discussin is that the brain is like a computer and that human memory stores information much like a hard disk. There is simply not enough evidence to suggest that the metaphor works. So, machines cannot dream because among many other things, we don’t have an adequate definition of what the mind does when it dreams. All we have is the language of metaphor and description, a semantically rich space that cannot be reduced to any single or singular process.

Response from Ronny Siebes, Free University of Amsterdam
I agree with your statement that there is simply not enough evidence to suggest that the metaphor works. But that does not 'prove' that machines cannot dream. I would like to turn your argumentation into my favor. We are limited in two ways which has one cause namely we humans can only understand complexity by breaking events up into simpler things (reductionism) and we pay the price by losing an overview of the big picture (holism): first we do not know what dreaming is (complex event), second, we do not know what kinds of events can emerge from complex structures like the Internet. This limitation is the reason that we still don't know if dreaming can physically be expained and we still cannot find out whether complex structures like computer networks (e.g. the Internet) can dream.

Ron Burnett
I think that we are agreed that there is very little evidence, in the scientific sense for dreaming (although we all dream). Don't misunderstand me, I support and am excited by research into the brain. I do think that the Internet is one of the best examples that we have of complex networks that far exceed their original design in scale and effect. I am not sure given what you have said, how we can overcome the opposition between holistic approaches and reductionism.

Part Four…

Can machines dream? (Part 2)

Quoting Ron Burnett, Emily Carr University
Imaging of the brain can provides pictures of the connections between different parts, but imaging cannot provide details of what Gregory Bateson has so aptly described as the set of differences that make relations between the parts of the mind possible. “The interaction between parts of mind is triggered by difference, and difference is a non-substantial phenomenon not located in space or time… (Bateson, 1972: 92)

Ronny Siebes, Free University of Amsterdam
If I understand you correctly, you mean that our imaging techniques only allow us to make snap-shots of a fixed state of our neurons (by doing terrible animal experiments) or energy levels (PET, CT, MRI or EEG-scans). I agree with that.

Besides this I still believe (rationally) that every state change has a physical cause, and therefore a physical change in neuron-activity also has a physical cause. I'm not sure what Bateson and you exactly mean by 'difference', but allow me to give my definition by introducing another example: difference is like a comparison between a high and low pressure areas in the weather domain. We 'see' wind indirectly by experiencing that the trees are moving. Also when we make satellite pictures of the clouds and measure other things like pressure and temperature, we can see that there is a 'difference' between the values of these properties when comparing the high- and low pressure area.

However this is not enough to *explain* the reason (cause) for this difference. Physics only allows us to come up with theories (models) and when they describe reality precisely (by doing direct or indirect observational experiments) we assume that we know the reason for the difference. In other words, physics only provides detailed discriptions and models of the input and output behaviour of certain physical 'objects' that we experience in reality.

Like wind, 'thoughts' and 'dreams' and other mental utterances of our brain are, in contrast with snap-shots of mental activity, words used for pointing to eventsand processes. A dream is only a dream when there is something going on, meaning that a snap-shot is not enough to describe it because the process includes transitions between the states.

Again, physics is limited in this sense that it can only try to explain the difference between two snap-shots of our brain (where the difference could indicate that the person was dreaming), by giving a detailed descriptions of the state of the objects that can be seen on the snap-shots and come up with theories that caused the changes.

To conclude: Although I know that we humans are able to experience events like wind and dreams, we have to deal with a limited toolkit (science) that only allows us to look at snap-shots and come up with theories that explain the causal differences between those different states.

I agree that difference is a non-substantial phenomenon but therefore it can also not be investigated by science and therefore must be researched by another method. I do not mean that non-scientific investigations are less important than scientific ones, however I am personally limited to the use of rational (i.e. scientific) argumentation in a discussion about our brains. I am completely aware of my limitation :-) and also am convinced that science allows us to explain only a (perhaps very small) subset of the things we experience.

Response from Ron Burnett
I agree that science provides us with models and that inevitably there are limitations to what can be described. The distinction for me is between investigations of the brain and how we research, talk about and explain the mind. The brain is a physical, biological object. It is in the simplest sense, matter. The question is whether scientific research into the brain using more and more complex imaging technologies ends up creating metaphors that overwhelm the complexity of what the mind does.

Science reduces to idealize which is what I understand by modeling. Rationalism looks at cause(s) and effect(s). In these instances, (for the purposes of this debate) the danger is that the many elements that make up the human body, from homeopathic pathways to the immune system as well as the complex networks of interaction between neurons that constitute brain activity, will be reduced to function (alism). Reasoned argument is essential, but can a reductive argument work here?

So, is it the limitations of science itself that we are discussing? Or are we dealing with models and paradigms that tend to focus on what can be researched and from which extrapolations are made that lead in potentially dangerous directions?

A large measure of what we describe as intelligence is derived from our own, quite self-reflexive understanding of thought processes. We understand intelligence from a very subjective point of view. We know very little about how the electrical and chemical activity of the brain translates into intelligence. We do know that we are capable of incredible mental feats. For example, our use of language is just one of many activities we engage in for which we have a fragmentary understanding. There may well be a part of the brain, for example that deals with language, but as Edelman and Tononi point out, it is likely that the complex processing of information of this sort is distributed throughout the brain.

This means that it takes millions of interactions among neurons across networks connected in millions of predictable and unpredictable patterns and ways for a simple sentence to be formulated. Ironically, we can only hypothesize that the sentences so produced actually relate to the thought(s) we have had. (Ramachandran, & Blakeslee, 1998)

Part Three…

Can machines dream?

This series is in FIVE parts.

Ronny Siebes is a researcher at the Free University of Amsterdam. He and I met recently in The Hague and the ensuing email exchange represents only a small facet of the longer discussion that we had.

Ronny Siebes
I thought about the question you asked "Can machines dream" and have the following answer:

First, I would like to give my definition of what human dreaming is. Most humans know that they sometimes dream and may remember what they have dreamt, like the images, sounds or other impressions. Obviously, these things like pictures are not really there in the head because we don't have eyes in our head to look at them and if we had, it is too dark to see it (Dennet:). I'm not an expert on neuroscience but I guess that the brain works like this: images (encoded in a parallel bundle of light beams) that our eyes receive trigger a set of neurons that are responsible for interpretating visual input and these interpretations are stored in our memory. When we dream, parts of our memory become active and are manipulated by a script generated by fears, angers or other chemical impulses.

For this information to be remembered, the outcomes of these manipultion processes which are generated by the scripts are stored back again into our memory. Our consciousness (whatever that may be) walks through our memory and recognises that there is new information, namely the new stuff that was added by the dream process.

Computers are also able to receive, store and manipulate information from the outside world. For example, take a computer that has a web-cam connected to it and stores the bitstream on a hard disk or other kind of memory. It is easy to build a program that reads out the bits that represent the movie and to manipulate it. This manipulation would currently be very rude (for example just change some colors, or cut/copy- and paste some shots), but also very advanced like algorithms that detect scenarios and are able to replace objects by other objects. These manipulated movies can be stored again and after a while be 'played' (my free definition of becoming conscious) in a macromeda or windows media player.

Thus to summarize my point: if we describe human dreaming by its functional properties, we can apply it to its artificial counterpart.

Response by Ron Burnett

Imaging of the brain can provides pictures of the connections between different parts, but imaging cannot provide details of what Gregory Bateson has so aptly described as the set of differences that make relations between the parts of the mind possible. “The interaction between parts of mind is triggered by difference, and difference is a non-substantial phenomenon not located in space or time… (Bateson, 1972: 92)

Difference is not the product of processes in the brain. Thought cannot be located in one specific location; in fact difference means that the notion of location is all but impossible other than in the most general of senses. Bateson goes on to ask how parts interact to make mental processes possible. This is also a central concern in the work of Gerald Edelman, particularly in the book he co-authored with Giulio Tononi (2000) where they point out how the neurosciences have begun to seriously investigate consciousness as a scientific ‘subject.’ (3) Edelman and Tononi summarize the challenge in this way:

What we are trying to do is not just to understand how the behaviour or cognitive operations of another human being can be explained in terms of the working of his or her brain, however daunting that task may be. We are not just trying to connect a description of something out there with a more scientific description. Instead, we are trying to connect a description of something out there — the brain — with something in here — an experience, our own individual experience that is occurring to us as conscious observers. (11)

The disparities between the brain and conscious observation, between a sense of self and biological operations cannot be reduced to something objective, rather, the many layers of difference among all of the elements that make up thought can only be judged through the various strategies that we use to understand subjectivity. Edelman and Bateson try and disengage a series of cultural metaphors that cover up the complexity of consciousness.

One of these metaphors is that the brain is like a computer and that human memory stores information much like a hard disk. There is simply not enough evidence to suggest that the metaphor works. So, machines cannot dream because among many other things, we don't have an adequate definition of what the mind does when it dreams. All we have is the language of metaphor and description, a semantically rich space that cannot be reduced to any single or singular process.

Part Two…

Hypochondriac Culture (5)

What happens when you eat junk food? Although most diets in the United States and Canada are based on a variety of prepared and junk foods, the reality is that people continue to eat as if their actions will produce no effect. The contrast between the fetish for health and the disregard for nutrition is one of the central paradoxes of Hypochondria.  

Part Six

Hypochondriac Culture (3)

What if the hyponchondriac body is an aesthetic projection?

Lets for a moment assume that our daily experiences are continuously in a kind of flux between awareness and loss of awareness. We engage with the world around us without being fully aware of our intentions, often without understanding what motivates us to do certain things or react in specific ways to people and to objects.

When someone looks at us we take that look and project it onto our bodies and into our minds. This is not a mechanical process and has no particular sequence to it. Nevertheless, a particular look can lead to any number of thoughts and from those any number of different projections.

Now, lets reverse what I just said. What happens when the feelings you are expressing towards a friend for example, don't play out in the way that you anticipated? How does your body deal with the impact of that experience?

Another way of thinking about this is to reflect on the fact that our bodies represent and express our histories, both personal and public. Pain becomes an interface between the internal and external images that we have of our biological selves. Irrespective of whether that pain is real or not, our bodies express and represent our thoughts—the internal becomes visible.

If the pain is a fiction, the only way to make it real is to rescuplt the body, remake it in the light of the artifice, mark it with evidence, in other words, transform it into an aesthetic object.

Part Four



Hypochondriac Culture (2)

Hypochondria is an insidious disease because it is a silent and often invisible part of the suffering of so many people. It is centred on fear and misinterpretation. Hypochondriacs are constantly worried about a variety of symptoms that they read into their bodies. A minor pain is scripted into a major illness and leads to thoughts of death. Yet, to describe this process as a disease is perhaps a grave error. The psychology of fear is not easy to pin down. Since so much of medicine is concerned with cause and effect, the idea that someone could imagine an illness seems to be outside of the pragmatic medical context of searching for cures. Imagination is the key here as is a process called projection. It is easy to imagine any number of problems in the complex biology of the human body. It is easy (but the consequences can be dire) to project an external problem into an internal space. At one point in some "jottings" that were found among Sigmund Freud's papers, he made the comment that "space is a projection of the psyche". What if the hyponchondriac body is an aesthetic projection? I will respond to this question in greater detail tomorrow.

Part Three


Hypochondriac Culture (1)

What happens when the boundaries between the artificial and the natural break down? Jean Baudrillard said the following in his book on Simulation (p. 167) "To simulate is to feign to have what one hasn't." Baudrillard then extends this statement and talks about hypochondria. The hypochondriac generates all of the symptoms of disease while not having it. The result is that the hypochondriac becomes ill . Baudrillard goes on to say: "Psychology and medicine stop at this point, before a thereafter undiscoverable truth of the illness. For if any symptom can be 'produced', and can no longer be accepted as a fact of nature, then every illness may be considered as simulatable and simulated, and medicine loses its meaning since it only knows how to treat 'true' illnesses by their objective causes." (Sim 168) This artificially induced disease cannot be treated because it has no apparent cause.

The challenge of this metaphor is a profound one. What happens when the hypochondriac loses control of their simulated disease? Baudrillard suggests that the threat posed by the counterfeit, by reproduction, by any simulation process (Sim 182) is that it may take on a life of its own. Suddenly, control can be lost.

In a hypochondriac culture any number of claims can be made about reality. It matters little if those claims are true or not. What then happens is akin to what happens with disease. The claims take on a life of their own and it becomes very difficult to trace the origins of the claims.

Part Two