Discussion Topic |
|
This thread has been locked |
yanqui
climber
Balcarce, Argentina
|
|
|
Dec 10, 2016 - 05:36pm PT
|
Too bad Rgold and Yanqui don't participate.
I'm checking in from time to time but I really need to finish up the real deal (an article I'm writing with my office mate) so I can get it sent off before I head out on summer vacation. Good luck with the mind thing!
|
|
MikeL
Social climber
Southern Arizona
|
|
|
Dec 10, 2016 - 06:17pm PT
|
Good luck with "the real deal."
|
|
MH2
Boulder climber
Andy Cairns
|
|
|
Dec 10, 2016 - 06:36pm PT
|
It isn't important if we reach a goal here, here, jgill, but it is fun to explore possibilities for progress. Thanks for your ideas.
I can see great difficulties in trying to match brain activity with what I understand some posters here to mean by experience. In one sense, experience is a seamless whole, or so it appears sometimes, and there might be objection to breaking it up into pieces, as in elements of a set.
If you were trying to align brain activity with so-called subjective experiences, how would you attempt to assess the subjective experiences? By asking for a running account from the subject?
Back in the 80s a European researcher put micro-electrodes into a nerve in the arm of volunteers. The nerve was known anatomically to connect sensory endings in the skin with the spinal cord and brain. By stimulating the whole nerve at a distance from the micro-electrode while using a micro-manipulator to move the micro-electrode he could isolate the signal from a single axon. Then he would stimulate the skin area innervated by the nerve until he had located the sensory receptor and determined the stimulus it was specialized for: heat, pressure, pain, etc. This could be done while the subject reported their sensations.
However, it would be far more difficult to keep track of the millions of sensory inputs to the brain, and the brain's own internal whirlpools of thought, if that were necessary to compare subjective experience with brain activity in real time.
Even if we had this:
Implants digitize thoughts and affect, routing them through the worldwide web. All knowledge is instantly accessible, every person a thought away.
How would you know that the thoughts you received were the ones that were sent? Would there be a translation problem? Everyone has a different history of life experiences, which may color their feelings about any particular experience.
But if it were possible, why stop with humans? Why not wire other creatures into your web?
|
|
jgill
Boulder climber
The high prairie of southern Colorado
|
|
|
Dec 10, 2016 - 08:20pm PT
|
From the staggering complexity you have described, I suspect any valid attempt to identify topological spaces and concurrent homeomorphisms would spin off into strange intellectual territory, like the following:
Probabilistic Topological Spaces
And that's just the simple objective side.
Tononi's Phi function (Integrated Information Theory) wanders around in a miasma of probability and is non-computable as well.
IIT addresses the mind-body problem by proposing an identity between phenomenological properties of experience and causal properties of physical systems: The conceptual structure specified by a complex of elements in a state is identical to its experience (Wiki)
Tononi has been working on this since 2004 and I don't think it has put a dent in the "Hard Problem." Too much for me.
|
|
MH2
Boulder climber
Andy Cairns
|
|
|
Dec 11, 2016 - 10:31am PT
|
When it comes to brain, we are still unable to look at large numbers of neurons while keeping track of each neuron's activity and its effect on the activity of other neurons.
Simulations that capture some of what neurons do have promise for answering questions relevant to perception, categorization, and memory.
However, it seems unlikely that there would be any point in trying to simulate human feelings in neural networks.
By accident I recently came across a curious sign of our times called a DARPA challenge. The 2015 challenge was to construct a human-like robot. It was not required that the robot should pass a Turing test or give care to nursing home residents. The successful robot should be able to:
1. Get into a standard human vehicle and drive it to a specified location.
2. Get out of the vehicle and travel across rubble.
3. Clear obstacles from a doorway.
4. Open the door, and enter the building.
5. Find a leaking pipe and close the associated valve.
6. Reconnect a hose or cable.
7. Climb a ladder.
8. Grab a tool from the site, break through a concrete wall and exit.
I like number 8, but I believe the 2015 DARPA Challenge found no robot able to perform the simple tasks above.
But if we ever do get a way to look at human brain activity with high resolution in space and time, it may still be possible to explore the connection between that brain activity and 1st person subjective experience. One approach might be to record the brain activity in a person whose behavior is considered to be closely connected to subjective feelings. Perhaps a composer, a conductor, a pianist, a dancer, an artist, a jazz musician. Then there might be a way to look for the sort of correspondences which jgill is talking about.
Of course, we might need to rely on good machines and software to do the analysis.
|
|
Largo
Sport climber
The Big Wide Open Face
|
|
|
Topic Author's Reply - Dec 11, 2016 - 11:27am PT
|
There are formidable challenges to trying to bridge the gap between 1st and 3rd person phenomenon, including the methods we use, seeming that most if not all known investigative modes derive exclusively from one or the other perspective, and use the criteria of a given perspective to evaluate success. What's more, the language and actual phenomenon are so qualitatively different that if there is a way to talk about both at the same time, that way is anything but clear. It only confuses the issue when one or the other camp insists that they ARE talking about both 1st and 3rd person phenomenon while camped in one or the other mode. Ergo inflation, or non-statements like "you only think 1st person experience exists." Bicker the point if you want, but it doesn't lead to new insight nor yet any breakthroughs.
Several obvious things to consider: No one can directly observe, from a 3rd perspective, the subjective life of anyone person. We can't read minds or feel what Ed is feeling or dream his dreams and understand his fears and desires and so forth. Not directly. And 3rd person inquiries are by design inquiries geared to frame external objects and phenomenon at mind-independent. So if we are trying to quantify or even "see" what is unseen, with a method geared to ignore the unseen "observer" and focus on the tangible, we are really up against it.
Even more difficult and the cause of all manner of confusion and conflation is the crucial difference between content and awareness, the crucial component is sentience and mind. 3rd person investigations can go far if not all the way in quantifying the process by which subjective content (feelings, thoughts, sensations, memories, etc) arises, but the mere process and existence of this content makes no matter unless a subject is aware of it happening. And the most crucial error made here (IMO) is that the default position is normally to posit awareness in terms of functionalism, a phenomenon that itself is task oriented and is entirely beholden to tasks - to doing this or that and responding to stimulus in determined ways. This is not at all true, but this can only be understood and made clear by doing some subjective adventures and empirically learning as much.
I think that one possible angle at starting to bridge the gap is tackle so called detection theory or signal detection theory - an attempt to quantify the ability to discern between information-bearing patterns - called stimulus in living organisms, and signal in machines. From a third person persepctive (note Ed's example of the driverless car) they look one and the same. But as you bore in and start comparing and contrasting, the stark differences start to shine through, and they are not issues of content or tasks at all, rather about awareness itself.
JL
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
|
Dec 11, 2016 - 01:20pm PT
|
We can't read minds or feel what Ed is feeling or dream his dreams and understand his fears and desires and so forth.
how do you know that "Ed" has feelings, dreams, fears, desires and so forth?
|
|
Largo
Sport climber
The Big Wide Open Face
|
|
|
Topic Author's Reply - Dec 11, 2016 - 01:49pm PT
|
From my 3rd person perspective, since there is no external object to measure per your direct experience, only the subject (you) can "know" in the way your are presenting this. But one wonders if there is such a thing as 3rd person knowing. Screwy question since there is no 3rd person, it's just what the 1st person does when discursively wrangling content, data, etc.
But how do You know you have impulses to ask that question and the experience of typing on the keyboard? Or do you "only think" you have the impulses, the self-awareness of same, and the experience of typing on the keyboard?
If you only think you do (experience = brain output), what criteria would you require to prove it was so - that you actually do have the impulses, the self awareness and the experience of typing?
CONFLATION WARNING! "What is REALLY going on is that the neuro substrate...." After all, what isn't physical? From the 3rd person perspective, what chance do you have of seeing anything BUT the physical, especially using a modality that seeks observer independent data? You're asking an apple to be an orange, and when it can't be one, instead of recognizing the limitations of a perspective, you cram it into square hole and say, Viola, THAT'S what IT is: Physical functioning. We just THINK we have experience.
Just note the switch from 1st to 3rd person, and the focus on content.
As I said, the machine model is a form of functionalism, built only to wrangle intake, processing, behavior and responses. Only when we start contrasting a sentient human being and a super duper Frankenstein AI "dood" do the difference hove into view, but it's not easy to do.
Basically the challenge, as mentioned, is to dive into detection theory or signal detection theory (information-bearing patterns), surveying the stimulus in living organisms, and the signal in machines, and the "awareness" involved, and only then do the differences start showing up.
JL
|
|
MH2
Boulder climber
Andy Cairns
|
|
|
Dec 11, 2016 - 05:05pm PT
|
As I said, the machine model is a form of functionalism, built only to wrangle intake, processing, behavior and responses. Only when we start contrasting a sentient human being and a super duper Frankenstein AI "dood" do the difference hove into view, but it's not easy to do.
Indeed.
What is it you propose to use as contrast in the sentient human being?
And where is this super duper Frankenstein AI "dood?"
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
|
Dec 11, 2016 - 05:58pm PT
|
Largo, was that an answer to my question?
maybe you could spend a little more time and make it a bit more understandable if it was...
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
|
Dec 11, 2016 - 06:17pm PT
|
maybe the premise of the story, the discovery of the "life force" and its use to "animate" the "creature" isn't very interesting...
...although it seems a concept very much alive in this thread...
|
|
MikeL
Social climber
Southern Arizona
|
|
|
Dec 11, 2016 - 07:46pm PT
|
MH2: it seems unlikely that there would be any point in trying to simulate human feelings in neural networks.
This would be an ignorant thing to say, especially for someone who knows about neuroscience research.
Maybe I don't understand what you wanted to say. I hope you are not saying that feelings are useless, irrelevant, or unimportant.
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
|
Dec 11, 2016 - 08:32pm PT
|
are there any female scientists posting to this thread?
probably shows better judgement that there are none.
|
|
jgill
Boulder climber
The high prairie of southern Colorado
|
|
|
Dec 11, 2016 - 08:40pm PT
|
Surely you science sausages . . .
Clearly an inability to imagine females as scientists, after all the efforts to recruit them into those disciplines. Women should probably read and write poetry - a higher calling.
;>)
Is it possible to dissect an experience? Even in a broad sense? An experience takes place over a span of time, and at each instant there is a kind of snapshot of the overall thing. Like Zeno's Arrow or freezing the film of a movie nothing occurs in each instant, although reintroducing the flow of time animates the process. An experience is like a function of many, many variables, some independent, others not so, F(x,y,z,...) where x=x(t), y=y(t), z=z(t), ...
More likely than not an experience is a continuous function, without dramatic changes from instant to instant. There are , however, what Rene Thom called "catastrophes" (Catastrophe Theory) where abrupt transitions do occur, or at least from our human perspective they seem abrupt. Is there any such thing in the real world as "completely abrupt", phenomena dramatically changing on the finest possible time scale?
It is easy to devise a mathematical function that does this. E.g., F(t)=0 if t<1 and F(t)=1 otherwise. However, reality may not work that way. If we were to freeze an experience at each instant (don’t ask me to be precise) and obtain a sequence of snapshots, could we call these slices “elements” of the experience? And if so, could we then begin to speculate on “open sets” of an experience? Could we then do a similar thing to the physical mechanisms of the experience, create a metric, and define open sets there? Then look for a homeomrphism?
Just BS ramblings.
|
|
jgill
Boulder climber
The high prairie of southern Colorado
|
|
|
Dec 11, 2016 - 08:43pm PT
|
Have you noticed that if you do a lengthy post, then try to make a few changes, you lose the bottom part?
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
|
Dec 11, 2016 - 11:29pm PT
|
you have to be careful using "greater than" and "less than" in your posts as they are interpreted as html tags and stripped out of the STForum server (only a subset of php tags are allowed).
Basically the "alleged" html tag and all the text after it are stripped.
|
|
MikeL
Social climber
Southern Arizona
|
|
|
Dec 12, 2016 - 07:32am PT
|
Jgill:
Hey, another reader of Thom!
Know anything about “path dependency” and “historical accidents?” (Think of Ashton Crutchner’s movie, The Butterfly Effect.) In my business, we think we see many industry developments / evolutions, as well as technological developments, that have turned on small, apparently insignificant events that could not have been foreseen or predicted. (Kinda like chaos theory, but not really.)
All these theories (highly mathematical) encourage us to see situations more like pure potentialities than linear cause-and-effect probabilities (although I suppose some folks will see potentials as probabilities).
What would an open system of open systems be like?
|
|
MH2
Boulder climber
Andy Cairns
|
|
|
Dec 12, 2016 - 11:42am PT
|
If we were to freeze an experience at each instant (don’t ask me to be precise) and obtain a sequence of snapshots, could we call these slices “elements” of the experience? And if so, could we then begin to speculate on “open sets” of an experience? Could we then do a similar thing to the physical mechanisms of the experience, create a metric, and define open sets there? Then look for a homeomorphism?
In a sense, this what correlative neurophysiology looks for. Nelson Kiang used micro-electrodes to isolate responses from single axons in the auditory nerve. Pure tones of differing frequency and amplitude were played into the ear and the responses of the axons were recorded in the form of spikes per second. By doing enough of these recordings, Kiang built up a picture of the information the auditory nerve carries once the ear has converted sound waves into nerve impulses. In addition to sine-wave tones, clicks can be presented to the ear, but in this early stage of processing in the nervous system nothing surprising happens when this approximation to a discontinuity occurs.
http://en.wikipedia.org/wiki/Nelson_Kiang
Further inside the brain, one may find abrupt transitions caused by small perturbations.
One of the most significant observations in the study of place cells during the past two decades is the discovery that place cells participate in multiple independent spatial representations. Under certain experimental conditions, place cells were found to totally alter their firing patterns in response to apparently minor changes in the sensory or motivational inputs to the hippocampus.
http://www.ncbi.nlm.nih.gov/pubmed/19021254
Sensitivity to apparently minor changes had been anticipated by earlier neuroscientists.
Even the simplest bit of behavior requires the integrated action of millions of neurons; the activity of any single neuron can have little influence on the whole, just as the path of an individual molecule of a gas has little influence on the gas pressure. It is questionable whether specific instances of behavior can ever be dealt with in terms of the activity of individual neurons; the complexity is too great. We shall probably have to use a different kind of model, a model which can be explained in principle by individual neuron action but which involves a somewhat different set of concepts and laws of action.
Why the Mind Is in the Head
Warren S. McCulloch
Even single neurons may also behave unpredictably because of smaller-scale events within them.
I think of the subconscious as a chemical soup that’s constantly making new combinations, and interesting combinations of ideas stick together, and eventually percolate up into full consciousness. That’s not too different from a biological population in which individuals fall in love and combine to produce new individuals. My guess is that all this activity takes place at a molecular level - like DNA and information storage in the immune system - not at the cellular level. That’s why the brain is so powerful, because that’s where the real information processing is, at a molecular level. The cellular level, that’s just the front end…
Meta Math! The Quest for Omega
Gregory Chaitin
It seems likely to me that any level of structure within an organism could be a site of information processing, and perhaps even information processing that affects the survival prospects and reproductive success of the organism, which would render that structure’s blueprint susceptible to selective forces in the environment outside the organism.
However, it also seems that none of our studies of however many variables on however many scales with howsoever sophisticated analyses would be good enough for Chalmers.
We have seen that there are systematic reasons why the usual methods of cognitive science and neuroscience fail to account for conscious experience. These are simply the wrong sort of methods: nothing that they give to us can yield an explanation.
Facing Up to the Problem of Consciousness
David J. Chalmers
[Published in the Journal of Consciousness Studies 2(3):200-219, 1995]
|
|
High Fructose Corn Spirit
Gym climber
|
|
|
Dec 12, 2016 - 11:51am PT
|
(1) "The central issue dividing the plant neurobiologists from their critics would appear to be this: Do capabilities such as intelligence, pain perception, learning, and memory require the existence of a brain, as the critics contend, or can they be detached from their neurobiological moorings?"
(2) "The question is as much philosophical as it is scientific, since the answer depends on how these terms get defined."
(3) "The proponents of plant intelligence argue that the traditional definitions of these terms are anthropocentric — a clever reply to the charges of anthropomorphism frequently thrown at them."
http://www.newyorker.com/magazine/2013/12/23/the-intelligent-plant
(4) "Descartes, who believed that only humans possessed self-consciousness, was unable to credit the idea that other animals could suffer from pain. So he dismissed their screams and howls as mere reflexes, as meaningless physiological noise."
...
“What we learned from Darwin is that competence precedes comprehension,” Dennett said when I called to talk to him about plant neurobiology. Upon a foundation of the simplest competences—such as the on-off switch in a computer, or the electrical and chemical signalling of a cell—can be built higher and higher competences until you wind up with something that looks very much like intelligence. “The idea that there is a bright line, with real comprehension and real minds on the far side of the chasm, and animals or plants on the other—that’s an archaic myth.” To say that higher competences such as intelligence, learning, and memory “mean nothing in the absence of brains” is, in Dennett’s view, “cerebrocentric.”
|
|
|
SuperTopo on the Web
|
Let us know!