Picture of the author
Topic :
Phenomenal Robot Design.
Thread Created at Nov 18th 2011, 10:06:55 am | Started by Brent_Allsop
Number of Post in this thread: 2
Please Sign In to comment on this Thread
Brent_Allsop replied 12 years ago (Dec 13th 2011, 9:44:31 am)
Theoreticians, [http://canonizer.com/support_list.asp?nick_name_id=243 James Carroll] and I have been continuing our productive debate between Material and Functional property dualism on the Mormon Transhumanist e-mail list. As usual, I'm having a hard time fully understanding competing theories to my own, and I sense James having the same problem with my theory. So I figured I could try to describe to him how I might create or simulate a "Functionalist Property Dualism" world within a basement world that adheres to the rules of Material Property Dualism. I noticed that imagining such simulations, that operate within a basement world that behaves according to the rules predicted by Material Property Dualism, is the way I try to think of other theories, so I thought this might be a good way to improve communication between two people like us with very different working theories. So, James, let's start in a basement world that follows the rules of Material Property Dualism, and within that world, let me attempt to create a simulation of a world that follows Functional Property Dualism rules, as I currently believe you imagine them to be. We want to do what would be required to build a similar phenomenal strawberry picking device to the one described above. But, since this one would be in a simulated FPD world, we want to describe the differences. The previous robot had a 1K x 1K x 1K 3D array of neuron like structures which, if these neurons fired Glutamate in their downstream synapses, it would experience a redness quale at that voxel location, while if it fired Aspartate, it would experience greenness at that voxel location, since it was the Glutamate and Aspartate that had these redness and greenness phenomenal properties. However, it is argued that in the FPD world, a string of bits could produce a redness quale experience. So let's say that we design this simulation to have, instead of a 3D array of neurons firing either glutamate or Aspartate, there are simply abstract string arrays that can contain a simple string value formatted as: "Magic:0xRRGGBB", where RR is a hex value from 00 to FF representing the level of redness color, and GG, and BB are the same for greenness and blueness. James admits he's not sure how any set of ones and zeros could result in an experience of a redness quale, he just claims he has proven it must be so. But in a simulated world you can alter the rules any way you want to make any theoretically possible reality. So that is exactly what we wire into this simulated FPD world. Since this simulated FPD world is running within a MPD world, it follows the assumption that you can't get something from nothing. In other words, just as you can't store information, without having something physical to represent it with, you can't have a phenomenal property, without something that has the phenomenal property. So, in order to have something that can experience a redness quale in the simulated FPD world, some additional stuff needs to be added to the simulation, to match up a string like "Magic:0xFF0000" that exists in the simulation, and wire in a real set of neurons in the basement world that fire Glutamate resulting in the redness quale experience that we need to exist in this FPD world. We do it in a way so that no matter what is representing the string "Magic:0xFF0000", whether it is simulated silicon, or simulated neurons, or even bear cans, the simulation hardware detects this "Magic:0xFF0000" string, and harness a set of neurons firing glutamate in the basement world, producing a true redness quale experience in the simulated world. Now, let's say there are two brains running in this functional property dualism world. They both propose two theories as to the way their world is operating and what is responsible for their phenomenal redness and greenness experiences. One proposes Material Property Dualism – predicting the only way to get redness, is to have the right matter. He predicts that silicone can't have a redness property. The other is predicting functional property dualism, that anything that has the correct set of ones and zeros, will result in the redness quale experience they are both capable of experiencing. So, they take their instruments, and start probing the makeup of these simulated brains and examining just how it is they work. Of course as predicted by the functional property dualists, they discover that any time they see the abstract string of "magic:0xFF0000", the brain experiences a point of redness qualia experience. This happens no matter what part of the brain it is working on, and no matter what that part of the brain is representing that string with – anything from a paper tape with the right bits on it, to a set of magnetic characters represented by bits of properly oriented ferrous material, and so on – will all produce a spot of redness quale experience (thanks to the magic wired into this simulated world). So, the brain predicting Functional Property Dualism shows to the other brain, this objective, demonstrate-able to both discovery, and says: I told you so, our world is a functional property dualist world. And the Material Property Dualist brain admits, yes, you are definitely right, I was wrong. I thought it would require effing magic, to achieve this, but low and behold, that is just what we see in reality, apparent effing magic, for any time we have a string like "Magic:0xFF0000", no matter what is being interpreted as representing that string. Who, but you, could have thought such? And the mistaken brain then abandon's the material property dualist camp, and starts supporting the functional property dualist's camp, and very much enjoys finally effing all his experiences with everyone else's brain, in their newly enhanced effing brains and finally grounded abstract communications. So, my question is, does such a simulated FPD world, work the way you are predicting our world operates, in a FPD way? If not why and how must it be different? And of course, if it is different, can you describe how you might do the reverse, and simulate a MPD world, where the MPD brain can be the one that turns out to be right? All so that I might better understand with more completeness, just what it is you understand the set of FPD rules to be? Brent Allsop
Brent_Allsop replied 12 years ago (Nov 18th 2011, 10:06:55 am)
Theoreticians, I've been having a great debate with [http://canonizer.com/support_list.asp?nick_name_id=243 James Carroll], the newest supporter of Functional Property Dualism, and "Software = Qualia" in another forum. It has been very productive and has helped my clean up many of the messy thinking arguments that I often use to support Material Property Dualism. My understanding is that James' argument as to why Software = Qualia is simply that the brain is a computer, and everything in it can simulated to any desired degree of accuracy, with any sufficient abstracted, ones and zeros based system. (James, let us know if I haven't adequately described your position.) I completely agree with this assertion, but am attempting to point out that this is entirely missing what ineffable phenomenal properties are, that can also be represented by enough ones and zeroes, but only if you properly interpret the appropriate set of ones and zeros as the redness and the greenness that they represent - and that - by design, such ones and zeros aren't like anything, they are just interpretations of anything, no matter what the ones and zeros are represented with. In another attempt to better communicate to him what I'm trying to say, I've adapted the phenomenal billiard ball system I described in the previously discussed [http://canonizer.com/thread.asp/88/1/24 Effing Billard Balls] thread to be a design for a phenomenal strawberry picking robot. In the other forum, we seemed to be the only two people interested in our debate, so we both wanted to find out what others thought about who was winning this debate, so I decided to move the conversation here. If any of you are interested in indicating what you do or don't believe, and why, we'd love to hear from you. My main questions to anyone in the Software = Qualia camp would be: Would you agree that if science turned out to be just like the predictions made by the below theoretically possible world, you'd have to assume Functional Property Dualism was false and abandon your camp as being falsified. Also, in this theoretically possible reality it is simply the neural transmitters that have, in addition to all their other causal properties, also have some ineffable phenomenal properties that can be experienced. So this is where the redness and greenness come from in this theoretically possible reality, and how they are discovered, and so on. My question to Software = Qualia people is what type of set of ones and zeros might correspond with, or cause something like redness to emerge? And how might you 'discover' new phenomenal properties, and what would dictate what these new properties would be like, would there be any kind of limit to the set of phenomenal properties that could be experienced or discovered, and so on? ============= '''Mechanical description of a theoretically possible phenomenal strawberry picking robot.''' The purpose of this robot is to pick red strawberries from a field of green leaves. There are traditional digital cameras for perception, and traditional, digitally controlled arms that can receive digital commands to move to a particular location, then grasp and pick any strawberry that is at that location. We will assume that in this theoretically possible reality, there are two different neural transmitters that, in addition to behaving identically to the actual ones in this world, also have phenomenal properties. Glutamate having a redness property, and Aspartate having a greenness property. If you shine light on them, common Glutamate in crystalline form, reflects white light, and Aspartate reflects gray, just like the real stuff in this world. It is possible in this reality to create neuron like mechanism with Axon terminals, that can emit these two neurotransmitters and Dendrites with receptors that can uptake them. The brain of this system has a 1K x 1K x 1K 3D array of these neuron like structures. The cameras are wired up to a perception system that feeds into this 3D array of neurons. When the cameras are pointing to a patch in the strawberry field, it causes a set of neurons in this volume, in the shape of the strawberries and levees, to fire. When any of these neurons are stimulated by the perception system indicating there is the surface of a strawberry at the corresponding location in the field of the camera, it fires Glutamate across all of its Axon terminals. When the perception system detect the surface of a green leaf at that same location, the axon terminals fire Aspartate. There is a hierarchical neuronal structured this entire 3D volume of neurons feeds into capable of recognizing not only the red strawberries and the green leaves, but is also self aware of the fact that its redness and greenness are phenomenally experienced very different from each other. It is the phenomenal difference between the redness of the Glutamate and the Greenness of the Aspartate, that enable the system to know where the strawberries are, and where to direct the hand to pick them. The system interprets the redness as representing what it wants to pick, and the greenness as representing what should be avoided. It also has a voice system, such that if something asks what color is the item at a particular location, it has enough intelligence such that it can answer red, if the neurons at that location are firing Glutamate, and it will answer Green, if the neurons are firing Aspartate. In this particular reality, you can completely simulate everything this computer does with a simple binary system. You can have a 3 dimensional 1K by 1K by 1K array of values that can hold a binary 0 when there is green at that location, and a binary one, when there is a strawberry. When you ask it what color something is, it behaves identically, saying the strawberries are red, and the leaves are green. But of course, this system is designed not to matter what represents red or green, as long as the downstream controlling system selecting what to pick, interprets whatever is dong the representation, appropriately. Anything that is represented by something interpreted as one is also interpreted as what should be picked. Now, this particular theoretically possible reality, is just as described by Material Property Dualism. There is nothing in it, except the Glutamate, that has the redness that can be experienced. If you replace any of the 3D volume of knowledge neurons with something else, if there is no Glutamate involved, the robot experiences nothing like redness for the particular location that that neurons represents in the visible space. Let's say there was a 3rd neurotransmitter in this world that had a blueness phenomenal property - say Dopamine. You go into this 3D volume of neurons, and replace a set of them with this different neuron that fires with Dopamine, when there is a strawberry, instead of the Glutamate. And let's say the recognition system has no words for this new blueness color. When you ask the system what color is at that location, it says I see a strawberry there, but it is not like any of the other red strawberries, nor any of the green leaves. I don't know what color that is. But, if you want me to pick it, I can.