2
0

How do animals experience pain?


 invite response                
2017 Jan 18, 4:43pm   5,560 views  32 comments

by Dan8267   ➕follow (4)   💰tip   ignore  

Comments 1 - 32 of 32        Search these comments

1   Rew   2017 Jan 18, 5:15pm  

When Sarah Palin shoots them from her Alaskan Helicopter Hunting Safari ... the little animals ... they experience the pain.
When Putin strips off his shirt, and wrestles them to the ground ... they feel the pain.
;)

2   Ceffer   2017 Jan 18, 5:31pm  

I don't know. I'll ask my brother-in-law.

3   Heraclitusstudent   2017 Jan 18, 6:14pm  

The real hard question is: can you program a computer to feel pain? What would that even mean?

Consider this:
Pain is not a behavior or a data processing: if you look at an organism feeling pain from the outside you will see a signal going to the brain, and triggering more signals there. At that's it. But a signal is not pain: it's just a signal. Receiving a signal is not the same as feeling pain. A program could be programmed to receive a signal and trigger further signals. I still couldn't say that program feels pain.
So if pain is not a behavior or a data processing, it is by definition outside the space of what is programmable.

Of course if you program some kind of conscious program, there is a chance that it actually has experiences based on what it is exposed to. These experiences maybe different than what they look like from the outside, looking at the processing happening. But since it can only be programmed from the outside, then we go back to the hard question: we can't program pain because pain cannot be defined as a behavior or a data processing, and therefore we wouldn't know what to program.

4   Patrick   2017 Jan 18, 7:03pm  

Pain is dependent on having the idea of self.

Pain is less self, threat to self, etc. One of my very first posts:

https://patrick.net/19/self

5   Dan8267   2017 Jan 18, 7:04pm  

Heraclitusstudent says

can you program a computer to feel pain?

Obviously yes. Anything you can do with a neural network, you can do with a Turing machine and vice versa. They are functionally equivalent. This is easy to prove as you can run a virtual Turing machine in a neural network and a virtual neural network in a Turing machine. Thus each can do what the other can. It's only a question of which is more efficient and easier to program for a given task.

6   Heraclitusstudent   2017 Jan 18, 10:59pm  

Dan8267 says

Thus each can do what the other can.

Yes but you are missing the point: since you can't possibly know what the program feels, you can't know what to program to make it feel something specific.

A neural network just crunches a bit of data. An assembly of neural network just crunches more data.
Tell us how you need to assemble it so this data crunching somehow amount to "pain".

7   Heraclitusstudent   2017 Jan 18, 11:10pm  

rando says

Pain is dependent on having the idea of self.

No it isn't.
For example, let's say that I have a computer here that can:
- form an image of the world including recognizing objects.
- Generate observations on any situation and learn generic facts
- Learn rules and make deductions.
- choose at each instant what to do next based on whatever reasoning is produced.
- observe these observations, deduction, choices being made at associate it with the concept self, by opposition to other agents.

That's essentially a conscious program. The point is consciousness is a certain behavior and a certain data handling. This clearly is in the realm of things that can be programmed.
Yet the same problem I described still arises: there is no pain. Just data crunching. Data crunching is not pain.

Pain is not something that can be looked at from the outside, reduced to a certain processing and programmed.

8   Dan8267   2017 Jan 18, 11:55pm  

Heraclitusstudent says

A neural network just crunches a bit of data. An assembly of neural network just crunches more data.

Tell us how you need to assemble it so this data crunching somehow amount to "pain".

You do realize your brain is a neural network, right? Empirically then that "crunching of data" is what pain is.

Heraclitusstudent says

since you can't possibly know what the program feels,

Why would you think that? There is no reason to believe it is impossible to understand what a mind is experiencing, whether that mind is implemented as a neural network or by any other means, if you have sufficient knowledge and understanding of that mind and its processes.

And if you mean is it possible to experience what another neural network is feeling, then yes, it's possible simply by augmenting your brain with a copy of that other neural network. Attach your neurons to those in the other network and you can literally expand your mind encompassing the other one. This can be done with either physical or virtual neural networks.

9   Heraclitusstudent   2017 Jan 19, 9:59am  

Dan8267 says

You do realize your brain is a neural network, right? Empirically then that "crunching of data" is what pain is.

Yes, and if this is the case then it is not morally reprehensible to torture you, right? After all, it's just a bit more of data crunching.
If no why not? What is the difference?

Dan8267 says

And if you mean is it possible to experience what another neural network is feeling, then yes, it's possible simply by augmenting your brain with a copy of that other neural network.

Nope, because if you attach you brain to the brain of CIC for example, nothing says that his experience will automatically become yours. How would you know how to attach it so it becomes the case? Since you don't know what the experience is in terms of calculation to start with, there is no way to engineer such augmentation.

10   Dan8267   2017 Jan 19, 3:32pm  

Heraclitusstudent says

Yes, and if this is the case then it is not morally reprehensible to torture you, right? After all, it's just a bit more of data crunching.

If no why not? What is the difference?

WTF point are you trying to make? Whether or not a sentient being's mind is implemented as a neural network, a Turing machine, or any other means is irrelevant. It would be immoral to torture it by definition of torture and immoral.

Heraclitusstudent says

Nope, because if you attach you brain to the brain of CIC for example,

First of all, thanks for putting that thought in my brain.

Second, combining neural nets would, by definition, make the experience of the both networks part of the experience of each as they would be a single network, at least for the time they are attached. Have you ever actually built a neural network? I have. It's one of the things compsci majors do. You do realize that arguing with me on how neural networks work is like arguing with a med school graduate on how the lymphatic system works, right? This is actually an area I've formally studied and have hands-on experience with.

11   Heraclitusstudent   2017 Jan 19, 3:47pm  

Dan8267 says

WTF point are you trying to make? Whether or not a sentient being's mind is implemented as a neural network, a Turing machine, or any other means is irrelevant.

It is indeed irrelevant. And the point is simple: you accept that there can be pain in a computer. Yet all a computer does is data manipulation. And pain cannot be explained a result of data manipulation. So how do you explain the presence of something that is not data manipulation in the computer?

Dan8267 says

neural nets would, by definition, make the experience of the both networks part of the experience of each as they would be a single network

Right. You don't know where the experience arises in the network. And we said nothing about how these networks are connected. But "by definition" you already know you would "feel" the experience that CIC feels.

I wonder how because clearly it isn't obvious.

Dan8267 says

This is actually an area I've formally studied and have hands-on experience with.

Me too. What type of neuron networks did you study?

12   Dan8267   2017 Jan 19, 3:51pm  

Heraclitusstudent says

Yet all a computer does is data manipulation.

All a neural network does is data manipulation. But that's a lot of shit.

All your body does is execute the laws of physics. You're body is composed entirely of atoms and nothing else. Can atoms feel pain? No. Can you? Yes. So just because you are nothing but atoms doesn't mean you can't feel pain. It's an emergent phenomenon.

Heraclitusstudent says

What type of neuron networks did you study?

I wrote virtual neural networks and install some of them on embedded systems, i.e. robots. It was part of graduate school.

13   Heraclitusstudent   2017 Jan 19, 4:22pm  

Dan8267 says

All a neural network does is data manipulation. But that's a lot of shit.

Granted but that is still math and data manipulation.

Dan8267 says

Can atoms feel pain? No. Can you? Yes. So just because you are nothing but atoms doesn't mean you can't feel pain.

Ok

Dan8267 says

It's an emergent phenomenon.

I would say it is a phenomenon that is linked to the point of view:
- If you look at it from the outside: you see no pain. Just physics unfolding.
- But if you ARE the 'network' then you have the experience. But how? this is million dollar question. This is not just perception. This is not just awareness of some data here. This is not just information.

Dan8267 says

I wrote virtual neural networks and install some of them on embedded systems, i.e. robots. It was part of graduate school.

As a result of your experience, would you honestly say you could assemble these networks in a way that you could say pain will be felt somewhere in there?

14   Patrick   2017 Jan 19, 5:52pm  

Heraclitusstudent says

- But if you ARE the 'network' then you have the experience. But how? this is million dollar question. This is not just perception. This is not just awareness of some data here. This is not just information.

My own theory is that it's all about recursion.

Consciousness is perception of perception.

Not sure if a computer can do that.

15   Heraclitusstudent   2017 Jan 19, 6:36pm  

rando says

Consciousness is perception of perception.

Not sure if a computer can do that.

It certainly can. I think consciousness is a behavior and can be coded. I don't see this as the hard problem.
But pain not like perceiving something, it is not information really. So how to code it is really a mystery. That is the hard problem for me.

Not only pain really, but all sensations. Even colors: When you see the color blue you don't see a number. You have a specific sensation.
A computer could of course perceive colors, as numbers, or some kind of bitmaps. This is information. But the sensation "blue" is not information and how this could make it into a computer is a mystery.

16   marcus   2017 Jan 19, 7:30pm  

Dan8267 says

How do animals experience pain?

Consider tomorrow an experiment using American humans. Republicans are the control group.

17   Dan8267   2017 Jan 20, 9:21am  

Heraclitusstudent says

Granted but that is still math and data manipulation.

Everything in the universe is math and data manipulation including you and me.

Heraclitusstudent says

Dan8267 says

Can atoms feel pain? No. Can you? Yes. So just because you are nothing but atoms doesn't mean you can't feel pain.

Ok

OK, you realize how this discredits any argument in the form "x is only y, therefore x has only the properties of y"?

Heraclitusstudent says

I would say it is a phenomenon that is linked to the point of view:

Irrelevant. This would not contradict anything I have stated. Everything that is subjective has an objective, physical manifestation in the universe. This includes everything your mind experiences such as the color blue, love, and consciousness. The physical manifestation of those things are electrical and chemical signals. The mind is entirely composed of the brain and nothing but the brain.

[more stupid comment limits]

18   Dan8267   2017 Jan 20, 9:26am  

Heraclitusstudent says

As a result of your experience, would you honestly say you could assemble these networks in a way that you could say pain will be felt somewhere in there?

It depends on what you mean by this question.

Can an artificial neural network, physical or virtual, be assembled such that it experiences pain? Obviously, yes. There is absolutely no difference between two things, neural networks or anything else, that are atom-by-atom the same but created by nature or man. There is nothing supernatural about neural networks or any other phenomenon in the universe. If nature can do it, so can intelligent beings. Natural or synthetic does not matter at all.

Does mankind currently have the knowledge to do this? No. At least not that I know of. I wouldn't be shocked if some research team is cracking the problem right now, but it's unlikely. Will we have the necessary knowledge in a hundred years? Probably. Will we have it in our lifetime? Maybe.

Can I personally do this right now? Obviously, no. If the knowledge isn't possessed by mankind, then it's not possessed by me since I'm part of mankind.

Could I figure out how to do it? Possibly. It's a tough problem to crack, and nature does not give up its secrets easily. However, it would take considerable time and effort. I have a day job and would not devote my time to doing this. In fact, I have no motivation, and actually quite a lot of repulsion, to create a neural network capable of experiencing pain. This would be unethical to me. I understand that there may be noble reasons for doing so, such as understanding how pain works so we can minimize or eliminate it. However, I'm not comfortable creating any mind and then inflecting pain on it. Even if that mind is simple and primitive. This is, however, a difficult ethical subject.

19   Dan8267   2017 Jan 20, 9:27am  

stupid comment limit won't let me edit the above post to fix the i tag. So doing that here.

20   Dan8267   2017 Jan 20, 9:31am  

rando says

Not sure if a computer can do that.

Computers have done that. You and I have done that and our brains are literally computers. They are devices that store, retrieve, and process information. They may be neural networks rather than Turing machines, but neural networks are computers. I would even consider every single cell in your body as a computer even though it's neither a Turing machine nor a neural network.

In any case, anything that a neural network can do, a Turing machine can do by running a virtual copy of that neural network. So trivially, anything you or I can do, a sufficiently advanced Turing machine can do.

21   Heraclitusstudent   2017 Jan 20, 11:36am  

Dan8267 says

In fact, I have no motivation, and actually quite a lot of repulsion, to create a neural network capable of experiencing pain. This would be unethical to me. I understand that there may be noble reasons for doing so, such as understanding how pain works so we can minimize or eliminate it. However, I'm not comfortable creating any mind and then inflecting pain on it.

If only ethics prevent you from doing it, then create a network that feels pleasure. Same problem.

22   Heraclitusstudent   2017 Jan 20, 11:46am  

Dan8267 says

Heraclitusstudent says

I would say it is a phenomenon that is linked to the point of view:

Irrelevant. This would not contradict anything I have stated. Everything that is subjective has an objective, physical manifestation in the universe. This includes everything your mind experiences such as the color blue, love, and consciousness. The physical manifestation of those things are electrical and chemical signals. The mind is entirely composed of the brain and nothing but the brain.

You are not acknowledging the core of the problem which is that pain is not felt by us like information.
Yet everything a computer (or a brain) does is manipulating information.
The only way this apparent contradiction could be reconciled is because the point of view is different. Therefore it is relevant.

All you keep repeating is yes everything is math and data manipulation, even pain. Nope, sorry, math and data manipulation are not pain. Information on a paper or in a computer is not pain.
Until you acknowledge the problem. It's not worth discussing it.

23   Dan8267   2017 Jan 20, 12:13pm  

Heraclitusstudent says

If only ethics prevent you from doing it, then create a network that feels pleasure. Same problem.

Not quite. But it is a difficult problem as well.

Nonetheless, my answers are the same. It is possible for such a neural network to be created artificially. Mankind does not yet know how to do this, but probably will in the next 100 years. I personally don't know how to do this, but it is plausible but not certain that if I worked at it diligently for many years, I could solve the problem. I'm not willing to do that without a solid guaranteed income for doing so because I already have a very good and stable job that I like.

I fail to see your point.

Heraclitusstudent says

You are not acknowledging the core of the problem which is that pain is not felt by us like information.

I don't even know what "felt by us like information" means. WTF is feeling information?

Heraclitusstudent says

Yet everything a computer (or a brain) does is manipulating information.

Everything a computer, a brain, a person, a animal, an ecosystem, a galaxy does is following the laws of nature. It's all atoms and quantum particles performing interactions govern by natural laws expressed mathematically. So if a feeling or emotion or consciousness can exist in our universe, then by definition, it's represented physically and obeys natural laws expressed mathematically.

Heraclitusstudent says

The only way this apparent contradiction

What contradiction?

[stupid comment limit]

24   Heraclitusstudent   2017 Jan 20, 12:18pm  

Dan8267 says

What contradiction?

Ok so we can stop this discussion here, if you fail to see the problem.

25   Dan8267   2017 Jan 20, 12:20pm  

Heraclitusstudent says

Nope, sorry, math and data manipulation are not pain.

I didn't say math and data are pain. I'm saying that pain is implemented by physical processes, acting on information, that obey natural laws that are expressed mathematically and therefore, by definition, pain can be implemented by copying those physical process and data either exactly or as something functionally equivalent. For example, I could copy a brain, or part of a brain that feels brain, cell by cell into an identical neural network by cloning the cells and connecting them in the same way and providing the same chemical cocktails. The newly constructed brain would feel pain. If I replaced the cells in the new brain with functionally equivalent cells of another species, even one that doesn't feel pain, then the resulting brain would still feel pain. Furthermore, if I replaced the cells with synthetic neurons on a chip, and had the chemical interactions also simulated or replaced by something that performs the same job, then again the resulting brain, now entirely silicon circuitry, would feel pain. Finally, I can replace the physical computer chip with a virtual chip simulated in software, and again, the virtual brain would feel pain.

Give me one reason why any of the above conclusions are wrong.

26   Dan8267   2017 Jan 20, 12:20pm  

Heraclitusstudent says

Ok so we can stop this discussion here, if you fail to see the problem.

That sounds like giving up.

27   Heraclitusstudent   2017 Jan 20, 12:35pm  

Dan8267 says

I didn't say math and data are pain.

Yes you said: "Everything in the universe is math and data manipulation". That includes pain.

Dan8267 says

Give me one reason why any of the above conclusions are wrong.

They are not wrong, it's just variation on the same thing you keep repeating "Everything in the universe is math and data manipulation".

You refuse to acknowledge that there is a dual nature of (1) looking at pain from the outside and seeing only maths and data manipulation, and (2) actually feeling pain.

Yes, I'm giving up. There is no point discussing a problem with someone who doesn't even see the problem.

28   Dan8267   2017 Jan 20, 3:01pm  

Heraclitusstudent says

Dan8267 says

I didn't say math and data are pain.

Yes you said: "Everything in the universe is math and data manipulation". That includes pain.

OK, I get it. You're reading comprehension skills are inadequate. Let me be pedantic.

The statement "everything in the universe is math and data manipulation" does not mean the same thing as the statement "math and data manipulation is everything in the universe". You see word order matters in English. By reversing the subject and object, the entire sentence means something completely different. Therefore, the statement "everything in the universe is math and data manipulation" does not imply that "math and data manipulation are pain" but rather that "pain is math and data manipulation".

Pain is a certain kind of data processing and is entirely implemented using physical objects obeying natural laws. There is nothing magical or supernatural about it, and there is no reason why an artificial device, physical or virtual, cannot be constructed that would experience pain. All subjective phenomenon are ultimately implemented as objective reality.

[stupid comment limit]

29   Dan8267   2017 Jan 20, 3:02pm  

Heraclitusstudent says

You refuse to acknowledge that there is a dual nature of (1) looking at pain from the outside and seeing only maths and data manipulation, and (2) actually feeling pain.

What are you looking for? Of course actually feeling pain is different than looking at a person feeling pain or looking at a real-time brain image of a person feeling pain. That is a stupid point that has no relevance to anything in this thread. It's like saying it feels different to eat a pie than to watch someone eat a pie. That doesn't mean you can't create a robot that eats pies and turn the chemical energy into electricity to run it's processors and motors. Digestion and the sense of taste can be implemented artificially as well.

30   Dan8267   2017 Jan 20, 3:03pm  

Heraclitusstudent says

Yes, I'm giving up. There is no point discussing a problem with someone who doesn't even see the problem.

I refuse to simply accept bullshit. I call bullshit whenever I hear it, and will continue to do so. Vague mumble jumbo has no place in serious discussions.

31   Heraclitusstudent   2017 Jan 20, 5:31pm  

Dan8267 says

Of course actually feeling pain is different than looking at a person feeling pain or looking at a real-time brain image of a person feeling pain.

Heraclitusstudent says

These experiences maybe different than what they look like from the outside, looking at the processing happening. But since it can only be programmed from the outside, then we go back to the hard question: we can't program pain because pain cannot be defined as a behavior or a data processing, and therefore we wouldn't know what to program.

Dan8267 says

the sense of taste can be implemented artificially as well.

Heraclitusstudent says

A computer could of course perceive colors, as numbers, or some kind of bitmaps. This is information. But the sensation "blue" is not information

32   Dan8267   2017 Jan 21, 8:59am  

Heraclitusstudent says

A computer could of course perceive colors, as numbers, or some kind of bitmaps. This is information. But the sensation "blue" is not information

Whether or not how computers currently process image data constitutes sensation is irrelevant. We are talking about whether or not sensations could be perceived by computers, specifically Turing machines, and the answer is a resounding yes. Your brain could be simulated perfectly inside a sufficiently powerful Turing machine. If you can feel a sensation, then that Turing machine can also simply by running a virtual copy of your brain. Hell, it's theoretically possible that right now your brain is a virtual machine running inside a Turing machine called the Matrix and there is no way you could tell otherwise. This is an unlikely scenario, but its mere possibility proves you wrong. In order for your statements to be right, it would have to be impossible for anything like the Matrix to be able to exist.

Please register to comment:

api   best comments   contact   latest images   memes   one year ago   random   suggestions