“Materialism” is an empty word. An incendiary article.

[LarryS]

Correct because snow is an aggregate of frozen ice crystals that are an external referents in the idiomatic self referencing system of symbols used in language between communicants.

I spit out my coffee reading this

 

[angrysoba]

Uh, no. That is not what it means. Which word are you having trouble with? "Semantic"? "Word"? "Game"? In any way, just look up those words in the dictionary and you will have a very clear idea of what I mean.

Ron, that's no where near as clever or as witty as you may like to think it is.

A lot of arguments require a clarification about the definition of terms being used. You cannot understand "natural selection" by simply looking up "natural" and "selection" in the dictionary. So it should not be assumed that looking up the three words in "semantic word games" will be very helpful either. In fact, it seems likely that you have not looked up the meaning of these words either or you may have simply said "semantic games" or "word games". Why "semantic word games" unless you thought it made you sound cleverer?

Essentially, phiwum is correct, isn't he. You are using the term "semantic word games" to dismiss something you are not interested in.

 

[Dancing David]

Can you link some article about this kind of measurement? Thank you.

That is a rather broad area actually, in Cognitive Behavioral Therapy an huge number of cognitions and moods can be tracked as part of the intervention.

I am reluctant to link to a specific tool kit (because they come from sources unknown to me) so I have to read some before I get back to you.


Aaron Beck started a trend with the Cognitive Therapy of Depression
http://www.amazon.com/Cognitive-Depression-Guilford-Psychology-Psychopathology/dp/0898629195


But I want to do some more reading before recommending mood assessments from the web.

 

[NotEvenWrong]

Qualia can't be shown to exist?

How do you know they exist?

Not sure if you're joking due to the smiley, but...

I do not believe qualia can be shown to exist in the empirical, physical, measurable sense because there is no way to directly or indirectly quantify nor measure an emotional state like "anger". Hard problem of consciousness and all that. Perhaps there will be a breakthrough one day, but I don't see how at this point. You're essentially trying to quantify a qualitative, internal state with a physical model. No idea what that would even look like.

You could assume neurons interacting across time domains gives rise to some kind of emotion. But this is a black box model and we would still have no idea "why" neural networks when in some configurations gives rise to consciousness. Not in the same way we understand how electrons moving around gives rise to a computer -- because we can actually quantify electron movement and map that to information content and build a computer as a result. Emotions and personal experience are not simply "information content"; subjective experience appears to be something different.

However, I still know qualia actually do exist because I experience them. No idea why I (or anyone else) experiences them if we are indeed just a mass of particles interacting, but nonetheless we do.

 

[Nick227]

Essentially, phiwum is correct, isn't he. You are using the term "semantic word games" to dismiss something you are not interested in.

Yes, accusing the other of semantic word games when you can't actually grasp the argument is popular. This is true.

Another very popular ploy is to accuse the poster of solipsism! That the argument isn't remotely solipsist doesn't appear to matter. Once you've uttered the word apparently the position has just lost any validity!

 

[Nick227]

I do not believe qualia can be shown to exist in the empirical, physical, measurable sense because there is no way to directly or indirectly quantify nor measure an emotional state like "anger". Hard problem of consciousness and all that. Perhaps there will be a breakthrough one day, but I don't see how at this point. You're essentially trying to quantify a qualitative, internal state with a physical model. No idea what that would even look like.

You don't actually need to overcome the hard problem to quantify emotional states. Not in principle anyway. There is neural correlation. That's enough to quantify.

 

[Dancing David]

I spit out my coffee reading this

Glad to help you do some cleaning.

 

[NotEvenWrong]

You don't actually need to overcome the hard problem to quantify emotional states. Not in principle anyway. There is neural correlation. That's enough to quantify.

I am talking about a mechanistic model (a why/how does a neural network give rise to an emotional state). Not a statistical correlation model. The correlation part is "easy" in principle given a powerful enough set of measuring tools.

An example of a mechanistic model would be general relativity -- Einstein's field equations explaining how matter causes spacetime to curve which gives rise to gravity. i.e. showing explicitly how a physical model leads to gravitation in a logically explicit way.

To do something like this with consciousness we would need to derive some set of equations using first principles and consciousness would "pop out" the way gravity comes about due to the curvature of spacetime. I don't see how this is possible since consciousness is a qualitative internal state (where everything that is physical is observable and quantifiable, at least in principle).

 

[Nick227]

I am talking about a mechanistic model (a why/how does a neural network give rise to an emotional state). Not a statistical correlation model. The correlation part is "easy" in principle given a powerful enough set of measuring tools.

But if you have the correlation you can probably measure the degree of activity. This does not overcome the hard problem, the how, but it does give a quantity. You could also do it with some form of subjective monitoring going along at the same time. How angry do you feel right now? - this kind of thing. Sounds reasonable to me.

To do something like this with consciousness we would need to derive some set of equations using first principles and consciousness would "pop out" the way gravity comes about due to the curvature of spacetime. I don't see how this is possible since consciousness is a qualitative internal state (where everything that is physical is observable and quantifiable, at least in principle).

Oh, there are still a whole heap of problems with getting from neurons to conscious awareness, I don't doubt it for a second.

I mean, the way people tie consciousness and the brain together with language, they set themselves up for problems if you ask me. There is inevitably a huge evolutionary bias and it doesn't get accounted for. First thing the brain does with conscious awareness is assign it an owner, a "me". Next thing is to partition the contents into me and not me - the body and the rest. Next thing is to arrange the contents to suggest that it's being observed from some position behind the eyes. All of this is evolutionary conditioning, it gives a survival advantage. None of it is a priori. For me you have to factor all this in before you can even articulate the hard problem properly.

 

[Marras]

Formal logic is thus a useful tool for making explicit philosophical claims. Of course, it doesn't work well in every domain, but it can be useful in certain cases.

I agree that it can be useful for eliminating ambiguity. But when it advances to constructing actual proofs, it often gets quite fishy. I have to admit that I haven't read that many papers in philosophy, I don't have strong enough background with the terminology so it's always an effort for me. But I've occasionally had to examine proofs that have been presented in such papers. And most of those have been cringe-worthy. Though, there may be some amount of selection bias involved with this where only the bad proofs catch my attention and I'm just unaware of the good ones. As I said, I haven't read much philosophy.

One minor question: you said that humans can't solve a SAT problem with 1000 variables. Is this currently feasible with automated SAT solvers?

Actually, I wrote a human can't create a problem with 1000 variables. For solving them the limit is much smaller. You need really good dedication and attention to details to solve a problem with something like 20 variables unless it is a trivial one. When it comes to difficult instances my personal limit for hand-solving would be somewhere around 5-6 variables, at that point the proof trees grow large enough that I don't trust myself to not make some simple stupid mistake somewhere along the way.

SAT problem is interesting in that there is a huge difference between difficulty in solving different instances. Most are easy. They either have lots of models so one can be found easily, are so overconstrained that they can be proved inconsistent fast, or they have simple structure that allows solvers to use truth-value propagation rules efficiently.

When you have a nice clean instance that has clean structure, then modern solvers can handle hundreds of thousands of atoms and millions of clauses.

However, some instances are very difficult. The most difficult area is called the phase-transition area for SAT (transition is from underconstrained to overconstrainted problems). It's an area of random instances that have no structure and approximately half of them have a model and a half are unsatisfiable. I'm not certain how well the best current solvers handle the area, but about 5 years ago they could get to about 700 variables.

 

[phiwum]

I agree that it can be useful for eliminating ambiguity. But when it advances to constructing actual proofs, it often gets quite fishy. I have to admit that I haven't read that many papers in philosophy, I don't have strong enough background with the terminology so it's always an effort for me. But I've occasionally had to examine proofs that have been presented in such papers. And most of those have been cringe-worthy. Though, there may be some amount of selection bias involved with this where only the bad proofs catch my attention and I'm just unaware of the good ones. As I said, I haven't read much philosophy.

My background is in philosophical logic, with considerable overlap with mathematical logic. I have not seen any especial problems with philosophers doing formal logic, but the proofs I tend to read are written by logicians.

 

[phiwum]

Actually, I wrote a human can't create a problem with 1000 variables. For solving them the limit is much smaller. You need really good dedication and attention to details to solve a problem with something like 20 variables unless it is a trivial one. When it comes to difficult instances my personal limit for hand-solving would be somewhere around 5-6 variables, at that point the proof trees grow large enough that I don't trust myself to not make some simple stupid mistake somewhere along the way.

SAT problem is interesting in that there is a huge difference between difficulty in solving different instances. Most are easy. They either have lots of models so one can be found easily, are so overconstrained that they can be proved inconsistent fast, or they have simple structure that allows solvers to use truth-value propagation rules efficiently.

When you have a nice clean instance that has clean structure, then modern solvers can handle hundreds of thousands of atoms and millions of clauses.

However, some instances are very difficult. The most difficult area is called the phase-transition area for SAT (transition is from underconstrained to overconstrainted problems). It's an area of random instances that have no structure and approximately half of them have a model and a half are unsatisfiable. I'm not certain how well the best current solvers handle the area, but about 5 years ago they could get to about 700 variables.

Interesting. I had assumed the problem blows up faster than that, so that 700 variables would be infeasible.

 

[Dancing David]

Can you link some article about this kind of measurement? Thank you.

Hi,
in general they look like this:
http://www.cqaimh.org/pdf/tool_edu_moodchart.pdf

they are qualitative tools where an individual rates their mood.

In CBT there are usually personal idiomatic markers that are added as well.

 

[Dancing David]

Not sure if you're joking due to the smiley, but...

I do not believe qualia can be shown to exist in the empirical, physical, measurable sense because there is no way to directly or indirectly quantify nor measure an emotional state like "anger". Hard problem of consciousness and all that. Perhaps there will be a breakthrough one day, but I don't see how at this point. You're essentially trying to quantify a qualitative, internal state with a physical model. No idea what that would even look like.

You could assume neurons interacting across time domains gives rise to some kind of emotion. But this is a black box model and we would still have no idea "why" neural networks when in some configurations gives rise to consciousness. Not in the same way we understand how electrons moving around gives rise to a computer -- because we can actually quantify electron movement and map that to information content and build a computer as a result. Emotions and personal experience are not simply "information content"; subjective experience appears to be something different.

However, I still know qualia actually do exist because I experience them. No idea why I (or anyone else) experiences them if we are indeed just a mass of particles interacting, but nonetheless we do.

What if you substitute 'perception' for 'qualia'?

Not perfect understanding but growing area. They are not just a mass of particles randomly interacting.

Can you see qualia of color if you are color blind. Or blind?

 

[NotEvenWrong]

What if you substitute 'perception' for 'qualia'?

Perception I would say is the act of becoming aware of a stimulus. So yeah I think that can be measured. A robot can perceive, for instance, under a broad enough definition, and there is probably a quantitative way to define and measure it. Perhaps you could call it the moment when sensory information (in a human, animal, or computer/robot) is integrated and the agent becomes aware of it. No need for subjective, qualitative definitions. Measure electrical or nerve impulses and figure out where they are integrated and the agent can act on them. (humans can perceive stimulus by reflex, for example. No need for a brain even.)

Not perfect understanding but growing area. They are not just a mass of particles randomly interacting.

Remove the word "randomly" and how could they not be? Under any reasonable definition of materialism?

Can you see qualia of color if you are color blind. Or blind?

If you are born blind I think the consensus is "no, you can't". The brain becomes wired differently and you need to experience color for the brain to be able to "see" it. Now if you are born with normal sight but become blind later then I think you can still experience color by visualizing mentally (but you can no longer see it through stimulus reaching your eyes).

 

[Roger Ramjets]

That computer is also imaginary.

1889: Heavier-than-air flying machines are imaginary!
1939: Bombs that convert matter into energy are imaginary!
1979: Personal computers with 20,000 million transistors that run at 8GHz are imaginary!

In 1979 I built my first computer with 1k of RAM and a 1MHz CPU. If you had told me back then that within a few short years I would be building a PC with Gigabytes of RAM and a CPU running at microwave frequency, I would not have believed you. And yet, here we are...

 

[tsig]

1889: Heavier-than-air flying machines are imaginary!
1939: Bombs that convert matter into energy are imaginary!
1979: Personal computers with 20,000 million transistors that run at 8GHz are imaginary!

In 1979 I built my first computer with 1k of RAM and a 1MHz CPU. If you had told me back then that within a few short years I would be building a PC with Gigabytes of RAM and a CPU running at microwave frequency, I would not have believed you. And yet, here we are...

My car can go from 0 to 60 MPH in 20 seconds so obviously at the end of one minute I'll be going 180 MPH and at the end of ten minutes 1,800 Mph.

 

[Fudbucker]

Qualia can't be shown to exist?

Well, you know your own subjective experiences exist. But with regards to others, no. The existence of anyone else's subjective experiences can only be inferred, and the content of those experiences will always be a mystery. There is no "qualia detector". There is no way for me to know what your experience of the color red is like, and vice-versa. For all we know, our subjective experiences of colors varies from person to person.

How do you know they exist?

If your brain is similar to mine, I can infer you experience things roughly like I do (even if the content of your experiences is unknowable to anyone other than you), but I can't be sure. If you make a mechanical brain that is functionally identical to a human brain, it should experience things, but for all we know it might be a zombie. There's no way to test for the existence of experience.

 

[Fudbucker]

'The researchers successfully measured the electrical impulses correlating with subatomic particles hitting the sensors, not the particles themselves'.

When an animal experiences fear, hate or love it is a measurable physiological effect - but when humans do it we insist that it is something else. To that argument I ask, what does it feel like to be a computer? Do they experience 'qualia' like us? Of course they do, just like humans do when specific parts of the brain are stimulated with electrical impulses. We only think we are different due to how our conscious mind interprets these 'feelings'.

How do you know computers can experience things? Can calculators? Light switches? Abacuses?

 

[Fudbucker]

You don't actually need to overcome the hard problem to quantify emotional states. Not in principle anyway. There is neural correlation. That's enough to quantify.

I think it's a pretty good tool (like facial expressions), but there's no way to prove that what we think a correlation of an experience is is actually causing that experience (unless you image your own brain). I.e., just because someone yells "ow" and winces does not prove they're in pain. Again, it's a good tool, and we can make a strong inference, but we can never know for sure.