Materialists......

[chriswl]

Because I've just spent a year studying the philosophical foundations of cognitive science and computationalism is as dead as the dodo. It is history. It went the same way as behaviourism, it's just some remnants of the computationalist rump (the slow ones) haven't yet cottoned on to the fact that it is in indeed dead. Connectionism delivered the first blow. Searle, Jackson and the "emobodied cognition" crew finished off the job. Computationism is dead. Let it rest in peace.

I'm not convinced by any of the formal arguments I've seen against computationalism and I don't think neural networks are any different in what they can accomplish logically from von-Neumann computers (after all, hardly anyone builds neural networks - they just simulate them on conventional computers because its easier).

Having said that, the idea of "embodied cognition" (as explained here http://www.iep.utm.edu/e/embodcog.htm) does seem to make a lot of sense. What puzzles me is why you think it is incompatible with materialism? The whole thing seems a lot more physically-grounded and less etherial than viewing cognition as computation.

If cognition isn't associated with a particular physical phenomenon (like computation) but is an umbrella term for a range of quite disparate concrete behaviours undertaken by organisms (or potentially machines) then perhaps there is no single, non-physical thing called "mind" that requires explanation? Perhaps we have created a mystery for ourselves by trying to find a single answer to a bunch of quite different questions?

 

[nescafe]

How exactly is the data retrieved, then ? (an answer beside "not very well" would be appreciated )

Well, if you want to know exactly how data is stored and retrived.... I do not know enough to describe the math that charecterizes the various classes of neural networks. However the way all neural networks work in general is that

• They consist of a bunch of connected neurons. The only salient things about these neurons for the purposes of this exposition is that they are connected to other neurons, and individually act as a function that takes a set of inputs and outputs a single bit of information. Real neural networks are a bit more complicated, but the principle is the same.
• They are trained on a series of inputs to yield a desired output.
• The training process essentially creates a function that maps the input to the output and encodes that function in the strengths of the synaptic connections.
• Once they have been trained, giving them part of an input that matches the input they were trained on is enough to elicit the output that maps to the input they were trained for.

... and now you are about as ignorant as I am on the subject of neural networks.

 

[Dark Jaguar]

So as a philosophy student, you claim that ALL the people working on AI programs at MIT, ALL those programmers, are wasting their time because what they are trying to do is impossible?

Also, what an odd sense of morality. Consequences of action are irrelevent? At least that's what the wikipedia article seems to be saying. Sorry, I can't go with that. Consequences seem to be all that MATTERS when it comes to an action, at least the most probable consequence as one who is deciding can judge it. It's not okay to steal because if one says "it is okay to steal" and universalizes it, then property can't exist and the statement is meaningless". Interesting, except it seems a nonsequiter, as well as making only two options when there are more. What about just CHANGING the statement to "propery does not exist"? Universalized THAT still makes sense. Stealing won't exist, but a lot of people would disagree with the person that does it and might stop him. Maybe consequences are important? It is "never okay to lie" because you can't universalize it? Fine, I'll remember that next time you are being hunted down by someone wanting to kill you and they demand your location from me. Still think consequences are irrelevent? Philosophy is all well and good until reality steps in and forces you to make a choice that doesn't always have a "perfectly logically formed solution that is applicable in all situations in perpetuity througout the universe".

And "all you proved are minds SUPERVENE on brains!". Well gee sherlock! Thanks for that. We don't HAVE 100% evidence that brains are where minds are stored. We have very compelling evidence in the FORM of the supervening.

If I were to say that clouds cause rain, would you respond "NO! All the science says is that rain SUPERVENES on clouds!", would that be a rational response?

If I were to say that science indicates evolution takes place, would it be rational to say "NO! All science can say is that a bunch of creatures mutated, but it can't PROVE that there's not some higher intelligence guiding the process!", would THAT be rational?

If you answered "no", then why is it rational to say that the evidence that minds ALWAYS seem to supervene on brains, and NEVER seem to occur anywhere else, is NOT somehow VERY compelling evidence that brains ARE our minds. Why invent some outside thing when it isn't needed to explain our mental functions? We have a sufficient model as it is. Remember occam's razor?

(Edit: Going back to the moral thing, why is it that no philosophical model of "objective morality" ever seems to satisfactoraly address the issue of how one can morally command their kids to do things or not do things but then the adult can do those very same things if they so chose? Are they "too simple" or "are animalistic barbarians that can't be trusted with logic"?)

 

[hammegk]

<pedant mode="extra snark">
Hammegk left out graviton exchange, though, which renders his entire worldview flawed.
</pedant>

Originally posted by hammegk

Our perceived-as-physical-structures communicate with what is not our specific body via photon exchange (and probably via graviton exchange). So far as science has ascertained other bosons do not effect or affect our bodies by communicating outside of them.

Granted, it is a few more layers of abstraction than is generally needed.

Who knows?

 

[hammegk]

In other words, for the confused reader, JustGeoff is presuming the immaterial nature of soul, mind, and experience right off the bat;

Yes, just as other confused readers "presume" physical reality exists.

thereby nullifying any scientistic or materialist perspective automatically.

Science works equally well in any ontological reality as long as that reality meets the objectivety required for the scientific method.

So, basically, he's assuming his conclusions automatically.

Exactly as the 'materialists' make the opposite assumption and reach their conclusion equally automatically.

How am I doing so far?

You appear to grasp the basic concept.

 

[nescafe]

Originally posted by hammegk

oooh, damn, caught not reading the full parent post chain. Time to resign my pedant certificate for a while.

 

[Belz...]

Yes, just as other confused readers "presume" physical reality exists.

Woah. You'll have to spoonfeed me the reason why this wouldn't be a rational, defensible assumption.

 

[I less than three logic]

Woah. You'll have to spoonfeed me the reason why this wouldn't be a rational, defensible assumption.

I’m sure hammegk will correct me if I’m wrong, but I believe he has said that this is rational, defensible assumption if you also reject the notion of God or Gods (and similar claims), since they contradict the notion of a solely physical reality.

 

[DreadNiK]

I’m sure hammegk will correct me if I’m wrong, but I believe he has said that this is rational, defensible assumption if you also reject the notion of God or Gods (and similar claims), since they contradict the notion of a solely physical reality.

If that is true, then it IS a rational, defensible assumption.

 

[hammegk]

And a way out of the cartesian dualism muddle is to accept "mind" as a non-causally-efficaceous epiphenoma. That is, it is useless froth generated after the fact by the physical system doing the processing.

 

[Tricky]

Woah. You'll have to spoonfeed me the reason why this wouldn't be a rational, defensible assumption.

The competing presumption is that there is something that is NOT a physical reality but which gives the illusion of a physical reality and simulates every single thing that we associate with a physical reality. And of course, it cannot be distinguished from a physical reality. What's unreasonable about that presumption?
(Shut up, Occam.)

 

[UndercoverElephant]

[*] How the network is layed out.
[*] The weights of the interneural connections.
[/list]
then copying a neural network wholesale is certianly feasible, and that copy of the neural network would still encode the same thing.

Yes, but this means that you still can't copy bits of one persons brain and transfer them to another persons brain. You could only replicate the whole thing. But this is no good for scientific study. It doesn't satisfy the requirements you needed at the beginning of this diversion. Your machine is still impossible.

 

[Tricky]

Yes, but this means that you still can't copy bits of one persons brain and transfer them to another persons brain.

Correction. We cannot at this time. There is nothing to say that it is impossible, only extremely difficult. We can't copy whole brains either, but for some reason, you don't seem to have a problem with future science managing that.

 

[UndercoverElephant]

I'm just trying to keep my emotions out of the way when discussing these things.

Why? Don't they matter? This is the R&P board. Emotions are allowed.

This said, do you have any solid indication that, indeed, we're not just the sum of our parts ?

I believe so, yes.

 

[UndercoverElephant]

I do believe this makes my point. Subjectivity and objectivity cannot be actual opposites, because one requires the other, while the reverse isn't true.

Subjective thoughts are based upon gathering objective data and interpreting it using already gathered objective data.

This is an interesting claim, philosophically. It begs the following question: "How do we gather objective data in the first place?" or "What allows us to gather objective data?" or "What allows us to have knowledge of an objective world?" Kant tried to answer questions like these. He tried to change the questions to these sorts of questions.

 

[UndercoverElephant]

HOW the information is stored doesn't change anything about the facts: information IS stored in the brain.

It makes a big fat difference if you want to extract parts of it from the whole.

 

[UndercoverElephant]

How exactly is the data retrieved, then ? (an answer beside "not very well" would be appreciated )

Depends whether you are talking about A.I. or brains. In A.I. it is retrieved by stimulating the inputs and reading the outputs, not knowing precisely what goes on inbetween. In brains, we pretty much don't know - apart from data gathered by various types of psychologist. We know enough about the way memories are stored from a subjective point of view to be able to devise techniques for people to be able to remember the order of four packs of cards, for example - by using associative symbols. That is content-addressable memory from a subjective viewpoint.

 

[UndercoverElephant]

Subjective experience is based on human emotion, bias, feelings, etc. Objective processes are free of human emotion, bias, feelings.

Any particular reason why you are saying "emotion" instead of "mind"? Emotions are part of human subjective experience, but surely not all of it. What about sensation? Is the experience of seeing the colour red an emotion? A belief?

How am I doing so far?

OK.

 

[UndercoverElephant]

Having said that, the idea of "embodied cognition" (as explained here http://www.iep.utm.edu/e/embodcog.htm) does seem to make a lot of sense.

Well then we have arrived at agreement. This bunch of people are the bridge-builders. They are influenced by people like Heidegeer and Merleau-Ponty as well as the mainstream of cognitive science.

What puzzles me is why you think it is incompatible with materialism?

Because people like Heidegger and Merleau-Ponty were the students/successors of Edmund Husserl, and Husserl more than anyone else in the history of philosophy was an enemy of reductive materialism. Husserl was the person who called the alarm on materialism. Heidegger and Merleau-Ponty followed him and stressed neutrality and the resolution of subjective and objective. The entire movement which the "embodied cognition" people are trying to build bridges to are the phenomenologists, and phenomenology is a branch of philosophy established as a direct reaction to reductive materialism.

Have a look at this:

http://en.wikipedia.org/wiki/Maurice_Merleau-Ponty

The whole thing seems a lot more physically-grounded and less etherial than viewing cognition as computation.

YES! The materialist theories of mind end up relegating "mind" to a process in a physical thing. But it isn't. Mind "reaches out".....that is the only way direct realism ends up being true after all....

If cognition isn't associated with a particular physical phenomenon (like computation) but is an umbrella term for a range of quite disparate concrete behaviours undertaken by organisms (or potentially machines) then perhaps there is no single, non-physical thing called "mind" that requires explanation?

There's a thing called subjectivity. Really there is.

Perhaps we have created a mystery for ourselves by trying to find a single answer to a bunch of quite different questions?

The questions are all inter-related. The mystery arises from the failure to relate the questions to each other properly. The questions end up cancelling each other out.

 

[UndercoverElephant]

So as a philosophy student, you claim that ALL the people working on AI programs at MIT, ALL those programmers, are wasting their time because what they are trying to do is impossible?

Far from it. If they wish to design a system which recognises breast cancer samples from non-cancer samples, then they will continue to have great successes. The problem is not with the engineering applications of A.I. but with the philosophical/cognitive applications. A.I. is a powerful tool, but not for studying minds. It is an engineering tool, not a research tool for pyschologists.

Also, what an odd sense of morality. Consequences of action are irrelevent?

No, they aren't irrelevant. But I believe intentions are more important than consequences, morally.

At least that's what the wikipedia article seems to be saying. Sorry, I can't go with that.

Then this deserves it's own thread.