On Consciousness

[Mr. Scott]

"...if it's true [what Tammet says he does] it blows away scientific theory"

I hate that sort of quote. It's either taken out of context or just plain dumb.

Agreed! It's usually written into scripts to stimulate anticipation of a shattering breakthrough -- a cliffhanger to keep the audience from tuning out during the commercial break. I'll scrutinize the context when I find the time and get back to you on that one.

In the meantime, I think "scientific theory" is pretty safe.

 

[PixyMisa]

I should have worded it more like this:
There are realms of science that remain beyond our present tools of inquiry.

No.

 

[Belz...]

I should have worded it more like this:
There are realms of science that remain beyond our present tools of inquiry.
Examining one's own consciousness; self-inquiry, remains legitimate, yet beyond the reach of our empirical understanding.

QM is relatively new, and it has opened up a whole new approach to how things behave, and why. The experiment is far from wrapped up.

No, really no.

Let's not get QM magic into this.

And there are ways to examine consciousness other than introspection. In fact, introspection is the least reliable one.

 

[Frank Newgent]

From another thread exploring the same topic (and when diesel cost less):

The majority of car engines manufactured in the last decade are conscious.

I have a 6.0 liter Ford diesel and I know what you mean.

Yesterday on the Ohio Turnpike my 2004 Ford E-350 began to question the posted price of diesel fuel as 'text' while subjecting the $2.59/gallon signs to a strenuous critical reevaluation as 'constructs' supplying myriad self-opposed 'meanings' derived from cultural signifiers drawing out assumptions from within the context of our media environment.

On the other hand my '96 Saturn is a Green Bay Packer fan.

 

[PixyMisa]

From another thread exploring the same topic (and when diesel cost less):

Yes. Have you managed to grasp the concept yet?

 

[Fudbucker]

I mentioned Dennett when asked for "an expert in a relevant field": http://www.internationalskeptics.com/forums/showthread.php?postid=8472869#post8472869

His relevant field is cutting through the nonsense propounded by other philosophers.

And of course, not all philosophy is worthless garbage. Apart from Dennett, there's Popper, and Hume, and... I had someone else in mind, but I've forgotten who it was. But at least three.

Hear, Hear!

 

[Frank Newgent]

Yes. Have you managed to grasp the concept yet?

Easily. My Saturn wasn't manufactured in the last decade

I bought it used. Its previous owner apparently ran it into a monolith.

 

[Fudbucker]

I think Pixy might be out on a limb re: conscious cars.

"Are cars conscious"
!No results found for "Are cars conscious".

 

[PixyMisa]

I think Pixy might be out on a limb re: conscious cars.

"Are cars conscious"
!No results found for "Are cars conscious".

Argument ad specious search queries?

Take a look at what is involved in automotive control systems, and try again.

 

[PixyMisa]

Hear, Hear!

Well, good to know we have something in common.

I think Hume is underrated precisely because he got so much right; we don't need to refer to his work because we live in his world. Also, he doesn't have a fancy name like Wittgenstein or Feyerabend or Sartre, when you can sound smart just by not mispronouncing the damn thing.

 

[quarky]

No.

Actually, yes.
Very yes.

Now you're just being obnoxious.
I suspect you know nothing about consciousness.
Is that the ruse?

Perhaps you'd like to learn something.
Give me a holler if you would.

 

[PixyMisa]

Actually, yes.

Hang on... Nope, sorry, still no.

Unless you'd like to present evidence to support your assertions? Hmm?

 

[quarky]

Hang on... Nope, sorry, still no.

Unless you'd like to present evidence to support your assertions? Hmm?

I've seen no evidence presented by you for an extremely weak and pointless argument.

 

[PixyMisa]

I've seen no evidence presented by you for an extremely weak and pointless argument.

I have shown conclusively that computers can be conscious. I have shown that self-referential information processing is what we mean when we talk about consciousness. I have shown that this is widespread in computers right now. I have explained exactly why and how the brain is a computer. All of these make it clear that consciousness is computational.

You have... What? Stamped your foot, if that?

 

[steenkh]

What animals are not self-aware?
Must we now define self?

A sphex wasp? You do not need a definition of self in order to be self-aware: it simply means to have a representation of itself in the internal model of the environment.

I sense a universal self.

And you can argue why this is not an illusion by referencing to some objective fact?

It is special. The universe is special.

Special in what sense? No other instances of it exist? You just happen to like the idea?

 

[!Kaggen]

Yes. Have you managed to grasp the concept yet?

Now vehicle management systems is something I know a few things about having a hobby re-mapping them for rally/track/drift cars.

For instance Subaru's have a neat ECU which uses different maps and 'a learning factor' to adjust ignition timing, fueling and valve timing depending on feedback from the engine knock sensor. From the factory this learning factor is pretty conservative to take into account all the variations in fuel quality and air pressure. It makes their cars very very reliable. With race cars we know at which altitude we are going to race and we know only use certified fuel with a minimum octane rating. So by changing the learning factor to be more aggressive and expanding the range of the fuel, valve and ignition maps the vehicle 'learns' to maximize its performance. What is neat about the system is that it still has the built in safety feature of constantly monitoring engine knocking which then allows adjustments which avoid engine failure. So using the factory ECU with a few adjustments makes the cars still very reliable as track cars.

The problem with this 'learning' system is that if you change the engine internals (pistons, conrods, valve shape/lift), the air flow meter and the intercooler, the standard ECU is useless as its 'learning' algorithm works against the changes to the hardware and defeats the purpose of improving performance with revised internals. It is easier using a fully programmable ECU which allows one to change every setting independently of each other. In other words get rid of the 'learning/SRIP' capabilities.

Just saying

 

[steenkh]

Yes, they're saying specifically that conscious choice is an illusion, that choice happens unconsciously and then is reflected consciously.

But do you agree with me that the part that makes the decisions can be conscious in the same way that a car can be conscious? It is the self-referential part that is taking longer to be informed.

 

[PixyMisa]

Now vehicle management systems is something I know a few things about having a hobby re-mapping them for rally/track/drift cars.

Cool!

For instance Subaru's have a neat ECU which uses different maps and 'a learning factor' to adjust ignition timing, fueling and valve timing depending on feedback from the engine knock sensor. From the factory this learning factor is pretty conservative to take into account all the variations in fuel quality and air pressure. It makes their cars very very reliable. With race cars we know at which altitude we are going to race and we know only use certified fuel with a minimum octane rating. So by changing the learning factor to be more aggressive and expanding the range of the fuel, valve and ignition maps the vehicle 'learns' to maximize its performance. What is neat about the system is that it still has the built in safety feature of constantly monitoring engine knocking which then allows adjustments which avoid engine failure. So using the factory ECU with a few adjustments makes the cars still very reliable as track cars.

Yep. I don't know if you've had to deal with this, but the point I'm making with respect to car engines and consciousness is that the ECU will also be monitoring itself (or it will have a secondary watchdog chip). So it's not only aware of (and constantly tweaking) all the things going on in the engine, it's also aware to some degree of its own activity. The last thing you want is for the engine to be perfectly fine but for the ECU to crash.

The problem with this 'learning' system is that if you change the engine internals (pistons, conrods, valve shape/lift), the air flow meter and the intercooler, the standard ECU is useless as its 'learning' algorithm works against the changes to the hardware and defeats the purpose of improving performance with revised internals. It is easier using a fully programmable ECU which allows one to change every setting independently of each other. In other words get rid of the 'learning/SRIP' capabilities.

Just saying

Yeah, adding SRIP to a system greatly increases the scope of possible behaviours, which makes it much better at adapting to new conditions to achieve its goals, but what you really want to do here is change the goals themselves. I don't know if you can tweak the learning algorithm in this case, and even in cases when you can it's often easier to just hard-code the settings as you've done.

 

[PixyMisa]

But do you agree with me that the part that makes the decisions can be conscious in the same way that a car can be conscious? It is the self-referential part that is taking longer to be informed.

Yes, that seems not only plausible but (given the organisation of the human brain) highly probable.

 

[!Kaggen]

Yep. I don't know if you've had to deal with this, but the point I'm making with respect to car engines and consciousness is that the ECU will also be monitoring itself (or it will have a secondary watchdog chip). So it's not only aware of (and constantly tweaking) all the things going on in the engine, it's also aware to some degree of its own activity. The last thing you want is for the engine to be perfectly fine but for the ECU to crash.

The point I was making is that the ECU is only as useful with respect to SRIP as the hardware it was designed to run. The ECU adds the possibility of SRIP to the vehicle, but this SRIP is dependent on the hardware of the vehicle(engine and its peripheries). If you change the hardware the SRIP is useless.

What we see with the brain is that even when we change the hardware the brain is not useless. It can even use hardware for one application to perform another.
http://www.scientificamerican.com/article.cfm?id=device-lets-blind-see-with-tongues

So what we need is an ECU that can be plugged into any car that will allow one to drive it from A to B even without wheels.

Oh and the ECU's do crash. Though not as often as the vehicles.

Yeah, adding SRIP to a system greatly increases the scope of possible behaviours, which makes it much better at adapting to new conditions to achieve its goals, but what you really want to do here is change the goals themselves. I don't know if you can tweak the learning algorithm in this case, and even in cases when you can it's often easier to just hard-code the settings as you've done.

We do change the goals by expanding the parameters within which the algorithm works. But the algorithm is not designed for every possible change in hardware. It was developed through many hours of testing using the standard hardware.

Modern race factory teams (F1 and WRC) certainly do produce many algorithms for each car set-up and using huge budgets for relentless track testing.

I look forward to the day when F1 teams have to develop ECU's which are then pooled, mixed up and then randomly handed out to teams to use in their vehicles just before a race.