[ExI] Strong AI Hypothesis: logically flawed

Dan danust2012 at gmail.com
Tue Sep 30 20:26:38 UTC 2014


On Saturday, September 27, 2014 9:18 AM, John Clark <johnkclark at gmail.com> wrote:
On Sat, Sep 27, 2014 at 1:23 AM, Dan <danust2012 at gmail.com> wrote:
>> John, regarding physicalism, you might want to read this:
>> 
>> http://plato.stanford.edu/entries/physicalism/
> 
> The very first words of this is "Physicalism is the thesis that
> everything is physical",

Yes, and then the rest and his book try to examine what that means.

> this was obviously written by a philosopher of mind which explains
> its uselessness. Now if it had said "all nouns are made of matter"
> I would have kept reading, if it had then said that matter is made
> of Bosons and Fermions or if it had said that matter is everything
> except for information I would have kept reading, but it didn't so
> I won't.

The problem is not with the definition though. Just as atomism was the thesis that all things are made of atoms, physicalism means all things are made ultimately of physical stuff. The onus would be on explain what physical stuff is in general as opposed at one point in history. I think there's a strong intuition of what "physical" means (across time) since which many people accept physicalism today. (Or materialism, which seems to be nothing more than the older term for the same thing. With that in mind, Ancient Greek atomism was just the physicalism of its day.)

Even your offer of "Bosons and Fermions" seems to be a variety of physicalism and would depend on what was meant by those terms. I don't see much in the way of rejecting these theses based on this -- based on the notion that calling something "physicalism" doesn't tell us much.

> I love philosophy but philosophers no longer do philosophy,
> scientists and mathematicians do.

I think people in general, scientists and mathematicians too (of the latter I think I can speak with some expertise because that's where my degree is), often do poor philosophy, just as most non-experts in any field do poorly in that field. This doesn't mean their work is to be rejected, but often when they stray beyond the confines of what they know -- and I mean they don't know philosophy; they tend to just go with their intuitions (not always wrong) and what little philosophy they've heard of (hence so many mathematicians adopting some watered down form of Platonism without understanding any of the deeper problems with any form of Platonism). This is little different than when anyone else chimes in on the latest findings in particle physics -- as if they spent time at physics conferences or reading all the journals and were able to hold their own at Fermilab or CERN, as if they were a collaborators in science rather than spectators.

This isn't to defend the whole profession of philosophy today or at any time, but I would dismiss all of it as you seem to do. (And, yeah, I'd separate philosophy from the philosophers, but in the same way I'd separate science from the scientists. No need for heroes or to follow any particular person. The problem for me is just as philosophers quite often do sloppy science, scientists often do sloppy philosophy. So, no, I don't trust the scientist, even the top level ones, to get the philosophy right here. And I do think it's important to understand and get right the latter -- that's it's neither trivial nor something one can ignore.)

>> Also, physicalism isn't trivial by way of being empty -- any more
>> than Ancient Greek atomism was similarly empty.
> 
> I think the ancient Greeks got more credit than they deserved
> for coming up with atomism, after all substances are either
> infinitely divisible or they aren't, Democritus said they
> aren't and Aristotle said they are. One said things were
> continuous and one said things were not, one of them had to
> be right although neither had a scrap of experimental evidence
> to support his guess.

They relied on observation and argument. Not bad tools to start with, and experiments are really ways of trying to control observations to test conclusions, no? But my point was that atomism is not empty even if it had only dubious support given the "science" of the time. By not being empty, I mean it wasn't merely like saying everything is made of "qwd'las'ashdhasjdahdasjdfjasd" without having any idea of what qwd'las'ashdhasjdahdasjdfjasd is. There was a fairly clear idea of what atoms were. (Yes, it wasn't completely clear, but one shouldn't expect infinite precision from the get go if ever.) Physicalism also seems to have a fairly clear idea of what the physical is -- in a way, at least, to distinguish it from its two major rivals today: dualism and idealism. Granted, it still has serious problems.

But my point in bringing up here -- and leaving aside Anders comments, which I agree with -- is it seems the strong AI hypothesis most proponents of AI hold seems to imply physicalism -- in the sense that what human intelligence relies on is how brains physically work -- what goes on in them in terms of electrical and chemical reactions and not on some notion of, say, that there's a mental realm separate from all physical moorings or that everything is really made of ideas and the physical stuff (neurons, biochemical reactions, etc.) is just our misreading an ideal realm. In other words, strong AI can succeed because there's no in principle reason why humans can't simply build non-biological brains. (Okay, there's also a multiple realizability thesis tucked in with this. This is just that intelligence supervenes on processes that can happen in a substrate other than a physical brain. This doesn't go against others' claims here that maybe AI might be better not to try to replicate everything and every function of a human intelligence.)

Regards,

Dan
My latest Kindle book, "Born With Teeth," can be previewed at:
http://www.amazon.com/gp/product/B00N72FBA2  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20140930/bbea85f0/attachment.html>


More information about the extropy-chat mailing list