This atheist website offers a refutation of the AFR, going not from actually reading my book, but from the Wikipedia commentary.
42 comments:
Anonymous
said...
"I didn’t buy the book, but I wanted to write a post about the Argument from Reason. The argument from reason was pretty aptly described by Wikipedia, so I will now quote shamelessly from it"
If she didn't buy or read the book, then how could she conclude that the wikipedia entry 'aptly described' the AFR?
"Even though the argument from reason may sound good, the argument is a good example of begging the question (I.E. circular reasoning). The premise (that physical sources cannot constitute a rational source) is the conclusion (That naturalism - which says physical sources can constitute a rational source - is wrong). The reason that this is hard to see initially is due to the way in which the argument is laid out."
Setting aside Ziztur's reworking of the argument for the moment, it seems to me as if she doesn't understand that all deductive arguments contain the conclusion in their premises. She also doesn't understand 'begging the question' very well, since the so-called instance of it she points out is anything but. That's pretty embarrassing: she doesn't understand what a deductive argument is, she set up a strawman, *and* she failed to knock it down!
Doctor Logic, the point is that an argument that concludes that naturalism is false isn't question begging if it uses that premise. That was her claim. That's quite a different issue from the one you're raising, viz. whether the premise can be supported.
The "gaps and incredulity" charge is inane. A person who points out that a housecat cannot lay an egg which will hatch into a tyrannosaurus rex is relying on "gaps and incredulity". They can hardly do otherwise no matter how much science or (good) philosophical argument is trotted out, because the point of their argument *IS* that there is a gap, and that this gap is unbridgeable. And if the gap really is unbridgeable, well then - it makes sense to be incredulous about the claim, now doesn't it?
The aspiring doctor may as well defend idealism or panpsychism or full-blown cartesian dualism on the grounds that all the opponents have to support their claims that these views are wrong amounts to "gaps and incredulity".
Sorry, the AfR needs a lot more than that. There's a reason why there have been moves in recent years to bless those who subscribe to Chalmers' views (so long as they don't utter the dreadful word 'soul') as naturalistic explanations despite his open hostility to materialism. And why there's been a growing warmth to panpsychism among naturalists as well. (And what did Strawson call panpsychism? Real materialism, of course.)
Incidentally, the comments on that site are atrocious. The idea that Victor argues that if the AfR is correct then God exists (or, worse yet, that if the AfR is correct then Jesus is God) is downright silly. Does anyone realize that naturalism can be false without Christianity (or even theism, certainly classical theism) being true? Or that Christianity is not the only alternative to naturalism? (I certainly believe Christianity is right, but still, the playing field is a bit broader than THAT.)
A person who points out that a housecat cannot lay an egg which will hatch into a tyrannosaurus rex is relying on "gaps and incredulity".
No, they're not. They're relying on induction. And their claim would be that it is extremely improbable that a house cat would lay a T-Rex egg.
The AfR isn't like that at all. What examples do AfR proponents hold up for their gap/incredulity arguments? Deep Blue.
Is Deep Blue supposed to show that mechanisms cannot think about things? That would be like arguing that house cats aren't material because rocks don't give birth. That's poor inductive reasoning because the behavior of matter varies with its structure. Likewise, Deep Blue does not have structure or behavior similar to a conscious human brain. For starters, Deep Blue does not do abstraction, nor does it employ propositions.
The main problem with the AfR is that it fails to define what intentionality actually is. After all, if you could define it, it would be pretty straightforward to implement it in a machine.
So... simple question: Imagine yourself thinking a thought that lacked intentionality. How would you know that this thought lacked intentionality? Obviously, the thought would seem confused, but in what way?
Did the argument from reason originate in "C.S. Lewis’ Dangerous Idea"? Sources seem to indicate that the argument from reason was developed by C.S. Lewis and NOT by Reppert and has been developed by several individuals, and thus I do not need to read Repperts book in order to understand it - I need to read C.S. Lewis (which I have). I used Reppert's book in my blog post because I had seen his book in a store and it made me think of the argument from reason, given that it mentioned it on the back cover.
Similarly, I do not need to consult a particular contemporary writer who expounds upon the ontological argument in order to understand the ontological argument, given that the contemporary writer is not the first to use it, especially if I can find multiple sources that lay the argument out. I used the Wiki version of the argument after reading several versions and concluding that it was decent enough to quote from.
If Dangerous Idea contains a more coherent formulation of the AFR that is missed in the various sources I consulted (including Lewis), then perhaps you could explain exactly why my sources are inadequate?
In mentioning specifically that I offered a refutation of the AFR going from Wiki and not your book, you are making an ad hominem attack - you are attacking my presentation of the argument rather than my counter-argument. It would be more persuasive to provide a counter argument, rather than simply imply that I have not done my homework.
Is my presentation inadequate? If so, please correct me, especially in the comments section of my blog, so that my commenters will know I am mistaken. If my presentation is adequate, then my source for the argument is irrelevant.
Can you explain why I don't understand "begging the question", rather than simply stating I do not understand it?
Can you please explain how a deductive argument can be useful if the conclusion is "contained" in the premise? To my knowledge, a deductive argument is one in which it is thought that the premises provide a guarantee of the truth of the conclusion. In this way, the premises are given as support for a conclusion in such a way that if the premises are true than the conclusion must be true.
One could describe a circular argument as deductive, but not all deductive arguments are circular (I.E. contain the premise in the conclusion) and a deductive argument which is circular fails precisely because the premise contains the conclusion.
@Anonymous - the commenter was being sarcastic with the AFR = Jesus comment. We get a little rowdy over there, but certainly people use arguments such as the AFR as a springboard to lead to Christianity, much like Lewis does with the argument from morality (among others) in Mere Christianity.
As I said, no matter how much science or philosophy is used in such a case (induction would clearly, for our purposes, fall under the former in this discussion), in the end what is had are 'gaps and incredulity'. Use as much induction as you please. All the induction will illustrate, and all it can possibly illustrate, are "gaps and incredulity". A gap, because it would be argued there is no detailed explanation for a housecat to become oviparous and lay an egg, especially an egg for a creature outside its species. Incredulity for obvious reasons (just plain unthinkable and seems superficially unlikely, etc.)
In other words, griping that all the the AfR proponents have on their side are "gaps and incredulity" is weak.
Second, your reference to Deep Blue is misleading. AfR proponents can and do use DB to illustrate some of what they mean - but you know what else is available to them? The human brain itself, associated neuroscience, and the fundamental commitments of materialism. Let's go ahead and talk about neurons, about the effects of harm or stimulation on the human brain, on modeling techniques, etc, all we please. Let's entertain the purely speculative notions of what "future science" may reveal according to materialist philosophers of mind, considerations of just how reason and intentionality (qualia aside) may be accounted for in the brain, without these things ultimately being defined away as illusions, illicitly smuggling in the very things they're attempting to explain, or leaving the questions ultimately unanswered anyway. AfR proponents are more than happy to engage all of these things, because they serve as beautiful illustrations of their point.
What's more, you just played a little bit of sleight of hand with Deep Blue. Why isn't Deep Blue (or any computer, really) capable of reason? Because DB's structure differs from a human brain of course. Wonderful. Care to explain this structure that is capable of explaining reason/intentionality in purely material terms? And if you're lacking purely scientific evidence here, would you care to tell us what such an explanation could even look like in theory?
As I see it, the AFR does not depend on the premise that 'rationality cannot arise out of non-rationality'. Non-rationality can, potentially, give rise to anything, just as a true and valid conclusion can be a result of false, even meaningless, premises.
The problem is that we cannot trust what is produced by non-rationality. It may be reliable, it may not. What reason do we have to think that the physical events in our brains give us a true picture of the universe? They may, they may not.
I wish Dr. Reppert would take part in this conversation
kbrowne, I don't think an AfR proponent would dispute that some non-rational source can (for example) by sheer coincidence string together words which could translate to an accurate statement. It's rather like admitting that a roomful of monkeys, given infinite time, may be able to produce the complete works of Shakespeare. But just as that sort of process isn't at all like what we think Shakespeare went through when he wrote his plays and sonnets, some non-rational source now and then getting something right isn't really the same as reason. In both cases all you're getting is a happy coincidence. And if reason is nothing but happy coincidences produced from blind, purposeless forces, then reason it ain't. (And if reason is more than happy coincidences from blind, purposeless forces, then naturalism has some splainin' to do.)
I also see the same problems apply to supernaturality or theism. Even if one were to potentially argue that naturalism leads to an untrustworthy source of rationality, positing a supernatural source for rationality will have the same and possibly more complex problems.
I think it would be best if we continued this discussion in one place rather than two, so I will stick to my blog since it is easier to keep track and respond to comments there. I'll still check back here from time to time, but if you see silence on my end it is not because I have nothing to say.
You appear unfamiliar with the meaning of "gaps argument" or "argument from personal incredulity".
They are both arguments from ignorance.
Care to explain this structure that is capable of explaining reason/intentionality in purely material terms? And if you're lacking purely scientific evidence here, would you care to tell us what such an explanation could even look like in theory?
You ignored my challenge, but I'm happy to explain for you.
What would it mean for you to have a thought that lacked intentionality?
It would mean that a proposition in your mind didn't mean anything, or didn't mean what you thought it meant.
Consider the proposition "the cat is on the mat". I know what this means to me because I know what experiences or sensations correspond to the proposition. I know the terms in the proposition for the same reason. The proposition means (in a simple case) that my cat recognition faculty and my mat recognition faculty will both fire in coincidence along with my "X on top of Y" recognition faculty. Or something along those lines.
Now, in order for this to happen physically, there needs to be a circuit in my brain that represents the abstraction for cat. As it happens, there is such a thing. It's called a hierarchical temporal memory, and the human neocortex is packed with them. You can train such memories to recognize cats in general by exposing them to a single cat. Once trained, the memory doesn't just recognize the sample cat, but other instances of cats, cats in other colors, cats of differing sizes and orientations, cats with missing elements (like a missing eye or leg), etc. In other words, an HTM trained to recognize cats is capable of recognizing cats that do not actually exist, were such fictitious cats to be placed at its inputs.
So, if the components of a proposition refer to HTMs, they refer to precisely the kinds of abstractions necessary for meaning in terms of past experiences. I know what "cat" is about because cat references a circuit in my brain that recognizes cat-like objects. Essentially, the circuit is like a universal for cats.
BTW, HTM's need not only be connected to sensory inputs, but can be connected with other HTM's. That is, there can also be abstract thoughts, abstractions of abstractions, or abstractions that cross between abstractions and sensory inputs. It's straightforward to see how there can be abstractions for logical manipulations.
As we know, DB does not possess HTM's, does not think in terms of propositions, and lacks abstractions. That is, DB is missing the ingredients of intentionality.
You and other dualists may object to this sort of analysis because you haven't defined what meaning is, and have no intention of doing so (no pun intended). If meaning is a form of magic that cannot be defined, then intentionality has to be a form of magic, too. However, your own refusal to define meaning would not be a good argument against my materialist definition.
I'm familiar with both "arguments" in the special little score-book of debating rules. As I said, your citing of them in this context is ridiculous - and falsely implies that critics of materialism aren't interacting with neuroscience, materialist arguments, first-person experience, etc on this topic. And one more time, that "gaps and incredulity" are what all of their arguments ultimately point to is one weak criticism. "Gaps and incredulity" are all materialists have against dualism. The question is what arguments and reasons they cite for their belief that such gaps are problematic, and for their reasons to be incredulous. That, critics of materialism (note: not the same thing as dualists) have in spades.
Second, fantastic idea! One problem. What does it mean for there being a "circuit in the brain that represents the abstraction for cat"? Represents to who or what - another part of the brain? How is this representation accomplished? What you've offered here is a nice prospective model for building/programming a theoretical computer that appears human-like to us. Valuable stuff, to be sure. Not what we're after on the topic of intentionality. What's more, if you're saying that 'aboutness' is real - that pattern X in the brain is objectively and certainly "about 'the cat is on the mat'" and that information is a real and certain part of the universe rather than a useful, if fictional model.. well done, citizen. You've buried materialism, and have opened the doors to hylemorphic dualism all over again! To make "information" a real constituent of the universe is to indirectly support the AfR, not fight it.
Third, I'm an AI fan - so I decided to google around for hierarchical temporal memory. I stumbled across the conversation you had over at New Empiricism - and I have to say, this guy gave a reply I like. I'll quote.
If your conscious experience is like looking at a photo at any instant then we will need to introduce extension in time into our reasoning (see Time and conscious experience), otherwise we will be no more than a frozen set of objects and incapable of observation. Once our physical theories include extension in time we have left the materialist paradigm and mind as simple machine.
...
What you are doing with the machine model of mind is looking at your machine with your human mind and using this to join up the stages of the machine into a continuous picture. Sure your mind can extend over time but your machine is just a jerky procession of frozen patterns. When a given frozen pattern is present at the end of a clock pulse the last pattern has entirely gone.
...
To use Ryle's terminology, equating conscious experience with processing or functions is a category mistake.
Apt observations. Loosely related of course, but hey, intentionality questions are loosely related to the AfR too.
Anonymous, for my most developed view on neuronal representational systems, follow the links in this post. I've thought a lot about this, researched it empirically, written papers implicitly about it as a neuroscientist. If you read them and have questions or criticisms, I'd be happy to respond.
IT took me a half hour to find that @#$# link. I think I need to make a post at my own blog with all the relevant links, so I never have to look for it again. How annoying. Grrrr... Waking up in 6 hours to go running. :O
Thanks. Interesting. Not really what I'm going after here (what's related to my response to DL doesn't offer anything new), but yet another fun resource to read up on for different topics.
Anon, it should be relevant to: >>What does it mean for there being a "circuit in the brain that represents the abstraction for cat"? Represents to who or what - another part of the brain? How is this representation accomplished?"<<
What does it mean for there being a "circuit in the brain that represents the abstraction for cat"?
My bad. It doesn't represent an abstraction. It IS an abstraction.
What's more, if you're saying that 'aboutness' is real - that pattern X in the brain is objectively and certainly "about 'the cat is on the mat'" and that information is a real and certain part of the universe rather than a useful, if fictional model.. well done, citizen.
No, the physical configuration in the brain isn't objectively about the cat in the sense that it lacks subjectivity. It objectively represents cats, is an abstraction of cat, to the subject that owns the brain. It is a structure that was formed by interaction with a cat, and a structure which will be triggered into similar activity when supplied with inputs resembling the cat.
To make "information" a real constituent of the universe is to indirectly support the AfR, not fight it.
This is question-begging. You're making an assumption about information. You're assuming that information is non-natural. Information is defined based on our own subjective perspective. It's not defined to be non-natural, so such an assumption is unwarranted.
Yep, and I didn't see that explained in your links. By all means, if you care to explain that, go for it!
DL,
My bad. It doesn't represent an abstraction. It IS an abstraction.
Fantastic: The pattern IS an abstraction. Great. So is a sentence containing an abstract proposition. As I asked, by what in the brain is this recognized? How is this accomplished? That a material medium can embody an abstraction (The words "The cat is on the mat" on a piece of paper) is not itself a mystery to me. The question is how this becomes an interpretation, a conceptualization in the/a mind - how is it grasped? A TXT file on a computer with the words "The cat is on the mat" in ASCII means diddly here.
No, the physical configuration in the brain isn't objectively about the cat in the sense that it lacks subjectivity. It objectively represents cats, is an abstraction of cat, to the subject that owns the brain. It is a structure that was formed by interaction with a cat, and a structure which will be triggered into similar activity when supplied with inputs resembling the cat.
Who said 'lacks subjectivity'? You're saying it "objectively represents cats" - which I can only take to mean is you declaring that information is a real constituent of the universe, not some useful fiction. Wonderful, I agree with you! You just sank materialism, and you're buying in to the AfR.
Unless you're saying that information doesn't "really" exist, and all you're describing here is some mechanical process that you're certain will explain the *appearance of* abstract thought, etc in a machine (in which case you're not talking about anything relevant) and eventually humans (in which case, pal, this ain't reason you're talking about. Ain't properly intention either.)
This is question-begging. You're making an assumption about information. You're assuming that information is non-natural. Information is defined based on our own subjective perspective. It's not defined to be non-natural, so such an assumption is unwarranted.
No, I'm assuming that information is non-material - and it is, even if substrates are involved. The AfR is aimed squarely at materialism; I could really care less if you suddenly want to define non-material existences, or even full-blown dualism(s), as natural. Go right ahead! Personally, I'd love to define hylemorphic dualism as natural. I'd love to define God as natural! Look at me, I'm an orthodox Catholic naturalist.
Now, I know from the thread over at New Empiricism that you think non-material/non-physical things can be filed under "naturalism". That's great, I don't even feel like fighting you on that. But if you're walking down that road here - if you're admitting that information/non-physical/non-material things is/are a real and actual constituent of our universe - then DL, you've left materialism behind. And that's central to the AfR's aim.
Don't worry, DL. You don't have say that God exists just because the AfR worked in that capacity. Victor has flatly said more than once that it mostly moves you to various other options in metaphysics and philosophy. But don't dig your heels in the ground and refuse to give credit where credit is due (assuming you're admitting these things are real, of course) just because a *gasp* christian or *GASP* apologist came up with a good argument. There are atheists out there who reject materialism. I'm sure they'll welcome you among their number. (Bring pizza. And wear a Discworld T-shirt.)
Anon: hmmm, perhaps you missed the links. For instance, this post entitled 'biorepresentations.'
Doctor Logic, that Numenta stuff is interesting theoretical engineering, but not particularly cutting edge or accepted as brain theory. I was not impressed by Hawkins talk on it at Society for Neuroscience. That said, it provides an interesting possibility argument.
Well the reason rationality cannot arise out of non-rationality is that rational effects cannot be produced by non-rational causes, and if whatever one designates as rational arises out of non-rational causes, then it is the effect of a non-rational cause and therefore non-rational as well.
Fantastic: The pattern IS an abstraction. Great. So is a sentence containing an abstract proposition. As I asked, by what in the brain is this recognized?
If I have a simple machine that generates 5 volts output in the presence of cyanide, and 0 volts otherwise, then this machine "recognizes" cyanide. The simple machine will have neither abstraction nor consciousness, but that's not important. It still recognizes.
If I create a simple HTM that recognizes cats, that HTM will not just recognize, but also *be* an abstraction for cats. It recognizes all cats, including the ones with eye patches and peg legs. Because HTM's learn, the simple HTM is abstracting (verb) from its sample training cat to the class of all cats (or most of them). A simple HTM will lack consciousness of its recognition, and lack consciousness of its abstraction. However, it both recognizes and abstracts. Humans would differ because they not only recognize and abstract from examples, but they are conscious and aware that they are doing this.
So, the primary difference between simple HTMs and humans is that humans also recognize and abstract their recognition and abstraction abilities. How could a machine make that work?
Well, we can imagine that an HTM that finds patterns in its own internal processes of recognition and abstraction, and which will learn to recognize when recognition and abstraction are taking place. Moreover, it will be possible for this meta-HTM to recognize when a cat is being recognized versus when a dog is being recognized.
The question is how this becomes an interpretation, a conceptualization in the/a mind - how is it grasped?
How do you grasp ideas for yourself? Consider soft velvet. When you think about this term, you think of it in terms of past experiences. If you had no sense of touch, you would not really know what soft velvet was. You would know it only by its appearance to your other senses. If you lacked both sight and touch, the term loses meaning even further. Soft velvet is defined in terms of the inputs to your recognition and abstraction processes. When you think of soft velvet, you can feel soft velvet with your fingertips, albeit less vividly than if you were actually touching velvet. Thinking the concept of soft velvet causes your mind to react as if you were actually touching velvet.
This is precisely what you would expect if you brain was full of HTM's. An HTM is a memory and a recognizer. It contains feedback loops, so that retrieving a memory is similar to an abstract rehash of all prior training experiences.
Meaning is in terms of past experiences.
The same applies to mental abstractions. An HTM that has other HTMs as inputs can find patterns in patterns, abstractions in abstractions. The meaning of a higher-level abstraction is in terms of lower level ones.
I'm assuming that information is non-material - and it is, even if substrates are involved. The AfR is aimed squarely at materialism;
You're taking a childish description of physicalism to an utterly absurd extreme.
Suppose that the universe is described by particle physics, and I have three Uranium atoms positioned such that they are equidistant from each other. This forms a triangle. Is the triangle real? By your definition, the triangle is non-material and not real to the materialist. This is nonsense. To a materialist, there is a triangle there. A triangle is a description of a configuration. It does no good to claim that, in descriptions of physics, there are no configurations, no abstract behaviors, etc.
Even if you were to say that descriptions of physical reality are not real, our own descriptions of physical reality would become real if we are physical beings.
Indeed, by your reasoning, there's no such thing as a planet in materialism. You practice some form of ultra-greedy reductionism in which a planet no longer exists because the planet is made up of quarks and leptons. But there's a difference between 10^30 quarks and a planet. A particular planet is a particular arrangement of those quarks. The behaviors of the planet depend on those arrangements. The planet is the word for that particular arrangement.
Indeed, by your definition, there aren't even laws. Consider Galileo's discovery that two different masses fall to Earth at the same rate (neglecting air resistance). Is this a law? Well, we would have to be talking about abstractions of mass, and abstractions of their initial position. We could be dropping the masses from the Tower of Pisa or from the Eiffel Tower, or dropping them from a tower we have yet to design and build. Furthermore, we would have to abstract the planet or gravitational field in which the masses are dropped. So are laws not allowed under physicalism? That would be a pretty peculiar physicalism, would it not?
Consider mathematics. By your definition, mathematics does not exist under materialism. Yet mathematics is a abstraction about computation. Mathematical systems are abstractions of systems. They're like physical laws of computing.
The idea that a mind is a physical system is the idea that a mechanism can create abstractions and think the way we humans are observed to think. If you could prove that machines could not create abstractions, then you would have a case against naturalism, but you don't have that proof. Indeed, the evidence is that not only can machines abstract, but that our brains are instances of such machines. (If dualism were correct, there would be no need for brains with memory or abstracting ability, nor need for a central nervous system at all.)
You have an argument from ignorance, i.e., "I can't imagine how a machine could think."
Getting back to your idea of information... Of course information exists under materialism. Suppose I have a configuration of atoms in a lattice, and their spins are either up or down. There is information in the configuration, i.e., it takes a certain number of bits to represent this.
There are two kinds of information. There's information in the sense of "how many bits does it take to record a configuration?", and also in the sense of "how much of this information has utility to me?" It takes a lot of bits to record the configuration of pixels in a black and white photo containing nothing but randomly black or white pixels. However, that photo has no utility to us, so it contains no information in the second sense. Even if you take information to be defined in the latter sense, there is still meaning to information in a physical world.
But don't dig your heels in the ground and refuse to give credit where credit is due (assuming you're admitting these things are real, of course) just because a *gasp* christian or *GASP* apologist came up with a good argument.
I wonder, could you possibly be more arrogant in your responses?
You're taking a childish description of physicalism to an utterly absurd extreme.
No, DL, I'm pointing out the utterly absurd extremes to which a consistent materialism leads. Have you ever thought that maybe - just maybe - materialism is absurd? And that maybe, just maybe, if your chosen way to demonstrate that materialism isn't absurd is to take advantage of slippery definitions to define materialism to include what amounts to hylomorphic dualism.. that maybe you're helping to demonstrate its absurdity?
You know that materialism and physicalism are not strictly the same things ("Slippery" as they may be), so shame on you for trying to substitute one for the other so casually. Double shame on you for purposefully obfuscating and switching in "naturalism", "materialism" and "physicalism" so interchangeably.
Anyway, let's pull out a representative sample of your materialism "defenses" here.
Consider mathematics. By your definition, mathematics does not exist under materialism.
Do you want to know what your biggest problem here is, DL? You seem to think that if materialism ever leads to utterly inane conclusions, that somehow the definition of materialism being used must therefore be wrong and dishonest. Because it can't possibly be that materialism is utter bunk and ridiculous, can it? So instead you try some bizarre debating tae kwon do where if a reductio ad absurdum demonstrates materialism to be absurd, what it actually demonstrates is that materialism is actually something else. Convenient!
I mean, to hear you put it, there's actually no such thing as a philosopher (much less a materialist philosopher) who explicitly argues for eliminative materialism, or for mereological nihilism, etc. And further, there's no possible way for someone's metaphysics to, when taken to its logical conclusions, reduce to positions like these. No, absurdity is only possible for philosophies attached to intellectual positions you disagree with, right?
The meaning of a higher-level abstraction is in terms of lower level ones.
I ask how someone grasps an abstract concept in their mind, you give a reply that explains how a machine can cycle through processes and give reactions to events that could, from our interpretation, make reference to what we take to be an abstract concept. But does the machine know what it's doing? Does it grasp it even if it has been trained to output "The cat is on the mat" when it's presented with a picture of a cat on a mat, and "The cat is not on the mat" when presented with a picture of same? You provide "meaning" as a given for the machine to have (it's just in terms of 'lower level' ones), but the problem is getting any meaning at all (back to the Chinese room.) And again, even for the machine - if you are really and truly a realist about information - adios materialism anyway.
Getting back to your idea of information... Of course information exists under materialism.
You can argue that information really exists - that it is an objective and real constituent of the universe, not just part of some nominal story made up for the purposes of pragmatic explanation. But if it's objective and real, then there is more to the universe than "colorless, odorless, featureless, totally homogeneous particles advancing in perfectly predicable lines with all the rigid and absolute predictability of the premises of a syllogism".. and therefore you reject materialism.
(If dualism were correct, there would be no need for brains with memory or abstracting ability, nor need for a central nervous system at all.)
Utterly incorrect even on Cartesian dualism. On hylomorphic dualism you could not be more wrong, since proponents explicitly argue for the presence and role of bodily systems for sensory operation (no sight without eyes, etc). Wait, let me guess: Hylomorphic dualism is just physicalism now, right? L-O-L.
I wonder, could you possibly be more arrogant in your responses?
I too am confused about this statement > You wrote: If I have a simple machine that generates 5 volts output in the presence of cyanide, and 0 volts otherwise, then this machine "recognizes" cyanide. The simple machine will have neither abstraction nor consciousness, but that's not important. It still recognizes.
"I suppose that sooner or later the physicists will complete the catalogue they’ve been compiling of the ultimate and irreducible properties of things. When they do, the likes of spin, charm, and charge will perhaps appear on their list. But aboutness surely won’t; intentionality simply doesn’t go that deep."
That's Jerry Fodor.
So do you agree or disagree with Fodor?
Is "forms a triangle" for your three uranium atoms an ultimate and irreducible property?
But does the machine know what it's doing? Does it grasp it even if it has been trained to output "The cat is on the mat" when it's presented with a picture of a cat on a mat, and "The cat is not on the mat" when presented with a picture of same?
Yes. As I've explained, the machine does not simply contain what is necessary to recognize "cats on mats." It has what is necessary to recognize its own recognizing of cats on mats, as well as the ability to abstract to "X on mats".
You provide "meaning" as a given for the machine to have (it's just in terms of 'lower level' ones), but the problem is getting any meaning at all (back to the Chinese room.)
The Chinese Room argument proves that individual neurons don't understand Chinese. It doesn't have anything at all to say about what brains do or do not understand. The Chinese Room argument is a simulation of a brain as it operates one neuron (or one subnetwork) at a time (as if the simulation were running on a single processor computer). The fact that an individual neuron doesn't understand Chinese is pretty irrelevant to cognitive science.
Again we're back at personal incredulity as an argument against AI/materialism.
But if it's objective and real, then there is more to the universe than "colorless, odorless, featureless, totally homogeneous particles advancing in perfectly predicable lines with all the rigid and absolute predictability of the premises of a syllogism".. and therefore you reject materialism.
You suggest that for something to be objective and real, it has to be irreducible. Well, that's a very peculiar definition of objective and real. I quite expect that in almost any philosophy there are things that are composite and real. Even in mereological nihiism (which is a very silly idea) arrangements of simples are real and objective.
Utterly incorrect even on Cartesian dualism. On hylomorphic dualism you could not be more wrong, since proponents explicitly argue for the presence and role of bodily systems for sensory operation (no sight without eyes, etc).
Oh, I'm sure dualists reject the idea *now*. But 500 years ago, if you theorized that mental faculties were non-material, there are lots of mental faculties that could be implemented non-materially with no support from the physical brain. Now that we know the brain is capable of doing everything by its onesy, the non-material part is more elusive than ever. Your situation is analogous to that of the conspiracy theorist. No matter how much hard evidence piles up in favor of materialism (or physicalism, because they're both equally deadly to your superstitious religious attachments), you keep pointing to gaps to support your case. You want it both ways. You want to interpret all the latest evidence against the defendant as merely expected evidence for an invisible conspirator who framed the defendant.
I mean, to hear you put it, there's actually no such thing as a philosopher (much less a materialist philosopher) who explicitly argues for eliminative materialism, or for mereological nihilism, etc.
Yes, I've heard of such things, but these are fringe ideas to say the least. They are far from mainstream. If that's what you're arguing against, save your breath/keystrokes.
"I suppose that sooner or later the physicists will complete the catalogue they’ve been compiling of the ultimate and irreducible properties of things. When they do, the likes of spin, charm, and charge will perhaps appear on their list. But aboutness surely won’t; intentionality simply doesn’t go that deep."
That's Jerry Fodor.
So do you agree or disagree with Fodor?
Sure I agree. I never said aboutness was irreducible. Why should it be? It is the dualist's view that aboutness is irreducible, not the materialist/physicalist's view.
If a mechanism responds only to a particular pattern in its input, then it recognizes that pattern, by definition. By this definition, hemoglobin recognizes oxygen.
I've also explained that when humans recognize something there's generally more than just recognition taking place. Humans have recognition, abstraction, and recognition of our own recognition. This is something that the simplest mechanisms (like hemoglobin) lack, but which more sophisticated mechanisms possess.
Hemoglobin does not learn to abstract, and does not recognize that it recognizes (because it cannot abstract).
There are different senses of 'recognize.' One sense DL is using, and another which requires conscious recognition. E.g., you see a friend's face but he is wearing a hat, so it takes two seconds before you have the 'click' of recognition. In AI, pattern recognition is simply the ability to reliably discriminate different patterns. For instance, the 'facial recognition' software can match a novel picture of Jill to a stored image template of Jill, and give a 'Jill' response (and can also do the same for lots of other face pictures). This involves a stored memory of Jill's face, and a comparison of this template with present data. (Of course there are other face recognition algorithms, this is just one type).
Clearly, such facial recognition software isn't conscious of faces, doesn't have that 'Aha there's Jill' experience that we would have.
The interesting question is how far such gizmos (implemented in brains) can get you toward full intentional contents and also toward consciousness (as I argue in the sites linked to above, I think they get you pretty close to the former, but it is less obvious how far they bring you to the latter. That is, it isn't obvious how to bridge the gap between recognition widgets and conscious states).
I agree. It may not be obvious how to get to consciousness, but there's nothing to suggest we can't. All the evidence points to the likelihood that we can. We know humans are machines, and we know humans can do it.
Now, if dualists could prove we weren't machines, they would have a case, but they can't prove it. I mean, it's not like there's astral projection or telepathy which would provide a compelling case for dualism.
This is why dualism is such a total sham. Dualists now admit that in every scientifically detectable way we're machines, yet they still insist that machines can't possibly think like we do. That's like saying that in every scientifically detectable way, the Apollo moon landings will appear to have occurred, but actually they were faked. (Oh, and, furthermore, we can never land on the moon!) Dualism is a ridiculous conspiracy theory.
DL: I agree to some degree but think you overstate things a bit. I said at another blog (the rest is pure quote):
[N]aturalists need a positive story, a biological mechanism to point toward and say 'See, that's how cognition works.' In arguments about phenotypes whose phylogeny is unclear, at least we can point to natural selection as a plausible mechanism. There is, as yet, no 'standard theory' of the biology of cognition, no default consensus mechanism we can point toward.
Because of this, even naturalists are presently forced to be speculative when it comes to specific mechanisms of cognition. Hence, many dualists believe they are on equal footing with the naturalists. "Hey, if the naturalists have no consensus, not even a plausible story, about how consciousness emerges from brain matter, then why should we take their word for it? Indeed, here are some arguments that cognition/consciousness cannot emerge from brain matter. Much less than leaving us on an equal footing, that puts me one up on these silly naturalists that can't even keep their stories straight!"
Such lines of thought are not easy [for the naturalist] to rebut. I think ultimately, after giving the general reasons for thinking that mental states are brain states, we have to say something like: 'In the absence of compelling data a good scientist is less confident, not more confident, in her conclusions. In that spirit, I'll only say that time will tell which general approach is more plausible.'
I said: " I think ultimately, after giving the general reasons for thinking that mental states are brain states"
This is nontrivial, because there are so many data to discuss, and so many reasons to believe the mind is neural. Unlike the antinaturalists, who tend to be one-hit wonders (basically repeating versions of Leibniz's windmill), the naturalists are sticklers for detail and evidence, and have lots of experiments they like to chew on (the antinaturalists should be chewing over the same experiments, but for some reason they like to stay in Leibniz's mill staring about in awe, feeding on their intuitions about what neuroscience will be like in 300 years).
I'm writing this up now but it is quickly expanding into a book unfortunately.
(for his mill example see Section 17 of the Monadology:
"Moreover, it must be confessed that perception and that which depends upon it are inexplicable on mechanical grounds, that is to say, by means of figures and motions. And supposing there were a machine, so constructed as to think, feel, and have perception, it might be conceived as increased in size, while keeping the same proportions, so that one might go into it as into a mill. That being so, we should, on examining its interior, find only parts which work one upon another, and never anything by which to explain a perception.").
It won't be long before we have arguments about whether these things created in labs are alive. Then the fights will be about whether the conditions in which such life is created are enough like those of prebiotic Earth. Then the fights will be over and another gap will be closed. :)
42 comments:
"I didn’t buy the book, but I wanted to write a post about the Argument from Reason.
The argument from reason was pretty aptly described by Wikipedia, so I will now quote shamelessly from it"
If she didn't buy or read the book, then how could she conclude that the wikipedia entry 'aptly described' the AFR?
"Even though the argument from reason may sound good, the argument is a good example of begging the question (I.E. circular reasoning). The premise (that physical sources cannot constitute a rational source) is the conclusion (That naturalism - which says physical sources can constitute a rational source - is wrong). The reason that this is hard to see initially is due to the way in which the argument is laid out."
Setting aside Ziztur's reworking of the argument for the moment, it seems to me as if she doesn't understand that all deductive arguments contain the conclusion in their premises. She also doesn't understand 'begging the question' very well, since the so-called instance of it she points out is anything but. That's pretty embarrassing: she doesn't understand what a deductive argument is, she set up a strawman, *and* she failed to knock it down!
Ziztur is right on target. One of the premises of the AfR is indeed that "rationality cannot arise out of non-rationality."
With what is this premise supported? At best, it's supported by gaps and incredulity.
Doctor Logic, the point is that an argument that concludes that naturalism is false isn't question begging if it uses that premise. That was her claim. That's quite a different issue from the one you're raising, viz. whether the premise can be supported.
The "gaps and incredulity" charge is inane. A person who points out that a housecat cannot lay an egg which will hatch into a tyrannosaurus rex is relying on "gaps and incredulity". They can hardly do otherwise no matter how much science or (good) philosophical argument is trotted out, because the point of their argument *IS* that there is a gap, and that this gap is unbridgeable. And if the gap really is unbridgeable, well then - it makes sense to be incredulous about the claim, now doesn't it?
The aspiring doctor may as well defend idealism or panpsychism or full-blown cartesian dualism on the grounds that all the opponents have to support their claims that these views are wrong amounts to "gaps and incredulity".
Sorry, the AfR needs a lot more than that. There's a reason why there have been moves in recent years to bless those who subscribe to Chalmers' views (so long as they don't utter the dreadful word 'soul') as naturalistic explanations despite his open hostility to materialism. And why there's been a growing warmth to panpsychism among naturalists as well. (And what did Strawson call panpsychism? Real materialism, of course.)
Incidentally, the comments on that site are atrocious. The idea that Victor argues that if the AfR is correct then God exists (or, worse yet, that if the AfR is correct then Jesus is God) is downright silly. Does anyone realize that naturalism can be false without Christianity (or even theism, certainly classical theism) being true? Or that Christianity is not the only alternative to naturalism? (I certainly believe Christianity is right, but still, the playing field is a bit broader than THAT.)
Anonymous,
A person who points out that a housecat cannot lay an egg which will hatch into a tyrannosaurus rex is relying on "gaps and incredulity".
No, they're not. They're relying on induction. And their claim would be that it is extremely improbable that a house cat would lay a T-Rex egg.
The AfR isn't like that at all. What examples do AfR proponents hold up for their gap/incredulity arguments? Deep Blue.
Is Deep Blue supposed to show that mechanisms cannot think about things? That would be like arguing that house cats aren't material because rocks don't give birth. That's poor inductive reasoning because the behavior of matter varies with its structure. Likewise, Deep Blue does not have structure or behavior similar to a conscious human brain. For starters, Deep Blue does not do abstraction, nor does it employ propositions.
The main problem with the AfR is that it fails to define what intentionality actually is. After all, if you could define it, it would be pretty straightforward to implement it in a machine.
So... simple question: Imagine yourself thinking a thought that lacked intentionality. How would you know that this thought lacked intentionality? Obviously, the thought would seem confused, but in what way?
Did the argument from reason originate in "C.S. Lewis’ Dangerous Idea"? Sources seem to indicate that the argument from reason was developed by C.S. Lewis and NOT by Reppert and has been developed by several individuals, and thus I do not need to read Repperts book in order to understand it - I need to read C.S. Lewis (which I have). I used Reppert's book in my blog post because I had seen his book in a store and it made me think of the argument from reason, given that it mentioned it on the back cover.
Similarly, I do not need to consult a particular contemporary writer who expounds upon the ontological argument in order to understand the ontological argument, given that the contemporary writer is not the first to use it, especially if I can find multiple sources that lay the argument out. I used the Wiki version of the argument after reading several versions and concluding that it was decent enough to quote from.
If Dangerous Idea contains a more coherent formulation of the AFR that is missed in the various sources I consulted (including Lewis), then perhaps you could explain exactly why my sources are inadequate?
In mentioning specifically that I offered a refutation of the AFR going from Wiki and not your book, you are making an ad hominem attack - you are attacking my presentation of the argument rather than my counter-argument. It would be more persuasive to provide a counter argument, rather than simply imply that I have not done my homework.
Is my presentation inadequate? If so, please correct me, especially in the comments section of my blog, so that my commenters will know I am mistaken. If my presentation is adequate, then my source for the argument is irrelevant.
Can you explain why I don't understand "begging the question", rather than simply stating I do not understand it?
Can you please explain how a deductive argument can be useful if the conclusion is "contained" in the premise? To my knowledge, a deductive argument is one in which it is thought that the premises provide a guarantee of the truth of the conclusion. In this way, the premises are given as support for a conclusion in such a way that if the premises are true than the conclusion must be true.
One could describe a circular argument as deductive, but not all deductive arguments are circular (I.E. contain the premise in the conclusion) and a deductive argument which is circular fails precisely because the premise contains the conclusion.
@Anonymous - the commenter was being sarcastic with the AFR = Jesus comment. We get a little rowdy over there, but certainly people use arguments such as the AFR as a springboard to lead to Christianity, much like Lewis does with the argument from morality (among others) in Mere Christianity.
P.S. Hi everyone!
P.S. For the sake of transparency, I wanted to let you know that am posting my thoughts on this exchange to my blog this afternoon.
DL,
As I said, no matter how much science or philosophy is used in such a case (induction would clearly, for our purposes, fall under the former in this discussion), in the end what is had are 'gaps and incredulity'. Use as much induction as you please. All the induction will illustrate, and all it can possibly illustrate, are "gaps and incredulity". A gap, because it would be argued there is no detailed explanation for a housecat to become oviparous and lay an egg, especially an egg for a creature outside its species. Incredulity for obvious reasons (just plain unthinkable and seems superficially unlikely, etc.)
In other words, griping that all the the AfR proponents have on their side are "gaps and incredulity" is weak.
Second, your reference to Deep Blue is misleading. AfR proponents can and do use DB to illustrate some of what they mean - but you know what else is available to them? The human brain itself, associated neuroscience, and the fundamental commitments of materialism. Let's go ahead and talk about neurons, about the effects of harm or stimulation on the human brain, on modeling techniques, etc, all we please. Let's entertain the purely speculative notions of what "future science" may reveal according to materialist philosophers of mind, considerations of just how reason and intentionality (qualia aside) may be accounted for in the brain, without these things ultimately being defined away as illusions, illicitly smuggling in the very things they're attempting to explain, or leaving the questions ultimately unanswered anyway. AfR proponents are more than happy to engage all of these things, because they serve as beautiful illustrations of their point.
What's more, you just played a little bit of sleight of hand with Deep Blue. Why isn't Deep Blue (or any computer, really) capable of reason? Because DB's structure differs from a human brain of course. Wonderful. Care to explain this structure that is capable of explaining reason/intentionality in purely material terms? And if you're lacking purely scientific evidence here, would you care to tell us what such an explanation could even look like in theory?
Hey, I've never read Logic and Theism, but Wikipedia says that Mackie's logical argument from evil goes like this:
(1) God exists
(2)God is omnipotent, omniscient and wholly good
(3) Evil exists
is an inconsistent triad
Here's my refutation:
It isn't - duh!
As I see it, the AFR does not depend on the premise that 'rationality cannot arise out of non-rationality'. Non-rationality can, potentially, give rise to anything, just as a true and valid conclusion can be a result of false, even meaningless, premises.
The problem is that we cannot trust what is produced by non-rationality. It may be reliable, it may not. What reason do we have to think that the physical events in our brains give us a true picture of the universe? They may, they may not.
I wish Dr. Reppert would take part in this conversation
kbrowne, I don't think an AfR proponent would dispute that some non-rational source can (for example) by sheer coincidence string together words which could translate to an accurate statement. It's rather like admitting that a roomful of monkeys, given infinite time, may be able to produce the complete works of Shakespeare. But just as that sort of process isn't at all like what we think Shakespeare went through when he wrote his plays and sonnets, some non-rational source now and then getting something right isn't really the same as reason. In both cases all you're getting is a happy coincidence. And if reason is nothing but happy coincidences produced from blind, purposeless forces, then reason it ain't. (And if reason is more than happy coincidences from blind, purposeless forces, then naturalism has some splainin' to do.)
I also see the same problems apply to supernaturality or theism. Even if one were to potentially argue that naturalism leads to an untrustworthy source of rationality, positing a supernatural source for rationality will have the same and possibly more complex problems.
I think it would be best if we continued this discussion in one place rather than two, so I will stick to my blog since it is easier to keep track and respond to comments there. I'll still check back here from time to time, but if you see silence on my end it is not because I have nothing to say.
Anonymous,
You appear unfamiliar with the meaning of "gaps argument" or "argument from personal incredulity".
They are both arguments from ignorance.
Care to explain this structure that is capable of explaining reason/intentionality in purely material terms? And if you're lacking purely scientific evidence here, would you care to tell us what such an explanation could even look like in theory?
You ignored my challenge, but I'm happy to explain for you.
What would it mean for you to have a thought that lacked intentionality?
It would mean that a proposition in your mind didn't mean anything, or didn't mean what you thought it meant.
Consider the proposition "the cat is on the mat". I know what this means to me because I know what experiences or sensations correspond to the proposition. I know the terms in the proposition for the same reason. The proposition means (in a simple case) that my cat recognition faculty and my mat recognition faculty will both fire in coincidence along with my "X on top of Y" recognition faculty. Or something along those lines.
Now, in order for this to happen physically, there needs to be a circuit in my brain that represents the abstraction for cat. As it happens, there is such a thing. It's called a hierarchical temporal memory, and the human neocortex is packed with them. You can train such memories to recognize cats in general by exposing them to a single cat. Once trained, the memory doesn't just recognize the sample cat, but other instances of cats, cats in other colors, cats of differing sizes and orientations, cats with missing elements (like a missing eye or leg), etc. In other words, an HTM trained to recognize cats is capable of recognizing cats that do not actually exist, were such fictitious cats to be placed at its inputs.
So, if the components of a proposition refer to HTMs, they refer to precisely the kinds of abstractions necessary for meaning in terms of past experiences. I know what "cat" is about because cat references a circuit in my brain that recognizes cat-like objects. Essentially, the circuit is like a universal for cats.
BTW, HTM's need not only be connected to sensory inputs, but can be connected with other HTM's. That is, there can also be abstract thoughts, abstractions of abstractions, or abstractions that cross between abstractions and sensory inputs. It's straightforward to see how there can be abstractions for logical manipulations.
As we know, DB does not possess HTM's, does not think in terms of propositions, and lacks abstractions. That is, DB is missing the ingredients of intentionality.
You and other dualists may object to this sort of analysis because you haven't defined what meaning is, and have no intention of doing so (no pun intended). If meaning is a form of magic that cannot be defined, then intentionality has to be a form of magic, too. However, your own refusal to define meaning would not be a good argument against my materialist definition.
It's called a hierarchical temporal memory, and the human neocortex is packed with them.
Where was this demonstrated?
DL,
I'm familiar with both "arguments" in the special little score-book of debating rules. As I said, your citing of them in this context is ridiculous - and falsely implies that critics of materialism aren't interacting with neuroscience, materialist arguments, first-person experience, etc on this topic. And one more time, that "gaps and incredulity" are what all of their arguments ultimately point to is one weak criticism. "Gaps and incredulity" are all materialists have against dualism. The question is what arguments and reasons they cite for their belief that such gaps are problematic, and for their reasons to be incredulous. That, critics of materialism (note: not the same thing as dualists) have in spades.
Second, fantastic idea! One problem. What does it mean for there being a "circuit in the brain that represents the abstraction for cat"? Represents to who or what - another part of the brain? How is this representation accomplished? What you've offered here is a nice prospective model for building/programming a theoretical computer that appears human-like to us. Valuable stuff, to be sure. Not what we're after on the topic of intentionality. What's more, if you're saying that 'aboutness' is real - that pattern X in the brain is objectively and certainly "about 'the cat is on the mat'" and that information is a real and certain part of the universe rather than a useful, if fictional model.. well done, citizen. You've buried materialism, and have opened the doors to hylemorphic dualism all over again! To make "information" a real constituent of the universe is to indirectly support the AfR, not fight it.
Third, I'm an AI fan - so I decided to google around for hierarchical temporal memory. I stumbled across the conversation you had over at New Empiricism - and I have to say, this guy gave a reply I like. I'll quote.
If your conscious experience is like looking at a photo at any instant then we will need to introduce extension in time into our reasoning (see Time and conscious experience), otherwise we will be no more than a frozen set of objects and incapable of observation. Once our physical theories include extension in time we have left the materialist paradigm and mind as simple machine.
...
What you are doing with the machine model of mind is looking at your machine with your human mind and using this to join up the stages of the machine into a continuous picture. Sure your mind can extend over time but your machine is just a jerky procession of frozen patterns. When a given frozen pattern is present at the end of a clock pulse the last pattern has entirely gone.
...
To use Ryle's terminology, equating conscious experience with processing or functions is a category mistake.
Apt observations. Loosely related of course, but hey, intentionality questions are loosely related to the AfR too.
Anonymous, for my most developed view on neuronal representational systems, follow the links in this post. I've thought a lot about this, researched it empirically, written papers implicitly about it as a neuroscientist. If you read them and have questions or criticisms, I'd be happy to respond.
IT took me a half hour to find that @#$# link. I think I need to make a post at my own blog with all the relevant links, so I never have to look for it again. How annoying. Grrrr... Waking up in 6 hours to go running. :O
BDK,
Thanks. Interesting. Not really what I'm going after here (what's related to my response to DL doesn't offer anything new), but yet another fun resource to read up on for different topics.
Anon, it should be relevant to:
>>What does it mean for there being a "circuit in the brain that represents the abstraction for cat"? Represents to who or what - another part of the brain? How is this representation accomplished?"<<
Anonymous,
What does it mean for there being a "circuit in the brain that represents the abstraction for cat"?
My bad. It doesn't represent an abstraction. It IS an abstraction.
What's more, if you're saying that 'aboutness' is real - that pattern X in the brain is objectively and certainly "about 'the cat is on the mat'" and that information is a real and certain part of the universe rather than a useful, if fictional model.. well done, citizen.
No, the physical configuration in the brain isn't objectively about the cat in the sense that it lacks subjectivity. It objectively represents cats, is an abstraction of cat, to the subject that owns the brain. It is a structure that was formed by interaction with a cat, and a structure which will be triggered into similar activity when supplied with inputs resembling the cat.
To make "information" a real constituent of the universe is to indirectly support the AfR, not fight it.
This is question-begging. You're making an assumption about information. You're assuming that information is non-natural. Information is defined based on our own subjective perspective. It's not defined to be non-natural, so such an assumption is unwarranted.
Eric,
Check out http://www.numenta.com.
HTM technology is based on Jeff Hawkins theory which is explained in the book "On Intelligence".
BDK,
Anon, it should be relevant to:
Yep, and I didn't see that explained in your links. By all means, if you care to explain that, go for it!
DL,
My bad. It doesn't represent an abstraction. It IS an abstraction.
Fantastic: The pattern IS an abstraction. Great. So is a sentence containing an abstract proposition. As I asked, by what in the brain is this recognized? How is this accomplished? That a material medium can embody an abstraction (The words "The cat is on the mat" on a piece of paper) is not itself a mystery to me. The question is how this becomes an interpretation, a conceptualization in the/a mind - how is it grasped? A TXT file on a computer with the words "The cat is on the mat" in ASCII means diddly here.
No, the physical configuration in the brain isn't objectively about the cat in the sense that it lacks subjectivity. It objectively represents cats, is an abstraction of cat, to the subject that owns the brain. It is a structure that was formed by interaction with a cat, and a structure which will be triggered into similar activity when supplied with inputs resembling the cat.
Who said 'lacks subjectivity'? You're saying it "objectively represents cats" - which I can only take to mean is you declaring that information is a real constituent of the universe, not some useful fiction. Wonderful, I agree with you! You just sank materialism, and you're buying in to the AfR.
Unless you're saying that information doesn't "really" exist, and all you're describing here is some mechanical process that you're certain will explain the *appearance of* abstract thought, etc in a machine (in which case you're not talking about anything relevant) and eventually humans (in which case, pal, this ain't reason you're talking about. Ain't properly intention either.)
This is question-begging. You're making an assumption about information. You're assuming that information is non-natural. Information is defined based on our own subjective perspective. It's not defined to be non-natural, so such an assumption is unwarranted.
No, I'm assuming that information is non-material - and it is, even if substrates are involved. The AfR is aimed squarely at materialism; I could really care less if you suddenly want to define non-material existences, or even full-blown dualism(s), as natural. Go right ahead! Personally, I'd love to define hylemorphic dualism as natural. I'd love to define God as natural! Look at me, I'm an orthodox Catholic naturalist.
Now, I know from the thread over at New Empiricism that you think non-material/non-physical things can be filed under "naturalism". That's great, I don't even feel like fighting you on that. But if you're walking down that road here - if you're admitting that information/non-physical/non-material things is/are a real and actual constituent of our universe - then DL, you've left materialism behind. And that's central to the AfR's aim.
Don't worry, DL. You don't have say that God exists just because the AfR worked in that capacity. Victor has flatly said more than once that it mostly moves you to various other options in metaphysics and philosophy. But don't dig your heels in the ground and refuse to give credit where credit is due (assuming you're admitting these things are real, of course) just because a *gasp* christian or *GASP* apologist came up with a good argument. There are atheists out there who reject materialism. I'm sure they'll welcome you among their number. (Bring pizza. And wear a Discworld T-shirt.)
Anon: hmmm, perhaps you missed the links. For instance, this post entitled 'biorepresentations.'
Doctor Logic, that Numenta stuff is interesting theoretical engineering, but not particularly cutting edge or accepted as brain theory. I was not impressed by Hawkins talk on it at Society for Neuroscience. That said, it provides an interesting possibility argument.
Well the reason rationality cannot arise out of non-rationality is that rational effects cannot be produced by non-rational causes, and if whatever one designates as rational arises out of non-rational causes, then it is the effect of a non-rational cause and therefore non-rational as well.
Anonymous,
Fantastic: The pattern IS an abstraction. Great. So is a sentence containing an abstract proposition. As I asked, by what in the brain is this recognized?
If I have a simple machine that generates 5 volts output in the presence of cyanide, and 0 volts otherwise, then this machine "recognizes" cyanide. The simple machine will have neither abstraction nor consciousness, but that's not important. It still recognizes.
If I create a simple HTM that recognizes cats, that HTM will not just recognize, but also *be* an abstraction for cats. It recognizes all cats, including the ones with eye patches and peg legs. Because HTM's learn, the simple HTM is abstracting (verb) from its sample training cat to the class of all cats (or most of them). A simple HTM will lack consciousness of its recognition, and lack consciousness of its abstraction. However, it both recognizes and abstracts. Humans would differ because they not only recognize and abstract from examples, but they are conscious and aware that they are doing this.
So, the primary difference between simple HTMs and humans is that humans also recognize and abstract their recognition and abstraction abilities. How could a machine make that work?
Well, we can imagine that an HTM that finds patterns in its own internal processes of recognition and abstraction, and which will learn to recognize when recognition and abstraction are taking place. Moreover, it will be possible for this meta-HTM to recognize when a cat is being recognized versus when a dog is being recognized.
The question is how this becomes an interpretation, a conceptualization in the/a mind - how is it grasped?
How do you grasp ideas for yourself? Consider soft velvet. When you think about this term, you think of it in terms of past experiences. If you had no sense of touch, you would not really know what soft velvet was. You would know it only by its appearance to your other senses. If you lacked both sight and touch, the term loses meaning even further. Soft velvet is defined in terms of the inputs to your recognition and abstraction processes. When you think of soft velvet, you can feel soft velvet with your fingertips, albeit less vividly than if you were actually touching velvet. Thinking the concept of soft velvet causes your mind to react as if you were actually touching velvet.
This is precisely what you would expect if you brain was full of HTM's. An HTM is a memory and a recognizer. It contains feedback loops, so that retrieving a memory is similar to an abstract rehash of all prior training experiences.
Meaning is in terms of past experiences.
The same applies to mental abstractions. An HTM that has other HTMs as inputs can find patterns in patterns, abstractions in abstractions. The meaning of a higher-level abstraction is in terms of lower level ones.
Anonymous,
I'm assuming that information is non-material - and it is, even if substrates are involved. The AfR is aimed squarely at materialism;
You're taking a childish description of physicalism to an utterly absurd extreme.
Suppose that the universe is described by particle physics, and I have three Uranium atoms positioned such that they are equidistant from each other. This forms a triangle. Is the triangle real? By your definition, the triangle is non-material and not real to the materialist. This is nonsense. To a materialist, there is a triangle there. A triangle is a description of a configuration. It does no good to claim that, in descriptions of physics, there are no configurations, no abstract behaviors, etc.
Even if you were to say that descriptions of physical reality are not real, our own descriptions of physical reality would become real if we are physical beings.
Indeed, by your reasoning, there's no such thing as a planet in materialism. You practice some form of ultra-greedy reductionism in which a planet no longer exists because the planet is made up of quarks and leptons. But there's a difference between 10^30 quarks and a planet. A particular planet is a particular arrangement of those quarks. The behaviors of the planet depend on those arrangements. The planet is the word for that particular arrangement.
Indeed, by your definition, there aren't even laws. Consider Galileo's discovery that two different masses fall to Earth at the same rate (neglecting air resistance). Is this a law? Well, we would have to be talking about abstractions of mass, and abstractions of their initial position. We could be dropping the masses from the Tower of Pisa or from the Eiffel Tower, or dropping them from a tower we have yet to design and build. Furthermore, we would have to abstract the planet or gravitational field in which the masses are dropped. So are laws not allowed under physicalism? That would be a pretty peculiar physicalism, would it not?
Consider mathematics. By your definition, mathematics does not exist under materialism. Yet mathematics is a abstraction about computation. Mathematical systems are abstractions of systems. They're like physical laws of computing.
The idea that a mind is a physical system is the idea that a mechanism can create abstractions and think the way we humans are observed to think. If you could prove that machines could not create abstractions, then you would have a case against naturalism, but you don't have that proof. Indeed, the evidence is that not only can machines abstract, but that our brains are instances of such machines. (If dualism were correct, there would be no need for brains with memory or abstracting ability, nor need for a central nervous system at all.)
You have an argument from ignorance, i.e., "I can't imagine how a machine could think."
Getting back to your idea of information... Of course information exists under materialism. Suppose I have a configuration of atoms in a lattice, and their spins are either up or down. There is information in the configuration, i.e., it takes a certain number of bits to represent this.
There are two kinds of information. There's information in the sense of "how many bits does it take to record a configuration?", and also in the sense of "how much of this information has utility to me?" It takes a lot of bits to record the configuration of pixels in a black and white photo containing nothing but randomly black or white pixels. However, that photo has no utility to us, so it contains no information in the second sense. Even if you take information to be defined in the latter sense, there is still meaning to information in a physical world.
But don't dig your heels in the ground and refuse to give credit where credit is due (assuming you're admitting these things are real, of course) just because a *gasp* christian or *GASP* apologist came up with a good argument.
I wonder, could you possibly be more arrogant in your responses?
Doctor Logic,
You're taking a childish description of physicalism to an utterly absurd extreme.
No, DL, I'm pointing out the utterly absurd extremes to which a consistent materialism leads. Have you ever thought that maybe - just maybe - materialism is absurd? And that maybe, just maybe, if your chosen way to demonstrate that materialism isn't absurd is to take advantage of slippery definitions to define materialism to include what amounts to hylomorphic dualism.. that maybe you're helping to demonstrate its absurdity?
You know that materialism and physicalism are not strictly the same things ("Slippery" as they may be), so shame on you for trying to substitute one for the other so casually. Double shame on you for purposefully obfuscating and switching in "naturalism", "materialism" and "physicalism" so interchangeably.
Anyway, let's pull out a representative sample of your materialism "defenses" here.
Consider mathematics. By your definition, mathematics does not exist under materialism.
Do you want to know what your biggest problem here is, DL? You seem to think that if materialism ever leads to utterly inane conclusions, that somehow the definition of materialism being used must therefore be wrong and dishonest. Because it can't possibly be that materialism is utter bunk and ridiculous, can it? So instead you try some bizarre debating tae kwon do where if a reductio ad absurdum demonstrates materialism to be absurd, what it actually demonstrates is that materialism is actually something else. Convenient!
I mean, to hear you put it, there's actually no such thing as a philosopher (much less a materialist philosopher) who explicitly argues for eliminative materialism, or for mereological nihilism, etc. And further, there's no possible way for someone's metaphysics to, when taken to its logical conclusions, reduce to positions like these. No, absurdity is only possible for philosophies attached to intellectual positions you disagree with, right?
The meaning of a higher-level abstraction is in terms of lower level ones.
I ask how someone grasps an abstract concept in their mind, you give a reply that explains how a machine can cycle through processes and give reactions to events that could, from our interpretation, make reference to what we take to be an abstract concept. But does the machine know what it's doing? Does it grasp it even if it has been trained to output "The cat is on the mat" when it's presented with a picture of a cat on a mat, and "The cat is not on the mat" when presented with a picture of same? You provide "meaning" as a given for the machine to have (it's just in terms of 'lower level' ones), but the problem is getting any meaning at all (back to the Chinese room.) And again, even for the machine - if you are really and truly a realist about information - adios materialism anyway.
Getting back to your idea of information... Of course information exists under materialism.
You can argue that information really exists - that it is an objective and real constituent of the universe, not just part of some nominal story made up for the purposes of pragmatic explanation. But if it's objective and real, then there is more to the universe than "colorless, odorless, featureless, totally homogeneous particles advancing in perfectly predicable lines with all the rigid and absolute predictability of the premises of a syllogism".. and therefore you reject materialism.
(If dualism were correct, there would be no need for brains with memory or abstracting ability, nor need for a central nervous system at all.)
Utterly incorrect even on Cartesian dualism. On hylomorphic dualism you could not be more wrong, since proponents explicitly argue for the presence and role of bodily systems for sensory operation (no sight without eyes, etc). Wait, let me guess: Hylomorphic dualism is just physicalism now, right? L-O-L.
I wonder, could you possibly be more arrogant in your responses?
Pot, kettle, black.
Dr-
I too am confused about this statement > You wrote: If I have a simple machine that generates 5 volts output in the presence of cyanide, and 0 volts otherwise, then this machine "recognizes" cyanide. The simple machine will have neither abstraction nor consciousness, but that's not important. It still recognizes.
It doesn't appear to follow =/
DL,
"I suppose that sooner or later the physicists will complete the catalogue they’ve been compiling of the ultimate and irreducible properties of things. When they do, the likes of spin, charm, and charge will perhaps appear on their list. But aboutness surely won’t; intentionality simply doesn’t go that deep."
That's Jerry Fodor.
So do you agree or disagree with Fodor?
Is "forms a triangle" for your three uranium atoms an ultimate and irreducible property?
Anonymous,
But does the machine know what it's doing? Does it grasp it even if it has been trained to output "The cat is on the mat" when it's presented with a picture of a cat on a mat, and "The cat is not on the mat" when presented with a picture of same?
Yes. As I've explained, the machine does not simply contain what is necessary to recognize "cats on mats." It has what is necessary to recognize its own recognizing of cats on mats, as well as the ability to abstract to "X on mats".
You provide "meaning" as a given for the machine to have (it's just in terms of 'lower level' ones), but the problem is getting any meaning at all (back to the Chinese room.)
The Chinese Room argument proves that individual neurons don't understand Chinese. It doesn't have anything at all to say about what brains do or do not understand. The Chinese Room argument is a simulation of a brain as it operates one neuron (or one subnetwork) at a time (as if the simulation were running on a single processor computer). The fact that an individual neuron doesn't understand Chinese is pretty irrelevant to cognitive science.
Again we're back at personal incredulity as an argument against AI/materialism.
But if it's objective and real, then there is more to the universe than "colorless, odorless, featureless, totally homogeneous particles advancing in perfectly predicable lines with all the rigid and absolute predictability of the premises of a syllogism".. and therefore you reject materialism.
You suggest that for something to be objective and real, it has to be irreducible. Well, that's a very peculiar definition of objective and real. I quite expect that in almost any philosophy there are things that are composite and real. Even in mereological nihiism (which is a very silly idea) arrangements of simples are real and objective.
Utterly incorrect even on Cartesian dualism. On hylomorphic dualism you could not be more wrong, since proponents explicitly argue for the presence and role of bodily systems for sensory operation (no sight without eyes, etc).
Oh, I'm sure dualists reject the idea *now*. But 500 years ago, if you theorized that mental faculties were non-material, there are lots of mental faculties that could be implemented non-materially with no support from the physical brain. Now that we know the brain is capable of doing everything by its onesy, the non-material part is more elusive than ever. Your situation is analogous to that of the conspiracy theorist. No matter how much hard evidence piles up in favor of materialism (or physicalism, because they're both equally deadly to your superstitious religious attachments), you keep pointing to gaps to support your case. You want it both ways. You want to interpret all the latest evidence against the defendant as merely expected evidence for an invisible conspirator who framed the defendant.
I mean, to hear you put it, there's actually no such thing as a philosopher (much less a materialist philosopher) who explicitly argues for eliminative materialism, or for mereological nihilism, etc.
Yes, I've heard of such things, but these are fringe ideas to say the least. They are far from mainstream. If that's what you're arguing against, save your breath/keystrokes.
"I suppose that sooner or later the physicists will complete the catalogue they’ve been compiling of the ultimate and irreducible properties of things. When they do, the likes of spin, charm, and charge will perhaps appear on their list. But aboutness surely won’t; intentionality simply doesn’t go that deep."
That's Jerry Fodor.
So do you agree or disagree with Fodor?
Sure I agree. I never said aboutness was irreducible. Why should it be? It is the dualist's view that aboutness is irreducible, not the materialist/physicalist's view.
normajean,
If a mechanism responds only to a particular pattern in its input, then it recognizes that pattern, by definition. By this definition, hemoglobin recognizes oxygen.
I've also explained that when humans recognize something there's generally more than just recognition taking place. Humans have recognition, abstraction, and recognition of our own recognition. This is something that the simplest mechanisms (like hemoglobin) lack, but which more sophisticated mechanisms possess.
Hemoglobin does not learn to abstract, and does not recognize that it recognizes (because it cannot abstract).
HTMs can and do learn to abstract.
I can see a causal relation but not that it literally recognizes.
normajean,
I can see a causal relation but not that it literally recognizes.
Tell me what you think is missing.
Isn't recognition a causal relation?
There are different senses of 'recognize.' One sense DL is using, and another which requires conscious recognition. E.g., you see a friend's face but he is wearing a hat, so it takes two seconds before you have the 'click' of recognition. In AI, pattern recognition is simply the ability to reliably discriminate different patterns. For instance, the 'facial recognition' software can match a novel picture of Jill to a stored image template of Jill, and give a 'Jill' response (and can also do the same for lots of other face pictures). This involves a stored memory of Jill's face, and a comparison of this template with present data. (Of course there are other face recognition algorithms, this is just one type).
Clearly, such facial recognition software isn't conscious of faces, doesn't have that 'Aha there's Jill' experience that we would have.
The interesting question is how far such gizmos (implemented in brains) can get you toward full intentional contents and also toward consciousness (as I argue in the sites linked to above, I think they get you pretty close to the former, but it is less obvious how far they bring you to the latter. That is, it isn't obvious how to bridge the gap between recognition widgets and conscious states).
BDK,
I agree. It may not be obvious how to get to consciousness, but there's nothing to suggest we can't. All the evidence points to the likelihood that we can. We know humans are machines, and we know humans can do it.
Now, if dualists could prove we weren't machines, they would have a case, but they can't prove it. I mean, it's not like there's astral projection or telepathy which would provide a compelling case for dualism.
This is why dualism is such a total sham. Dualists now admit that in every scientifically detectable way we're machines, yet they still insist that machines can't possibly think like we do. That's like saying that in every scientifically detectable way, the Apollo moon landings will appear to have occurred, but actually they were faked. (Oh, and, furthermore, we can never land on the moon!) Dualism is a ridiculous conspiracy theory.
DL: I agree to some degree but think you overstate things a bit. I said at another blog (the rest is pure quote):
[N]aturalists need a positive story, a biological mechanism to point toward and say 'See, that's how cognition works.' In arguments about phenotypes whose phylogeny is unclear, at least we can point to natural selection as a plausible mechanism. There is, as yet, no 'standard theory' of the biology of cognition, no default consensus mechanism we can point toward.
Because of this, even naturalists are presently forced to be speculative when it comes to specific mechanisms of cognition. Hence, many dualists believe they are on equal footing with the naturalists. "Hey, if the naturalists have no consensus, not even a plausible story, about how consciousness emerges from brain matter, then why should we take their word for it? Indeed, here are some arguments that cognition/consciousness cannot emerge from brain matter. Much less than leaving us on an equal footing, that puts me one up on these silly naturalists that can't even keep their stories straight!"
Such lines of thought are not easy [for the naturalist] to rebut. I think ultimately, after giving the general reasons for thinking that mental states are brain states, we have to say something like: 'In the absence of compelling data a good scientist is less confident, not more confident, in her conclusions. In that spirit, I'll only say that time will tell which general approach is more plausible.'
I said:
" I think ultimately, after giving the general reasons for thinking that mental states are brain states"
This is nontrivial, because there are so many data to discuss, and so many reasons to believe the mind is neural. Unlike the antinaturalists, who tend to be one-hit wonders (basically repeating versions of Leibniz's windmill), the naturalists are sticklers for detail and evidence, and have lots of experiments they like to chew on (the antinaturalists should be chewing over the same experiments, but for some reason they like to stay in Leibniz's mill staring about in awe, feeding on their intuitions about what neuroscience will be like in 300 years).
I'm writing this up now but it is quickly expanding into a book unfortunately.
(for his mill example see Section 17 of the Monadology:
"Moreover, it must be confessed that perception and that which depends upon it are inexplicable on mechanical grounds, that is to say, by means of figures and motions. And supposing there were a machine, so constructed as to think, feel, and have perception, it might be conceived as increased in size, while keeping the same proportions, so that one might go into it as into a mill. That being so, we should, on examining its interior, find only parts which work one upon another, and never anything by which to explain a perception.").
BDK, off-topic sorry. Is there any news on abiogenesis you're aware of?
NJ: I only have an amateur's interest in this topic, but I recall thinking this was cool:
Mansy SS, Schrum JP, Krishnamurthy M, Tobé S, Treco DA, Szostak JW. Template-directed synthesis of a genetic polymer in a model protocell. Nature. 4 June 2008.
(Summary and implications discussed here).
It won't be long before we have arguments about whether these things created in labs are alive. Then the fights will be about whether the conditions in which such life is created are enough like those of prebiotic Earth. Then the fights will be over and another gap will be closed. :)
http://www.youtube.com/user/mazierephilippe#play/all/uploads-all/1/mojV17Y0iu4
Is this really the maverick philosopher?
Post a Comment