Recently, neurosurgeon Michael Egnor offered a parable about whether machines really learn. The tale features a book that “learned” to fall open at the right places. Computer scientist Jeffrey Shallit responded, claiming that machines really CAN learn!, and Dr. Egnor responded to him, pointing out that a baseball glove can “learn” the game if adjustment to circumstances is all we are counting.
But he also wanted to make clear to Dr. Shallit, brains don’t learn either. Only minds learn:
Shallit implies that the reinforcement and suppression of neural networks in the brain that accompanies learning means that brains, like machines, learn. He is mistaken. Brains are material organs that contain neurons and glia a host of cells and substances. Brains have action potentials and neurotransmitters.
Brains are extraordinarily complex, and brain function is a necessary condition for ordinary mental function.
But brains don’t have minds, and brains don’t have knowledge, and brains don’t learn. Reinforcement and suppression of neural networks in the brain are not learning. They are a necessary condition for learning, but learning is an ability of human beings, considered as a whole, to acquire new knowledge, not an ability of human organs considered individually. Human organs don’t “know” or “learn” anything. This error is the mereological fallacy. It is the same mereological fallacy [mistaking the part for the whole] to say that my brain learns as it is to say that my lungs breathe or my legs walk. I learn and I breathe and I walk, using my brain and lungs and legs… Michael Egnor, “Do either machines—or brains—really learn? A further response to Jeffrey Shallit” at Mind Matters Today
See also: Can machines really learn? Michael Egnor offers a parable.
Machines really CAN learn! A computer scientist responds to my parable
Also: Inner peace: Is there software for that? Tech billionaire funds neuroscience in a search for the secret of contentment
and
Google is collecting data on schoolkids. Some say it’s okay because the firm supplies a lot of free software and hardware to schools
Children are watching much less TV. But what we learned from children’s TV is coming back to haunt us.
Will AI triumph? Will that phone end up smarter than your kid? If so, it might not happen in quite the way we are told to fear. U.S. kids who spend more than two hours a day looking at screens “perform worse on memory, language and thinking tests than kids who spend less time in front of a device.”
Much of biological research has advance by interfering with the function of an organ/tissue/gene to see what the impact is. By doing this we learned that the pancreas is needed to process sugars, the kidney to remove nitrogen wastes, the retina to see, the cochlea to hear. But when the same approach is used on the brain we are told that there is something else.
By R J’s reasoning computers are just hardware cuz by messing with the hardware we can get the computer to malfunction.
ET
If you think that is what my reasoning is, feel free to proceed along these lines. I suspect that there are others here (BA77, KF) who don’t want you to go down this rabbit hole. But, please, chase the rabbit.
RJ@1
Interesting transition you made there. From processing sugars and removing wastes to seeing and hearing.
There’s a ‘necessary but not sufficient’ in those last two.
As both an AI researcher and an ID proponent, I must say that I am disappointed in Egnor. This is total nonsense. The only part of the mind that learns is the brain. Brains learn by trying new connections and disconnecting them if they fail a learning test. The test is invariably dependent on the timing of pulses or signals. For example, some neurons learn concurrent patterns of signals, i.e., groups of pulses that frequently occur simultaneously.
The idea that brains do not learn is pure nonsense. Sorry.
FF@5
Interesting….and how would the brain know of the learning failure? Where is that standard set?
Nor does any of this “pulse timing” or “signal patterns” explain consciousnesses required to understand the signals.
R J:
Nope, I don’t care to chase anything your mind conjures up. I am OK with correcting your mistakes
ET
When you start doing so, please let us know. 🙂
FF
Agreed. We know that certain physical brain anomalies are related to learning impairments or enhancements. We know that certain chemical imbalances can affect our ability to learn. We know that certain chemical treatments can reduce the impact of learning impairments. All of these are material in nature.
At post 5 FourFaces states:
Well, contrary to what you believe, material brains can learn nothing. PERIOD.
To learn anything requires a subjective immaterial conscious mind that can possibly know something new.
Brains know nothing. Immaterial minds know. To believe otherwise is, to use your term, ‘pure nonsense’.
Dr. Egnor succinctly sums up the ‘pure nonsense’ that atheists, (and apparently a few ID proponents such as yourself), are prone to in the following quote:
Moreover, the ability to learn something new requires the ability to create new information within a immaterial mind.
Yet the material processes of a computer, (or the material processes of a brain in this case), as Dembski, Marks, and company have shown, can never create new information over and above the information that was originally programmed into it by a immaterial mind.
And as Dr. Stephen Meyer stated,
FourFaces, It is a simple and profound category error on your part to confuse the complex material processes of the brain (or a computer) with the actual learning and knowing that is possible only with a immaterial mind.
Verse:
Let’s keep this simple. When Egnor finds a mind without a brain in sight, he’ll have evidence for his dualism. Otherwise it’s just wishful theological thinking.
Seversky, (a Darwinist who would be a neuronal illusion instead of a real person if Darwinism were actually true), has repeatedly refused to accept evidence from Near Death Experiences as proof that the conscious immaterial mind can exist apart from the material brain even though the evidence for the validity of Near Death Experiences is far more compelling than the nonexistent evidence is for Darwinian evolution:
Bottom line, We have far more observational evidence for the reality of souls than we do for the Darwinian claim that unguided material processes can generate functional information. Moreover, the transcendent nature of ‘immaterial’ information, which is the one thing that, (as every ID advocate intimately knows), unguided material processes cannot possibly explain the origin of, directly supports the transcendent nature of the soul:
As Stuart Hameroff states: “it’s possible that this (conserved) quantum information can exist outside the body. Perhaps indefinitely as a soul.”
Seversky’s repeated refusal to accept the validity of NDE testimonies, and yet his unquestioning acceptance of Darwinian evolution despite the abject failure of the material processes of Darwinian evolution to create ‘immaterial’ information, is unfortunately a defining characteristic of the uncritical Darwinian mindset.
R J:
I started when you started posting here. 😛
Seversky:
And when Seversky or anyone else can demonstrate materialistic processes can produce the brain they will have evidence to refute Egnor.
That will never happen because the brain is not reducible to necessity and chance.
R J:
We know that certain hardware anomalies are related to computer failures. We know that certain electrical imbalances can affect a computer’s ability to function. We know that certain conditions can reduce the impact of computer functionality.
And yet we also know that computers are NOT reducible to their hardware.
Latemarch @6
Interesting….and how would the brain know of the learning failure? Where is that standard set?
Nor does any of this “pulse timing” or “signal patterns” explain consciousnesses required to understand the signals.
The learning mechanism is in the neurons. It’s called Hebbian learning. Synaptic inputs that fire concurrently with a target neuron are strengthened until they become permanent connections. Those that don’t are weakened and eventually disconnected. Neurobiologists have known about this for decades.
The brain, by itself, does not explain consciousness because it takes two entities to have consciousness, a knower and a known. In this case, the knower is the spirit/soul and the known comprises certain parts of the brain (cerebrum or neocortex). It’s called dualism.
The idea promoted by some in the Christian community that the mind is just the soul is false. Knowledge resides in the brain and is known by the spirit or soul. I, too, am Christian but I refuse to park my brain at the door when I walk into a church.
Seversky @11
Let’s keep this simple. When Egnor finds a mind without a brain in sight, he’ll have evidence for his dualism. Otherwise it’s just wishful theological thinking.
Apparently Egnor does not believe in dualism. He believes in a form of monism whereby the mind is just the spirit or soul where knowledge resides and the brain has nothing to do with mind. Materialists believe in another form of monism: the mind is just the brain.
Both forms of monism are dead wrong, in my opinion. It takes two opposite and complementary entities to have a mind, a knower and a known. This is what true dualism is all about. For example, there are no colors in the brain. Colors are activated parts of the spirit/soul that some call qualia. Also, there is no 3D world in the visual cortex, only firing neurons. The fabulous and colorful 3D vista you think you see in front of you is a supernatural phenomenon. It does not exist.
It is not so clear what Egnor means by the merelogical fallacy. Is he saying that an isolated brain could not learn? But that would be similar to saying a computer needs external inputs to learn. That doesn’t seem problematic. I think he means there is a unified ‘I’ that governs everything, which is a good argument since an ‘I’ is inherently a different thing that a computation, but it is not so clearly worded here.
@FF reinforcement does not seem to be the same sort of thing as learning. Learning has to do with truth, but an error can be reinforced. So, Egnor’s distinction seems valid.
@ET why do you say computers are not reducible to their hardware?
@Seversky, why must there be a mind without a brain for dualism to be true? There are no living people without food and water, but people are not food. On the other hand, a disembodied mind would indeed be evidence for dualism.
EricMH- Last I knew a computer without software is just a big paper weight or door stop.
Of course matter itself is immaterial, so the whole mind-body dualism is obsolete:
http://nonlin.org/im-materialism/
…and of course if ‘machine learning’ is as thing, then your regular hammer can also learn:
http://nonlin.org/ai/
…but traditionally, the word “learn” was reserved for intelligent beings – organisms created directly by God and endowed with free will.
I can’t believe that, even though I believe in intelligent design, I am agreeing with Darwinists and materialists that brains and computers do learn. Something is not right in this picture.
FF
Welcome to the dark side my young apprentice. 🙂
A mind actually implies a person (even if it is angelic, and thus pure spirit). It is a sine qua non of personality, of personhood.
The soul consists of memory, will and understanding. Computers don’t possess the second faculty, and the other two need to be operated by a live entity, i.e with a soul.
It’s the plainest common-sense to most of the population, but, apparently, less and less by arch-worldlings, materialists, te higher up you go. To mix metaphores, their Achilles’ heel is their heart, long considered the seat of the will.
R J Sawyer at #9:
I think we could simply admit that we are discussing two different meanings of “learning”:
1) First meaning: if some algorithm can process new information from the outside, according to the rules it was programmed to implement, we call that “learning”.
2) Second meaning: a conscious being “learns” when he becomes aware of some new meaning represented in his consciousness.
So, a computer cam learn in the first sense, but not in the second sense, because it is not conscious.
If we accept that the brain is a physical interface to cosnciousness, then the brain can learn in the first sense, but not in the second.
Cosciousness, whatever it is, can learn in the second sense.
Now, you say, correctly, that brain structures are certainly important for our conscious learning process. That is no news, of course.
But then, even a computer is important for many of our learning processes. Or a book. Or a TV set.
IOWs, consciousness does use physical instruments to learn. In the case of a computer (or, for that, an abacus), consciousness uses intelligently designed algorithms to learn things that it could not easily learn in other ways.
So, if my computer is out of order, my learning is impaired. The same thing happens if my brain is out of order, in some way.
The point is: we become aware of new meanings using computing algorithms, but a computing algorithm never becomes aware of any meaning.
Four Faces,
To which RJ quips
That quip ought to send a chill right down your spine FF.
FF in 17 you state, “It takes two opposite and complementary entities to have a mind, a knower and a known.”
And in 16 you state, “Knowledge resides in the brain and is known by the spirit or soul.”
And in 21 you state that ‘brains and computers do learn.’
So I hold that you believe that material brains and computers, minus any spirit or soul, do not have a ‘knower’ associated with them.
So my question is this, if computers and brains, minus any spirit or soul, do not have a ‘knower’ associated with them, then what exactly is it that is doing the ‘learning’ when you say that ‘brains and computers do learn’ ?
It seems obvious that minus any ‘knower’, i.e. minus the immaterial conscious mind, there can be no learning.
So again, what exactly is doing the ‘learning’ when you say that ‘brains and computers do learn’?
Another point that might help clear this up is that you seem to believe that memories reside solely in the material brain, i.e. ‘the known’, and that the immaterial mind, i.e. ‘the knower’, can know this memory only as long as the immaterial mind is associated with the material brain.
Yet, if memories resided solely in the material brain, as you seem to hold, then none of the following evidence makes any sense.
For instance, people who have had hemispherectomies retain all their memories even though half their brain has been removed,
Retention of memories and a person’s personality after hemispherectomies simply makes no sense on your ‘knower/known’ model FF.
Moreover, Dr. Pim van Lommel states that “For decades, extensive research has been done to localize memories (information) inside the brain, so far without success,,,”
Moreover, if memory, and/or ‘the known’, resided solely in the brain as you hold FF, then a full life review, which is a common feature of many Near Death Experiences, should be completely impossible.
Around the 20 minute mark of the following Near Death Experience documentary, the Life Review portion of the Near Death Experience is highlighted, with several testimonies relating how every word, thought, deed, and action, of a person’s life (all the ‘information’ of a person’s life) is gone over in the presence of God:
Verse:
FF, I think your basic mistake is that you are confusing a representation of a memory in a brain, or in a computer, i.e. the known, with memory itself:
Although matter can represent immaterial information, it is clear that it is impossible for matter to actually ever be the immaterial information that it represents. This distinction between the representation of immaterial information and the actuality of immaterial information is made all the more clear by the abject failure of unguided material processes to ever generate immaterial information. i.e. Immaterial information, although it can be represented by material particles, simply can never be completely reduced to material particles. Thus, since memories at their most essential nature are immaterial information, then it is impossible for material brain states to ever actually be the memories that they represent to a mind.
In regards to this irreducible aspect of immaterial information, Dr. Stephen Meyer states:
And Dr. Stephen Meyer’s contention that immaterial information is real and is ‘not reducible to matter. It’s not reducible to energy’ is now empirically verified.
The fact that immaterial information is its own distinct physical entity that is separate from matter and energy, and that it can exist apart from its material representation, has now been established by the following.
First off, information can now be transferred, i.e. teleported, between particles completely independent of the material particles themselves physically interacting:
Secondly, immaterial information has now been shown to have a thermodynamic content and immaterial information has also now been shown to be able to accomplish physical work:
Thus FF, immaterial information is a real physical entity that has now been shown to be its own distinct entity that is separate from matter and energy and thus your belief that immaterial information resides solely in the brain, i.e. the known, is shown to be without warrant. To repeat, material brain states can represent immaterial information but they can never actually be the immaterial information that they represent.
Of supplemental note: I hold that Dr. Egnor, (who I remind is a brain surgeon as well as a professor of brain surgery), has the most accurate model, via Aristotle and Aquinas, as to how the immaterial mind and the material brain actually interact, (i.e. particulars and universals), than any other model that I have ever seen presented.
Verses:
BA77
Humour does not run deep in this one, my young apprentice. 🙂
My interpretation of what gpuccio wrote @24:
No human-designed technology (computers, robots) can become aware of any meaning, simply because they lack consciousness.
However, we conscious beings, can benefit much from intelligently designed “machine-learning” algorithms that allow the robots to widen their range of data-processing capabilities.
I like the clear distinction gpuccio made between two concepts of “learning” that are being used in this discussion. Let’s keep that important difference in mind when we express our ideas about “learning”.
Robots and computers aren’t aware of what they are, much less of the meaning of the information they process in increasingly sophisticated ways.
Only conscious beings are aware of meaning. In the case of humans, we’re also aware of our own existence.
BA77 @25
So my question is this, if computers and brains, minus any spirit or soul, do not have a ‘knower’ associated with them, then what exactly is it that is doing the ‘learning’ when you say that ‘brains and computers do learn’ ?
It seems obvious that minus any ‘knower’, i.e. minus the immaterial conscious mind, there can be no learning.
This is faulty logic, IMO. Knowledge is what has been learned by the brain. Building knowledge does not need a knower. It is based on physical causes and effects and on temporal regularities or order. Sure, the knower in a human brain can decide what the brain learns (music and the arts, for example) but it does not do the learning. Otherwise, we would know how learning occurs.
By the way, I don’t believe animals have a soul or spirit. I believe they are unconscious meat robots regardless of how conscious they appear to be. But animals can certainly learn. I have written AI programs that can listen to sounds (speech, animals or musical instruments, etc.) and learn to recognize them. I am certain that my programs are not conscious and do not have souls.
I predict that robots with human-level intelligence will be created in our lifetimes. By this, I mean a robot that can walk into a generic kitchen, even one it has never seen before, and prepare a breakfast of scrambled eggs with bacon, pancakes, toast and coffee.
R J Sawyer @22
Welcome to the dark side my young apprentice. ????
I can assure that my feet are firmly planted on the well lit side. “So sure of this, you are.” Yoda once said to me. I answered, “Yes, master Yoda.” He replied, “May the Force be with you, young Jedi.” ????
as to
BA77: It seems obvious that minus any ‘knower’, i.e. minus the immaterial conscious mind, there can be no learning.
FF: This is faulty logic, IMO. Knowledge is what has been learned by the brain.
Strange logic you have there FF,,, a brain, minus a ‘knower’ somehow learns knowledge.
Hmmm, but as a 5 year old might ask, would that not make the brain that supposedly learns knowledge the ‘knower’ ???
Anyways,,, you go on to claim.
“Building knowledge does not need a knower. It is based on physical causes and effects and on temporal regularities or order.”
And the law of conservation of information, which says purely physical causes can never create information over and above what was put there from the start, by a mind, figures into your scheme of ‘building knowledge’ how exactly?
Because of such strict limits on the creation of new information, the future of AI is greatly overrated.
BA77,
I said all I wanted to say in this discussion.
FourFaces: “I said all I wanted to say in this discussion.”
What you have said thus far makes no sense to me.
To reiterate:
People learn knowledge, brains and computers merely store representations of knowledge and/or information that can be accessed and manipulated by the minds of people.
By saying “brains and computers do learn” you are making a category error and are imparting ‘personhood’ to a brain and/or computer.
Definition:
To acquire knowledge in something, i.e. to ‘learn’ something, requires, by your own model, a ‘knower’. But, again by your own model, the brain is not the ‘knower’ but is the ‘known’.
Of one final note, since God himself possesses all wisdom and knowledge then, by your model, I suggest our ‘knower’ get to know God.
Verses:
Of supplemental note: