Header 1

Our future, our universe, and other weighty topics


Sunday, June 25, 2017

Physicists Should Have Invested in Super Heroes Rather Than Supersymmetry

A recent article in the New York Times discusses the failure of attempts to find evidence for the physics theory called supersymmetry.

These are difficult times for the theorists,” Gian Giudice, the head of CERN’s theory department, said. “Our hopes seem to have been shattered. We have not found what we wanted.” What the world’s physicists have wanted for almost 30 years is any sign of phenomena called supersymmetry, which has hovered just out of reach like a golden apple, a promise of a hidden mathematical beauty at the core of reality.

The Standard Model of physics has fewer than 30 independent parameters. But according to one scientific web site, supersymmetry has more than 100 independent parameters. According to another page at the same web site, more than 10,000 scientific papers “reference” the theory of supersymmetry.

With all that work by physicists, you would think that there must be some evidence for supersymmetry. But efforts to find evidence for the theory have been a complete failure. The Large Hadron Collider has completely failed to support the theory.

Why have physicists spent so much time on such a theory? It was to try to explain a case in which nature seemed to be extraordinarily fine-tuned. The wikipedia article on supersymmetry states the following:

In the Standard Model, the electroweak scale receives enormous Planck-scale quantum corrections. The observed hierarchy between the electroweak scale and the Planck scale must be achieved with extraordinary fine tuning. In a supersymmetric theory, on the other hand, Planck-scale quantum corrections cancel between partners and superpartners (owing to a minus sign associated with fermionic loops). The hierarchy between the electroweak scale and the Planck scale is achieved in a natural manner, without miraculous fine-tuning.

The “miraculous fine-tuning” being talked about here (what is known as the hierarchy problem) is a kind of matching of two unrelated numbers so that they end up canceling each other out – rather like what you might have if you had to pay on Friday a $5000 payment to save your house from foreclosure, and you coincidentally won $5000 in the lottery on Friday morning.

But this hierarchy problem is actually much more of a coincidence. Because according to this scientific web site, “one has to hypothesize that the several correction terms cancel out to a part in 1034 (a hundred billionths of a billionth of a billionth of a billionth), if one is to make the Higgs mass smaller than a lead brick.” So maybe our analogy should be that you own 6 houses that are each behind $5000 on the mortgage, with Friday being the last day for you to save them; and you coincidentally on Friday morning buy 6 different lottery tickets that each win $5000. That's the kind of fine-tuning that seems to be involved in the case of the hierarchy problem.

Supersymmetry (also known as SUSY) is an attempt to explain away this “miraculous fine-tuning.” But supersymmetry has always been a ridiculously ornate contrivance. For example, it imagines that almost every known type of particle has a corresponding “superpartner.” It would be quite the fantastic coincidence if nature was set up in such a way. So supersymmetry is basically a kind of gigantic case of “robbing Peter to pay Paul.” It tries to get rid of one fine-tuned coincidence (the hierarchy problem) by introducing a whole bunch of other fine-tuned coincidences, involving all these cases in which a known type of particle happens to have a matching “superpartner” particle.

It's kind of the same approach taken by a similar theory, the theory of cosmic inflation or exponential expansion. That theory tries to get rid of a case of extreme fine-tuning in the universe's first second, but it does it by making assertions that require their own fine-tuning in numerous ways. So it's just robbing Peter to pay Paul. There is no real net reduction in the amount of fine-tuning required.

For years scientists were hoping that the Large Hadron Collider would produce some evidence for supersymmetry. But no evidence has been produced. The chance of supersymmetry being confirmed in our lifetimes now seems almost zero. That's no surprise. Supersymmetry is a case of trying to explain away an example of cosmic fine-tuning, and there has never been a confirmed success in any effort to explain away any case of cosmic fine-tuning. The cosmic inflation theory (the theory of exponential expansion in the universe's first second) has been sociologically successful, a case of a thought virus that spread widely. But it has not been scientifically successful, because no evidence has been produced for it (and there are many problems with the theory).

Rather than investing so much time and effort on supersymmetry, physicists would have got a lot better return if they had invested instead in super heroes. There are several ways physicists could have done that. The first way would have been to invest in physics technologies that might have given us high-tech gadgets that would allow someone to have the equivalent of a comic super-power. Physicists might have invented some fancy gadget that would give some of the powers of Batman's utility belt. Or they might have invested in some bullet-stopping force field that might have worked kind of like Wonder Woman's bracelets. Or physicists might have invested in some super-strong material that would give someone some of the powers of Spider Man or Iron Man.

Or, physicists could simply have invested in super heroes without doing any research. They could have taken all that money wasted on supersymmetry papers, and invested the money by either buying super hero comic books, or investing in companies such as Marvel that published comic books. Comic books have long been collector's items, and a comic book which sold for 12 cents back in the 1960's may sell for hundreds of dollars today. Disney paid 4 billion dollars for Marvel Entertainment. If physicists had invested in super hero comic books or the companies that published them, the physicists could have made gigantic returns for themselves or for the colleges or universities where they work. 

supersymmetry

I must confess that I myself am guilty of failing to see how much money could be made from comic books and from another collector's item: baseball cards. When I lived in a dull Maryland suburb in the early 1960's, I had lots of comic books and many baseball cards. In the suburb I lived, elementary school students were preoccupied with baseball cards. The kids would gamble the cards in various ways, which made a fun pastime. One game worked like this: you would take your stack of cards, and face your friend who also had a stack of cards. After each of you shuffled your stack, both of you would deal them into a common stack. First you would deal a card, then your friend. If your friend dealt a card with one top color bar, and you then dealt a card on top of that stack with the same color bar, you would win the entire stack that had been dealt.

In my suburb it seemed every family with boys had a cardboard box of baseball cards. But when I was eleven I moved to a very different environment: the more sophisticated locale of Washington D.C. Of course, I packed my comic books and baseball cards, and I hoped the kids there would be as interested in these as the kids in my Maryland suburb. But it seemed that none of the Washington children had the slightest interest in baseball cards. And they didn't seem too interested in comic books. I kind of thought to myself: I guess the city kids are too sophisticated for these things – maybe they're just silly suburb things. So not very long after moving to Washington D.C, I threw away all my baseball cards and comic books. What a mistake! I could have made thousands if I had kept them.

Whether children at a particular school gamble with baseball cards is a sociological consideration, a vogue of a particular locale. When groundless theories such as supersymmetry become all the rage among little academic tribes, it seems to be also a sociological consideration, a case of some vogue that went viral when it shouldn't have.

Wednesday, June 21, 2017

Shermer's Faulty Case Against an Afterlife

Scientific American columnist Michael Shermer has a new column entitled “Why the 'You' in an Afterlife Wouldn't Really Be You.” Arguing against both spiritual concepts of an afterlife and technological concepts of a “digital afterlife,” Shermer attempts to give three arguments against the possibility of an afterlife.

His first argument is without merit, as it depends on an unproven assumption. Shermer argues:

First, there is the assumption that our identity is located in our memories, which are presumed to be permanently recorded in the brain: if they could be copied and pasted into a computer or duplicated and implanted into a resurrected body or soul, we would be restored. But that is not how memory works. Memory is not like a DVR that can play back the past on a screen in your mind. Memory is a continually edited and fluid process that utterly depends on the neurons in your brain being functional.

This argument is not valid against any of these three ideas of an afterlife: (1) the idea that you have some immaterial soul that will survive death in a kind of natural way, without any divine intervention required; (2) the Christian idea that the dead will be physically resurrected by some divine agent; (3) the idea that minds may be uploaded into computers in the future, providing people with a digital afterlife. The first of these ideas does not depend on the idea that memories are permanently recorded in the brain. A person believing in a soul may believe that memories are stored largely in some soul, and may deny the claim that memory depends on neurons (a claim scientists haven't proven). The argument also does not debunk the idea of a physical resurrection of the dead. In such a case the neurons of individuals would presumably be recreated. The argument also does not debunk the idea of uploading minds into a computer. If our memories now depend on neurons (and there are reasons for doubting that), that is merely a current dependence, that could in theory be overcome if some new type of computer could be created that could store the equivalent of human neural states.

In his second argument, Shermer attempts to stretch out one of the arguments made against mind uploading, and turn that into an argument against believing in any type of afterlife. Mind uploading is the idea that in the future it will be possible for people to transfer their consciousness into a computer. The idea is that we will be able to scan the human brain, and somehow figure out some neural pattern or synaptic pattern that uniquely identifies an individual. Then, it is argued, it will be possible to recreate this pattern in some super-computer. It is claimed that this could be a method of getting a digital afterlife. Some futurists claim that if someone were to have his brain scanned and then have his neural patterns transferred to a computer, that person could discard his body, and continue to live on indefinitely within a computer.

Here is how Shermer presents his second argument:

Second, there is the supposition that copying your brain's connectome—the diagram of its neural connections—uploading it into a computer (as some scientists suggest) or resurrecting your physical self in an afterlife (as many religions envision) will result in you waking up as if from a long sleep either in a lab or in heaven. But a copy of your memories, your mind or even your soul is not you. It is a copy of you, no different than a twin, and no twin looks at his or her sibling and thinks, “There I am.” Neither duplication nor resurrection can instantiate you in another plane of existence.

Minus the clumsy twin reference, this argument has considerable force against the idea of a digital afterlife through mind uploading. If I were to scan your brain and recreate your neural pattern inside a computer, that seems to be making a copy of your consciousness rather than a transfer of your consciousness, for two different reasons. The first reason is that a transfer has been made to a totally different medium (from a biological platform to an electronic platform). The second is that the mind upload seems to leave open the possibility of your biological body continuing after your mind upload has completed, and that would seem to be the creation of a copy of your consciousness rather than a transfer.

But the copying argument has much less force (and perhaps no force) when used against the idea of a physical resurrection of the dead. In that hypothetical possibility, there is no transfer to a different medium. A physically resurrected body would be the same (or mostly the same) as the body that existed before someone died. Also, a physical resurrection would presumably only occur for people who had already died. So there would be no issue that both a source and a target (or copy) could exist at the same time.

The copying argument has no force at all against the idea of a soul that continues to exist after a person's body dies. In such a case, presumably there would be no copy at all made of anything. A person who believes that the soul survives death does not tend to believe that the soul suddenly appears at the moment of death, suddenly having a copy of consciousness and memory that was stored in the brain. Such a person will instead tend to believe that the soul was a crucial component of the human mind all along (or perhaps was equivalent to the mind), and that such a soul simply continues to exist when a person dies.

Imagine a person who has worn a heavy lead coat all his life, and has worn a dark glass fishbowl over his head all his life. Under the soul concept, we may regard death as being rather like a person shedding that heavy lead coat and dark-glass fishbowl, with the heavy lead coat and dark-glass fishbowl being the restrictions of movement and perception associated with a human bodily existence. Under such a concept, there is no copying at all, but more like a kind of jettisoning, rather like a soaring rocket jettisoning a no-longer-needed fuel tank, with the body being what is jettisoned. Such a concept of the survival of the soul is completely free from difficulties involving copying, because no copying at all is assumed.

Shermer's third argument has no force against any concept of an afterlife other than a digital afterlife. Using the odd term “POVself,” he argues as follows:

If you died, there is no known mechanism by which your POVself would be transported from your brain into a computer (or a resurrected body). A POV depends entirely on the continuity of self from one moment to the next, even if that continuity is broken by sleep or anesthesia. Death is a permanent break in continuity, and your personal POV cannot be moved from your brain into some other medium, here or in the hereafter.

Such reasoning based on continuity might have some force against the concept of a digital afterlife by means of mind uploading, but probably not very much force as Shermer has stated it. For he's used the phrase “there is no known mechanism,” which is hardly going to discourage futurists who will claim that such a mechanism will be invented in the future. The reasoning here also has no force against either the possibility of a physical resurrection of the dead or the idea of a soul surviving death. A Christian believing that the dead will be physically resurrected will believe that this is done through divine agency, so it is futile to argue that such a thing cannot occur because there is “no known mechanism” by which it would occur. The person believing in a soul that survives death need not believe that any transfer, copying or moving occurs to allow a person to survive death. Such a person will tend to believe that the essence of a person – what makes you you – has already resided in your soul all along, not in your brain; and such a person is not required to believe that anything is transferred from the brain to the soul when a person is died. Under the idea of a soul, there is no break in continuity when a person survives death.

Neuroscientists have always followed the principle: explain every conceivable mental activity as being something caused by the brain. But what we must remember is that Nature never told us that all our memories are stored in brains, or that our thoughts are generated by brains. It was neuroscientists who told us that, not Nature. For example, we have no understanding of how 50-year-old memories can be stored in brains, given all the rapid molecular turnover that occurs in brains. Below is a quote from a recent scientific press release, citing a comment by a neuroscientist.

Neuroscience has also been struggling to find where the brain stores its memories. “They may be ‘hiding’ in high-dimensional cavities,” Markram speculates.

Such a quote (which has a “grasping at straws” sound to it) gives a very strong impression that neuroscientists have no real basis for being confident claiming that long-term memories are stored in the brain. The type of evidence neuroscientists cite for their claims is often weak evidence that doesn't hold up to critical scrutiny, such as dubious brain scanning studies which typically take minor 1% differences in brain activity, and try to make them look like compelling signs of what the brain is doing, when they are no such thing.

The idea behind a soul or spirit can be summarized as follows. You have a soul or spirit that is not at all a brain thing. You also have a brain, which serves largely for the purpose of localizing or constricting your mental activity, making sure that it stays chained to your body. The main purposes of the brain are things like control of autonomic functions, response to tactile stimuli, coordination of muscle movement, the coordination of speech, and the handling of sensory and auditory stimuli. There may also be some brain function relating to storing kind of what we may call “muscle memories,” which we use for performing particular physical tasks. These are all things related to living and surviving as a corporeal being. But things such as abstract thinking, conceptual memory and long term memory may be functions of a human soul. So when you die you may lose those things that you needed to continue walking about as a fleshly being, but may still have (as part of your soul) those things that were never necessary for such an existence (no cave-man needed to form abstract ideas, think philosophical thoughts, or remember his experiences as an 8-year-old).

This idea may seem old-fashioned to some, but a strong argument can be made that it is compelled by fairly recent evidence, and that in such a sense it is not at all old-fashioned. It is only in recent years that we have discovered what a very high degree of molecular turnover occurs in the brain, which makes it so hard to maintain that 50-year old memories are stored in the brain, as discussed here. It is only in the past 50 years that we had research such as John Lorber's, showing astonishingly high mental functioning in patients who had large fractions of their brains (or most of their brains) destroyed by disease. It is only in the past 130 years that experimental laboratory evidence has been repeatedly produced for phenomena such as ESP, which cannot be accounted for as a brain effect. It is only in recent decades that we have had reports of near-death experiences, in which many observers have reported floating out of their bodies, often verifying details of their medical resuscitation attempts that they should have been unable to observe  while they were unconscious. Such observations are quite compatible with the idea of a soul that survives death, and may force such a conclusion on us.

Shermer seems rather to be thinking of the idea of the soul as something kind of like a USB flash drive to back up the brain. But those who have postulated a soul have more often supposed it as a crucial component of human mental functioning, not some optional accessory. Under the concept of a soul that accounts for a large fraction or most of human mental functioning, there is no requirement for any sudden copying from the brain to occur for someone to survive death; and there is also continuity, as death merely means discarding what you don't need to survive beyond death. As none of Shermer's three arguments damages such a concept, Shermer has not succeeded in closing the door to an afterlife.

We know not what lies at the end of the misty bridge

Saturday, June 17, 2017

Nonsense in NatGeo's “Year Million”

The National Geographic channel has a new TV series entitled “Year Million.” A recent episode in the series was entitled “Never Say Die,” and was about the prospects of human immortality. The episode was long on slick visuals, and short on intelligence. The episode presented the ideas that we will greatly extend human longevity by using nanobots and genetic engineering, and that later people will upload themselves into computers.

The episode presented a simplistic view of biological life, one that is purely mechanistic, epitomized by a person who said, "We are still biologically hindered by the squishy computer in our heads.” The assumptions seemed to be that biological life is merely some mechanical phenomenon that humans can completely master through technologies such as nanobots and genetic engineering. These assumptions are quite dubious because of all the profound mysteries involving biological life.

It is generally admitted that we do not understand the origin of life. It is generally admitted that we do not understand the mystery of protein-folding (why proteins conveniently form into three-dimensional shapes needed for our existence). It is generally admitted that we do not understand the mysteries of morphogenesis, how it is that a fertilized egg is able to progress to become a baby who then grows into a full-grown human. We also don't really understand the origin of species, as the theory of evolution by natural selection is merely a theory of accumulation rather than what we actually need to explain the mountainous degrees of coordination, structural fine-tuning, and functional coherence in our bodies: a theory of organization explaining the arrival of the fittest, not just the survival of the fittest. As discussed here, we also don't really understand where it is that the body plans of organisms are stored, because DNA is a minimal “bare bones” language unsuitable for describing three-dimensional body plan blueprints; so while DNA might be storing a list of chemical ingredients, it does not actually seem to have a specification of our body plans.

We therefore have every reason to suspect that biological life involves some mysterious forces or factors very far beyond our understanding. If that is the case, then the prospect for vastly increasing the human lifespan through injecting humans with tiny robots called nanobots and tinkering with DNA may be much less bright than many think.

But there is some reasonable hope that genetic engineering or nanobots might extend the human lifespan, so this part of the TV show wasn't nonsensical. The show did become nonsensical when it discussed the very dubious concept of mind uploading in a completely uncritical manner. Among the reasons why mind uploading is such a dubious concept is that we have no understanding of how a brain could even be storing memories that last for decades, given all the rapid molecular turnover and protein turnover that occurs in brains; and we have no idea of how brains can generate abstract ideas. Mind uploading cannot possibly work unless some reductionist model of consciousness and memory is correct; and there are very good reasons (discussed here and in this series of posts) for doubting that any such model is correct.

The show quoted neuroscientist Michael Graziano, who made this very dogmatic proclamation:

A lot has to be done before we can figure out whether we can take a pattern of connectivity from a brain and upload it or copy it. It's going to happen; I'm positive of that. Everything's moving in that direction.

Graziano's certainty on this matter is laughable. There is no basis for believing that minds can ever be uploaded. There is no progress at all in mind uploading. Instead of “everything's moving in that direction,” the situation is actually, “nothing's moving in that direction.” And why is Graziano saying he's sure it's going to happen, just after saying, “A lot has to be done before we can figure out whether we can take a pattern of connectivity from a brain and upload it or copy it”? Those two thoughts contradict each other.

The show then quotes string physicist Michio Kaku, who nowadays seems to have his head popping up on every science or futurism documentary put on television. Kaku states, "Believe it or not, we can actually upload memories, and record memories in mice.” This statement is false. The actual claims that have been made by certain mice researchers (researchers in optogenetics) is that memories in mice can be blocked or activated. Such claims are not well founded, and are based on a few doubtful studies using a dubious methodology. The studies typically report weak levels of statistical significance. See here for a discussion of the flaws in some of these studies.

Making claims about mind uploading, the show's narrator says this:

Once we shed our physical bodies, we will exist only as a computer copy of our brains. It's hard to imagine exactly, but basically you'd be digital information stored on a server.

The TV show assures us that mind uploading is the key to eternal life. The narrator says this:

But when we start talking about immortality, that's a whole other can of worms. If we really are serious about getting to forever-land, we'll have to replace our carbon based human bodies and upload ourselves to a super-computer.

The show then tells us this existence will be like some afterlife heaven. We are told:

We're talking about everlasting life in a digital paradise....All of society will live digitally in a virtual world called the metaverse that's a thousand times more intense than the one you live in now.

There are quite a few reasons why such a thing is very unlikely to happen. First, the brain does not store digital information, so the mind is very likely not something that can be uploaded into a computer. Second, there is not the slightest evidence that the brain uses any type of readable code to store information. DNA uses the genetic code, something we can understand and read. But we have zero knowledge of anything like a brain code that we can use to read information from the brain. Nor can we plausibly imagine how such a code could have originated, as it would have to be something almost infinitely harder to explain than the genetic code (the origin of which is very hard to explain).

Third, you could never electronically capture the exact state of a particular person's brain, even if you tried to use microscopic nanobots to do such a thing. Imagine trying to exactly map every synapse and neuron in a brain with nanobots. Each of the brain-mapping nanobots would have to somehow be aware of its own exact position in the brain, so it can record the exact position of each neuron and brain connection it encounters. So if a nanobot comes to a neuron that is 1.334526 centimeters from the left edge of the skull, and 2.734538 centimeters from the right edge of the skull, and 5.292343 centimeters from the back of the skull, then those exact coordinates must be recorded. But how can a microscopic nanobot do that? You can't supply a microscopic nanobot with a little GPS system allowing it to tell its position. So it would seem that nanobots are completely unsuitable for any such job as mapping the exact physical microscopic structure of an organ with billions of cells and synapses packed together in a small space.

Then there's the duplication problem. Imagine if there was a machine that could scan your brain, and then upload your mind to a computer. Even if that process was done perfectly, the computer would not have your mind. It would instead have a copy of your mind. You may realize this just by considering that if this uploading process didn't kill you, and the computer with your “mind upload” existed at the same time as you, there wouldn't be two you's. There would be one you, and a copy of you living in the computer. If you then died after this upload, it would not at all be true that you had survived death by the fact that a copy of your mind was in the computer.

None of these difficulties are discussed on the “Year Million” TV show. The show presents mind uploading as a sure thing, ignoring all the reasons for thinking that it can't ever happen. Not only did the TV show present mind uploading as something that will likely happen; the TV show assured us that mind uploading is the key to true immortality – on the grounds that once you are living in a computer server, you are guaranteed to live forever.

This is nonsensical, because the programs running on computer servers sometimes crash and stop working; the programs running on computer servers are sometimes deliberately halted; and the computer servers themselves sometimes crash, lose power, or are turned off. If you were living in a computer server, there is no reason why you should be confident that you will be around for more than a century or two. 

mind upload
 After the father and daughter mind-uploaded

The next episode of the “Year Million” TV series continued to be largely about mind uploading, speaking as if this extremely doubtful idea was on solid ground. There was also a pitch for the idea that we are already living inside computers. The “Year Million” TV series is on track to be the silliest documentary series ever put on about the human future.

Given our very low state of knowledge about the most basic riddles of biology, memory and consciousness, fantasies of mind uploading are rather like the fantasies of some boy who knows almost nothing about the contents of the sun, but who fantasizes that he will one day reorganize the sun into a shape and color more pleasing to him.

Tuesday, June 13, 2017

Seven Levels of Mentality, and Why Brains Can't Explain the Last One

Can we account for all of man's mental activities by assuming that they can all be explained by the brain? When considering this question, it may make sense to distinguish between different levels of human mentality, with each level being more more difficult to explain than the previous level. For it might be that our brains are sufficient to explain one level, but insufficient to explain a more advanced level.

Below are seven levels of mentality we can distinguish.

Level 1: Let us imagine a very young baby lying in a crib near a window. A red bird flies and lands on the windowsill. The baby looks out at the bird, and perceives the bird's color. But there is no recognition, and the baby feels neither fear nor wonder at the sight of the bird. There is no thought or emotion, but merely sensory perception.


Level 2: Let us imagine the same event happens. The baby is lying in its crib, and a a red bird flies and lands on the windowsill near the crib. But let us suppose the baby is a little older, and now something a little more advanced happens. The baby feels an emotion of wonder or delight. But there is still no recognition, and no memory is involved.

Level 3: Let us imagine the same event happens. The baby is lying in its crib, and a a red bird flies and lands on the windowsill near the crib. Let us imagine this baby is now almost a toddler. The baby feels an emotion of delight or wonder at seeing the bird, but now some memory is involved. Without thinking any words, the child has a vague feeling of recollection, the feeling that he has seen such a bird before.

Level 4: Now let us imagine the child can walk, and he sees the red bird out in his back yard. When he sees the red bird, not only is there a feeling of delight, and a recollection, but also a use of language. Now the child points to the bird, and says, “I see birdy.”

Level 5: Now let us imagine the child is seven years old. Now he sees the red bird in his backyard, and the sight produces not just emotion and language, but also abstract thinking. The child makes an observation, “Wow, that sure is a beautiful bird,” or perhaps asks a simple question such as “I wonder whether birds like that get tired when they fly.”

Level 6: Now let us imagine the child is now thirteen years old, and is old enough to engage in philosophical thinking. So when he sees the red bird, he asks an advanced question such as “I wonder how it is that birds or any other species ever started to exist,” or “Is it morally right for us humans to be putting birds in cages?”

Level 7: Now let us imagine the person is no longer a child, but is now sixty years old. He can write in depth about many advanced philosophical questions; he can feel many refined emotions; and he is quite capable of writing an entire book about birds. Moreover, he can instantly recall memories that happened five decades ago. So perhaps he remembers the day 50 years ago when he first saw a beautiful red bird in his backyard.

Now, which one of these levels can we explain by assuming brain activity? Some philosophers would deny that even Level 1 can be explained by brain activity. As rudimentary as Level 1 is, it is still an example of Mind, and some philosophers say that we cannot explain how Mind can arise from mere matter. But perhaps they are wrong, and perhaps we can explain Level 1 by assuming that it involves merely sensory perception, which parts of the brain might explain. Conceivably we also might explain Level 2 also by imagining that some kind of hormone or chemical produces a feeling of wonder or delight.

But we cannot explain Level 7 through brain processes. For we cannot explain any way in which the brain could store memories for decades; we cannot explain how brains could instantly recall distant memories; and we cannot explain how a brain could generate abstract thoughts.

Consider the storage of memories. It has been proven through the work of scientists such as Bahrick that humans can store memories fairly reliably for more than 50 years. In order for you to have a workable theory for how brains can be storing memories, a neuroscientist would need to plausibly explain how human memories could be stored in a brain for 50 years. No scientist has done any such thing.

The main theory for how the brain stores memories is the idea that our synapses store memories. But there is a reason why this theory does not work. As discussed here, synapses are subject to very strong molecular turnover and structural turnover which should prevent them from storing memories for longer than a year. The proteins in synapses have an average lifetime of only about a week.  As a scientific paper puts it, "Thus, the constituent molecules that subserve the maintenance of a memory will have completely turned over, i.e. have been broken down and resynthesized, over the course of about 1 week."

Not only do neuroscientists fail at explaining how very long-term memories can be stored in the brain; they also fail to explain human memory retrieval. The two main problems in explaining human memory retrieval are these: (1) the problem of explaining how a human being could find the exact location where a memory might be stored in the brain; (2) the problem of explaining how we are able to recall obscure and old memories with such blazing speed.

If human beings took 30 minutes to retrieve a memory, the difficulty would not be so great. We might imagine that a brain simply scans all of our memories, looking for the right one, like a person flipping through the pages of a book looking for a topic. But given the instantaneous memory recall of distant, trivial memories shown on TV shows such as Jeopardy, it is obviously not true that you read through all your memories until you find something.

How are computers able to retrieve information so quickly? Through indexing, which requires sorting. But there is zero evidence that the brain does any type of physical sort or physical indexing. The brain does not have the type of architecture to support physical sorting or indexing. So our memories cannot be indexed in the way that computer information is indexed.

In short, there is no viable explanation as to how a brain could be able to find so quickly a spot in the brain where a memory was stored. Nor does it work to claim that memories are stored everywhere in the brain. The idea that every memory is stored everywhere in the brain is nonsensical.

Although neuroscientists are very bad at explaining how a brain could store or instantly retrieve memories lasting for decades, neuroscientists are even worse at explaining how brains could possibly be generating ideas and abstract thought.

If you go to this page on the “Expert answers” site quora.com, you will be treated to some answers illustrating the utter failure of modern neuroscience on how a brain could possibly generate new ideas. The page gives some answers to the question, “How does our brain form new ideas?”

The first answer by Tanush Jagdish begins by saying, “We don't know,” but then suggests “synaptogenesis,” the creation of new synapses. This is not a plausible idea. Some text such as “tall blue cold triangle” can instantly create an idea in your head that you have never had before, and you certainly did not have to wait for new synapses in your brain to form before you had such an idea.

The second answer by Jeff Nosanov is a circuitous answer that explains nothing, while claiming incorrectly that “ideas cannot be made to happen.” Nosanov ends by saying “that was not a physiological explanation.”  Then the page has some answers by non-scientists which are not illuminating.

A similar page on the “Ask science” sub-reddit at www.reddit.com offers equally little illumination. A Google search for “how the brain generates ideas” will result in a vast wasteland of results failing to offer a single neuroscience study offering any real illumination on this topic. One of the items you will get is a Harvard news story with the title “How the brain builds new thoughts.” But the story is discussing some research that does nothing to explain such a thing – just another one of those oh-so-dubious brain scanning studies in which some scientists scan brains and find trivial differences in blood flow (see here on why such studies are typically of little value).

Essentially the only idea that neuroscientists have to explain a brain creating ideas is some idea of combination, kind of the idea that you create a complex idea by combining simpler ideas. This does not explain the miracle of abstract thought. Let us imagine a savage who experiences 100 cold days, and who then reaches the abstract concept of coldness. This idea is not reached from any type of combination – it is reached by abstraction. Similarly, a person who sees 100 other humans may reach the abstract idea of a human being. But that idea is not reached by combination.

Computers offer not the slightest clue as to how abstract thinking can occur, because no computer has ever had a concept, an idea, or an abstract thought, something that requires a conscious mind. Don't be fooled by the type of computer program called an idea generator. Such programs are typically just programs for combining words into novel combinations. Until a human reads the output of such a program, an idea is not actually created. 

In short, there is no plausible brain explanation for how humans could be storing memories for 50 years; there is no  plausible brain explanation for how humans can instantly retrieve distant memories; and there is no plausible brain explanation for how humans can even create abstract ideas.  See here and here for additional support for these claims. It is sometimes argued that we need to postulate something like a human soul or spirit (something beyond the brain) to account for paranormal phenomena such as ESP and near-death experiences.  It may also be argued very forcibly that we need to postulate something like a human soul or spirit to account for certain types of normal mental functioning that we observe every day. 

Postscript: Just after finishing this post, I read this new scientific article, which quotes a neuroscientist speculating about memory storage. The article says:

Neuroscience has also been struggling to find where the brain stores its memories. “They may be ‘hiding’ in high-dimensional cavities,” Markram speculates. 

High-dimensional cavities? Cavities are holes, not information storage media.  I think the quote bolsters my claim that scientists do not have any plausible explanation of how brains can be storing memories for 50 years.