How the brain builds sentences, neuron by neuron

[John Clark]

It looks like the human brain and chatbots use the same basic method for generating and understanding language. The following article appeared in the June 17 2026 issue of the journal Nature: 

How the brain builds sentences, neuron by neuron

John K Clark    See what's on my list at  Extropolis

i7x

[Brent Allsop]

On Mon, Jun 22, 2026 at 1:14 PM John Clark <johnk...@gmail.com> wrote:

On Mon, Jun 22, 2026 at 11:47 AM Brent Allsop <brent....@gmail.com> wrote:

> The system must detect physical qualities when it is on, and not when it is off.  Objectively observing all that will tell you when it is on and when it is off,

Yes.

 

> but it will not tell you what it is like.

That is also true, but why is it true? The only answer I can think of is that a color qualia must not be "like" anything except itself. An explanation is supposed to make something that is complicated simpler, and you just can't get any simpler than a brute fact. After all, what is the alternative? If you discovered that a color qualia is "like" X then you'd want to know what X was "like", and if X was "like" Y then....

Yes, exactly.

It is a brute objectively and subjecdtively observable fact that glutamate behaves the way it does, in a synapse, because of it's redness quality.

(Note: glutamate behavior is just a temporary, easily falsifiable, standin.  But there must be something in the brain for which a fact like this will be true)

 

PS: why did you decide to take this off list? 

Oops, my mistake.  Thanks for pointing this out, and I'm putting back on the lists.

 

 John K Clark

On Mon, Jun 22, 2026 at 5:24 AM John Clark <johnk...@gmail.com> wrote:

On Sun, Jun 21, 2026 at 5:42 PM Brent Allsop via extropy-chat <extrop...@lists.extropy.org> wrote:

> AI does binding/computing with brute force discrete logic gates, which is obviously a far less efficient way than whatever we are doing with bound phenomenal qualities.

Unless you want to invoke magic, and I don't think you want to do that, then whatever we are doing must be logical; and in fact for years you seem to be demanding a logical explanation for the existence of qualia. And a good explanation means making something that is very complex and hard to understand less complex and easier to understand. And you can't get any simpler than turning on to off, or off to on. 

John K Clark

[John Clark]

On Mon, Jun 22, 2026 at 3:38 PM Brent Allsop <brent....@gmail.com> wrote:

 >there must be something in the brain for which a fact like this will be true

Not if that "something" is a color qualia and it is a brute fact, you may not like that but if it doesn't terminate then the only alternative is that an iterated sequence of "what" questions goes on forever, and you'd probably like that even less. Even pure mathematics can't start from nothing, it needs axioms that seem reasonable but have no proof.  

John K Clark

On Mon, Jun 22, 2026 at 1:14 PM John Clark <johnk...@gmail.com> wrote:

On Mon, Jun 22, 2026 at 11:47 AM Brent Allsop <brent....@gmail.com> wrote:

> The system must detect physical qualities when it is on, and not when it is off.  Objectively observing all that will tell you when it is on and when it is off,

Yes.

 

> but it will not tell you what it is like.

That is also true, but why is it true? The only answer I can think of is that a color qualia must not be "like" anything except itself. An explanation is supposed to make something that is complicated simpler, and you just can't get any simpler than a brute fact. After all, what is the alternative? If you discovered that a color qualia is "like" X then you'd want to know what X was "like", and if X was "like" Y then....

Yes, exactly.

It is a brute objectively and subjecdtively observable fact that glutamate behaves the way it does, in a synapse, because of it's redness quality.

(Note: glutamate behavior is just a temporary, easily falsifiable, standin.  But there must be something in the brain for which a fact like this will be true)

 

PS: why did you decide to take this off list? 

Oops, my mistake.  Thanks for pointing this out, and I'm putting back on the lists.

 

 John K Clark

On Mon, Jun 22, 2026 at 5:24 AM John Clark <johnk...@gmail.com> wrote:

On Sun, Jun 21, 2026 at 5:42 PM Brent Allsop via extropy-chat <extrop...@lists.extropy.org> wrote:

> AI does binding/computing with brute force discrete logic gates, which is obviously a far less efficient way than whatever we are doing with bound phenomenal qualities.

Unless you want to invoke magic, and I don't think you want to do that, then whatever we are doing must be logical; and in fact for years you seem to be demanding a logical explanation for the existence of qualia. And a good explanation means making something that is very complex and hard to understand less complex and easier to understand. And you can't get any simpler than turning on to off, or off to on. 

John K Clark

--

[Brent Allsop]

On Mon, Jun 22, 2026 at 3:18 PM John Clark <johnk...@gmail.com> wrote:

On Mon, Jun 22, 2026 at 3:38 PM Brent Allsop <brent....@gmail.com> wrote:

 >there must be something in the brain for which a fact like this will be true

Not if that "something" is a color qualia and it is a brute fact,

I must not be understanding what you mean here, because it seems obviously wrong to me.

It must be a brute fact that something must be behaving the way that we objectively observe it behaving, because of it's qualia or subjective quality.

 

--
You received this message because you are subscribed to the Google Groups "extropolis" group.
To unsubscribe from this group and stop receiving emails from it, send an email to extropolis+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/extropolis/CAJPayv24JYnvozYryyKzQbFNTq80jfuwHTtTWhnfqFvS-Brryw%40mail.gmail.com.

[John Clark]

On Wed, Jun 24, 2026 at 9:39 PM Brent Allsop <brent....@gmail.com> wrote:

>> Not if that "something" is a color qualia and it is a brute fact,

> I must not be understanding what you mean here,

Yes.

 

> because it seems obviously wrong to me. It must be a brute fact that something must be behaving the way that we objectively observe it behaving, 

Yes, and at the most fundamental level the "thing" that we can objectively observe is a simple switch. And an explanation means telling the same story in a way that is fundamentally simpler. So if you want a better explanation for the existence of a color qualia than the one I have provided then you're going to need to find something simpler than on and off. Good luck with that! And even if by some miracle you somehow succeeded in doing that then you'd be right back where you were before and asking "what causes qualia?".    

 

> because of it's qualia or subjective quality.

So you think a better explanation for the existence of subjectivity and qualia is by invoking the existence of something else that also has subjectivity and qualia? You don't get dizzy going round and round on that merry go ride? 

John K Clark

 

 

you may not like that but if it doesn't terminate then the only alternative is that an iterated sequence of "what" questions goes on forever, and you'd probably like that even less. Even pure mathematics can't start from nothing, it needs axioms that seem reasonable but have no proof.  

John K Clark

On Mon, Jun 22, 2026 at 1:14 PM John Clark <johnk...@gmail.com> wrote:

On Mon, Jun 22, 2026 at 11:47 AM Brent Allsop <brent....@gmail.com> wrote:

> The system must detect physical qualities when it is on, and not when it is off.  Objectively observing all that will tell you when it is on and when it is off,

Yes.

 

> but it will not tell you what it is like.

That is also true, but why is it true? The only answer I can think of is that a color qualia must not be "like" anything except itself. An explanation is supposed to make something that is complicated simpler, and you just can't get any simpler than a brute fact. After all, what is the alternative? If you discovered that a color qualia is "like" X then you'd want to know what X was "like", and if X was "like" Y then....

On Mon, Jun 22, 2026 at 5:24 AM John Clark <johnk...@gmail.com> wrote:

On Sun, Jun 21, 2026 at 5:42 PM Brent Allsop via extropy-chat <extrop...@lists.extropy.org> wrote:

> AI does binding/computing with brute force discrete logic gates, which is obviously a far less efficient way than whatever we are doing with bound phenomenal qualities.

Unless you want to invoke magic, and I don't think you want to do that, then whatever we are doing must be logical; and in fact for years you seem to be demanding a logical explanation for the existence of qualia. And a good explanation means making something that is very complex and hard to understand less complex and easier to understand. And you can't get any simpler than turning on to off, or off to on. 

John K Clark

--

[Brent Allsop]

Hi John,

Thanks for the continued help with this; I think we are making progress, and I now better understand what I'm still failing to communicate about objectively observing subjective qualities.

On Fri, Jun 26, 2026 at 5:17 AM John Clark <johnk...@gmail.com> wrote:

On Thu, Jun 25, 2026 at 6:15 PM Brent Allsop via extropy-chat <extrop...@lists.extropy.org> wrote:

> does a hole in a tape equal ON or OFF?

That question is of absolutely no importance because either will work equally well. The only important thing is that "on" is different from "off", and zero is different from 1, and having a hole in a tape is different from not having a hole in a tape.

It only doesn't make a difference if you rewrite the dictionary to be different.  For example,  If the dictionary says physical redness represents off and greenness represents on, and if you switch the two, there is an important difference in quality, or what it is like.  And recognizing this difference is important for objectively understanding what consciousness is like.

 

> It is a brute physical fact that some strawberries are 'red' (because they reflect light in a particular way).

No. That is NOT a brute fact because quantum mechanics can explain why certain chemicals in strawberries reflect red photons of light but absorb blue and green photons of light.  

> Whatever has a red (or redness) quality is just a brute physical fact 

OK, so what are we arguing about? A certain arrangement of on and off switches produces the redness quality.  

I wouldn't use the term "produces" because that separates qualia from objectively observable physical reality.   The functionalist only cares about what is causally downstream from the redness quality, and they fail to realize that the strawbery still has the red quality, which is different from the different type of light that is reflected.  i.e. the light isn't what has the red quality, it is the strawberry.  In other words, using "has quality" communicates better what we are talking about.  It behaves the way we objectively observe it behaving because of the redness quality we are only objectively observing. (doesn't tell you what it is like.)

All objective observation can detect is what is causally downstream from the physical behavior.  But the subjective binding does more than detect causally downstream behavior.  It must have some way to direclty aprehend qualities, as if they were all one subjectively bound gestalt direct awareness of all the qualities at the same time.  Sure, you can objectively observe the causally downstream behavior of all the qualities and the binding, but that doesn't tell you what any of it is like.

Once we demonstrate the true color qualities of stuff, and know the true color qualities of stufff, once we finally understand color (not just seeming color) then the objective will be subjective.  All we need is the dictionary between the subjective and the objective brute physical facts.

 

> It also must be a physical fact that something in the brain behaves as it does because of its subjective quality.

You're never going to get a good explanation for the existence of subjectivity by simply invoking more subjectivity, objectivity must come into the picture someplace! You want an explanation for the existence of the objectivity-subjectivity bridge, you want to start with something objective and end with something subjective, but postulating subjectivity at both ends of the bridge, as you're doing in the above, is not going to get you there. 

As I said before, a good explanation means breaking up something that is complicated and confusing into its simplest and easiest to understand parts, so you have only 3 possible courses of action:

1) Accept my explanation.  

2) Find something simpler and easier to understand than on and off (I don't think that's possible but you're welcome to try). 

3) Claim that brute facts do not exist because iterated sequences of "why" and "how" questions always go on forever. 

But if #3 is true then your investigations are not going to make any progress; or to put it another way, in the next thousand years you'll make exactly as much progress thinking about the nature of consciousness as philosophers have made in the last thousand years thinking about the nature of consciousness, and that would be ZERO. It would be a complete waste of time, pointless navel gazing.

What you are saying is that consciousness is not approachable via science.  You are saying there really are 'hard problems' that we will never solve.  There is a 4th option, neurons can do subjective binding, allowing us to directly apprehend subjective qualities.

We simply need to realize there is some objectively observable subjective binding which can directly apprehend these qualities.  They will soon be able to bind whatever has my elemental redness into your subjective experience, then you will know with infallible certainty the true color quality of things, and you will finally know if my elemental redness is like yours.  And you will know that AI systems just run on different voltages, which aren't like anything, so they aren't qualitatively conscious.

 

John K Clark

 

On Wed, Jun 24, 2026 at 12:37 PM John Clark via extropy-chat <extrop...@lists.extropy.org> wrote:

On Wed, Jun 24, 2026 at 10:29 AM Ben Zaiboc via extropy-chat <extrop...@lists.extropy.org> wrote:

> I just wanted to reiterate that nothing physical, like for instance a neurotransmitter, has the kind of 'qualities' we talk about when referring to colours (e.g. redness, etc.).This should be obvious when you consider how many types of red there are (many, many more than the number of neurotransmitters our brains use), then how many different types of green, and blue, and yellow, etc., etc. Then there are all the other sensations we are capable of experiencing.
There are probably more different types and variations of possible qualia than there are atoms in the universe, so they cannot possibly be inherent in or dependent on specific pieces of matter. There simply isn't enough of it /in the entire universe/, let alone in a single brain.
If you come up with some way for a small number of molecules to combine together in different ways, though, things are very different. Molecule x in molecular aggregate A can signify something entirely different from the same molecule in aggregate B. This is how an atronomical number of different proteins can be made from 22 amino acids. Unending complexity can be, and is, built from small numbers of simple things. We know this from physics, from biochemistry, and from biology. It's a pattern we see again and again in nature.
So rather than asking "what molecules have what physical qualities that are equivalent or relevant to mental states?" (which leads to the above problem of there not being enough molecules in existence), we need to ask "what simple processes in the brain can be combined in endless ways to produce the vast variety of mental states that brains can experience?". The answer is glaringly obvious, when you know how neurons work. To drive this point home, imagine that in the early days of computing, instead of using holes punched in cards or tape to represent zeros and ones which can be combined in various ways to produce a hierarchy of codes that could be assigned to an endless variety of meanings, we instead started with the meanings themselves, and used chemicals in small areas on the cards, with each meaning being assigned to a different chemical compound. Waving away the problems of producing and reading the different chemicals, how far would computing have got? Would "sodium borate = homesickness, silver nitrate = black with the faintest tinge of crimson, etc., etc., etc..." actually be practical? Complex things are made of simpler things. All the ordinary matter in the universe is made of neutrons, protons and electrons. All the staggering complexity of physical things boils down to that.
In the same way, all mental states, no matter how complex, are made of two pieces of information, ON and OFF, combined together in a huge variety of ways. This is what red is, this is what homesickness is. The infinitely large space of possible thoughts boils down to ON and OFF, nothing else. The physical properties of molecules are not only unnecessary, but totally insignificant in comparison.

I agree with all of that. 

John K Clark

[John Clark]

On Fri, Jun 26, 2026 at 10:46 AM Brent Allsop <brent....@gmail.com> wrote:

Hi Brent

>>> does a hole in a tape equal ON or OFF?

>>That question is of absolutely no importance because either will work equally well. The only important thing is that "on" is different from "off", and zero is different from 1, and having a hole in a tape is different from not having a hole in a tape.

> It only doesn't make a difference if you rewrite the dictionary to be different.  For example,  If the dictionary says physical redness represents off and greenness represents on, and if you switch the two, there is an important difference in quality, or what it is like. 

That is incorrect. If the color qualia labels for red light with wavelengths of 650 nanometers and green light with a wavelength of 550 nanometers where suddenly switched there would be no objective difference in behavior and there would be no subjective notice that something had changed either unless a memory had been retained about how those labels used to represent different wavelengths of light. And a new memory, regardless of whether the brain is biological or electronic, cannot be formed unless something objective has changed.

 

>>> Whatever has a red (or redness) quality is just a brute physical fact 

>> OK, so what are we arguing about? A certain arrangement of on and off switches produces the redness quality.  

> I wouldn't use the term "produces" because that separates qualia from objectively observable physical reality. 

How is that a separation? I thought you were trying to find a way to unite the objective and the subjective, in other words explain, in the simplest way possible, how one can produce the other.  

 > The functionalist only cares about what is causally downstream from the redness quality, and they fail to realize that the strawbery still has the red quality,

A strawberry has a 650 nanometer light reflecting quality and a 550 nm light absorbing quality, but that is NOT a qualia, it is the property of a chemical in the strawberry that can be explained by quantum mechanics. A strawberry is not in the qualia producing business, a mind is, and mind is what a brain does. And it's not important what label a brain uses to represent red light, the only important thing is that it is different from the label it uses for green and blue light.

>>> It also must be a physical fact that something in the brain behaves as it does because of its subjective quality.

>>You're never going to get a good explanation for the existence of subjectivity by simply invoking more subjectivity, objectivity must come into the picture someplace! You want an explanation for the existence of the objectivity-subjectivity bridge, you want to start with something objective and end with something subjective, but postulating subjectivity at both ends of the bridge, as you're doing in the above, is not going to get you there. 

As I said before, a good explanation means breaking up something that is complicated and confusing into its simplest and easiest to understand parts, so you have only 3 possible courses of action:

1) Accept my explanation.  

2) Find something simpler and easier to understand than on and off (I don't think that's possible but you're welcome to try). 

3) Claim that brute facts do not exist because iterated sequences of "why" and "how" questions always go on forever. 

But if #3 is true then your investigations are not going to make any progress; or to put it another way, in the next thousand years you'll make exactly as much progress thinking about the nature of consciousness as philosophers have made in the last thousand years thinking about the nature of consciousness, and that would be ZERO. It would be a complete waste of time, pointless navel gazing.

> What you are saying is that consciousness is not approachable via science. 

What I'm saying is that consciousness is not approachable via science unless you make the assumption that at least some iterated "how" questions do not go on forever but terminate in a brute fact. And that's OK because even pure mathematics needs to start with axioms, which is just a fancy name for assumptions. 

 

> You are saying there really are 'hard problems' that we will never solve. 

The "hard problem of consciousness" will never be solved but not because of anything I said in the above, the question will never be answered because it has never been clearly elucidated exactly what the question is, nor is it clear what sort of answer would be considered satisfactory.   

 

> There is a 4th option, neurons can do subjective binding, allowing us to directly apprehend subjective qualities.

Huh? I can already directly apprehend subjective qualities, I'm doing it right now, philosophers could do that too but that hasn't stopped them from arguing about consciousness for over 3000 years, and yet all that bafflegab has not resulted in the production of one particle of new knowledge. 

> you will finally know if my elemental redness is like yours 

NO. Brent Clark (or John Allsop) might learn something, but you and I won't, and even he won't learn anything new about how to bridge the subjectivity/objectivity gap without making a new assumption. 

 

> And you will know that AI systems just run on different voltages, which aren't like anything,

Well, the color qualia that I experience aren't "like" anything either except for themselves, so I and AI systems must be pretty similar. 

And by the way it's not just AI systems, the neurons in the human brain also fire because of differential voltages, the only difference is that the voltage is caused by potassium and sodium ions instead of electrons.

 

> so they aren't qualitatively conscious.

Brent, I'd really like to hear your ideas about how you think random mutation and natural selection managed to produce consciousness at least once and probably many billions of times.  I'd like to know how natural selection could select for something that it can't see unless it was the inevitable byproduct of intelligence. 

John K Clark