How do atoms build thoughts and emotions?

[A.T.]

A subjective experience is not, at least definitionally, the same thing as its objective manifestation in the brain. If you want to make that statement, you will have to argue it.

Yes, I am actually arguing exactly that. For me, consciousness and subjective experience are emergent phenomena of the physical brain state. So even the subjective experience is a manifestation happening due to physical processes in the brain.

"Is a manifestation happening due to" is different from "is the same thing". So you are not "arguing exactly that" here.

 

[A.T.]

I see subjective experience the same way—an emergent pattern that can be compared at a higher system level, even if the underlying individual neurons are arranged differently.

You can compare brain activity patterns on all levels you want. But this is not the same as comparing subjective experiences.

Scanning the retina only shows the input. The fMRI scans of the visual cortex show the actual processing of that input.

There is no such clear cut distinction. The retina cells are also already doing processing.

And the process is what defines the subjective experience.

How do you know that those neurons, that the machine learning model identified here, weren't just doing some basic pre-processing instead?

The point of the study isn't that the brain region is doing a different task. The point is that it is doing the same task (processing the concept and sensation of color), just receiving the data from the ears instead of the eyes.

Or, the blind-from-birth subjects were simply processing audio-input in the location where seeing subjects are processing visual-input, because their brains developed differently to make use of that location, which would otherwise be idle. Just because it happens in the same location in two different brains, doesn't imply that it objectively gives the same subjective experience.

Just think about this part: "devices that translate color into sound". There is no objectively unique way to do that translation. So why should we assume the resulting subjective experiences of the blind-from-birth was the same as seeing people?

You seem to be putting subjective experience in a box that is fundamentally isolated from physical measurement.

Yes. We can develop objective proxy measures, and assume they indicate the subjective experience. But won't be able to check that they actually do objectively.

 

[wrobel]

Another fun question to ask is: what about free will?

 

[Motore]

"Is a manifestation happening due to" is different from "is the same thing". So you are not "arguing exactly that" here.

Fair point, my wording was a bit sloppy there. Let me be precise: I am arguing they are exactly the same thing. The subjective experience is the complex emergent pattern of the physical brain.

You can compare brain activity patterns on all levels you want. But this is not the same as comparing subjective experiences.

It is the same thing, if you accept the premise that the subjective experience is the emergent physical pattern. You are starting from the assumption that they are two fundamentally different things, which is exactly the assumption I am questioning.

Or, the blind-from-birth subjects were simply processing audio-input in the location where seeing subjects are processing visual-input, because their brains developed differently to make use of that location, which would otherwise be idle. Just because it happens in the same location in two different brains, doesn't imply that it objectively gives the same subjective experience.

I never claimed it gives them the exact same subjective experience as a seeing person. Of course the translation of color to sound is different than photons hitting the retina.

My point is that whatever they are experiencing, it has a physical footprint that we can measure and objectively compare to a seeing person. If they experience it differently, the complete physical brain state will reflect that difference. We are comparing the experiences by comparing the physical states.

Yes. We can develop objective proxy measures, and assume they indicate the subjective experience. But won't be able to check that they actually do objectively.

If subjective experience is fundamentally isolated from physical measurement, how does it cause your physical fingers to type about it on a physics forum?

If it influences the physical world, it must have a physical mechanism. As I said before, I see the brain as an incredibly complex AI machine. We simply lack the tools yet to read its entire state at once. But that is a technological limit, not a fundamental one. Once and if we overcome it, we will be able to interpret and objectively compare everyone's unique subjective experiences.

 

[A.T.]

It is the same thing, if you accept the premise that the subjective experience is the emergent physical pattern.

It is the same thing if you accept that it is same thing.

If subjective experience is fundamentally isolated from physical measurement, how does it cause your physical fingers to type about it on a physics forum?

Whether consciousness causes anything is another can of worms: the question of free will.

As I said before, I see the brain as an incredibly complex AI machine. We simply lack the tools yet to read its entire state at once.

You can read the entire state of AI-systems. How would you tell if they are conscious from that?

 

[gentzen]

If subjective experience is fundamentally isolated from physical measurement, how does it cause your physical fingers to type about it on a physics forum?

What you can measure is a physical state.
My own personal subjective experiences are not such quasi-static states, but more dynamical processes. Even when I perceive color, where I personally would prefer a more static experience. (Not sure whether this is related to me being color blind.)

From what you write, I get the impression that your subjective experiences are indeed more quasi-static:

For me, consciousness and subjective experience are emergent phenomena of the physical brain state.

For me, it is less that such quasi-static experiences are

happening due to physical processes in the brain

Instead, what I do experience subjectively really feels like dynamic processes.

---

Even before I learned physics at school, I wondered how those feelings could be compatible with a determined state at every point in time. When I learned special relativity, the extension of the "now moment" to all spatially separated spacetime-points felt like some sort of relief to me, because it made that perceived gap a bit smaller, that gap between what I experience and the mathematical description of our physical world.

 

[gentzen]

My own personal subjective experiences are not such quasi-static states, but more dynamical processes. Even when I perceive color, where I personally would prefer a more static experience.

Actually, this seems to be related to me looking at a computer screen. If I just look at a normal object and let my eyes adapt a bit, I do get a quasi-static experience.
I think when I was younger, before I stared at computer screens day and night, this overly dynamic less quasi-static experience only happened when it was really dark (after my eyes adapted to the darkness), or when it was "too" bright.

 

[Motore]

It is the same thing if you accept that it is same thing.

Yes comparing subjective experiences is the same as comparing brain patterns (in the crude sense of the word). Clearly you do not accept that subjective experiences are emergent physical pattern. I do.

You can read the entire state of AI-systems. How would you tell if they are conscious from that?

I do not know, but this is actually a massive area of current research. For example a proposed checklist for how to examine an AI's state and architecture for those exact emergent properties:
https://arxiv.org/abs/2308.08708

So the scientific community is treating it exactly the way I am describing: as an objective, measurable, and structural property of a highly complex system. It is a mathematical and physical problem, not an unmeasurable subjective box.

 

[Motore]

What you can measure is a physical state.
My own personal subjective experiences are not such quasi-static states, but more dynamical processes. Even when I perceive color, where I personally would prefer a more static experience. (Not sure whether this is related to me being color blind.)

From what you write, I get the impression that your subjective experiences are indeed more quasi-static

I think we are just getting caught up on the word "state". When I say "physical brain state," I don't mean a frozen, static photograph. I mean it in the physical and computational sense.

Think of it like the AI machine analogy: when an AI processes a live video feed, its internal state is continuously updating, flowing, and computing. The brain is the same—neurons are constantly firing, action potentials are moving, hormones are fluctuating.

So it is an incredibly dynamic physical process. The fact that your subjective experience feels like a continuous flow perfectly matches the fact that the physical brain is a continuous, dynamic processor.

Even before I learned physics at school, I wondered how those feelings could be compatible with a determined state at every point in time.

But physics is fundamentally about dynamics. Whether it's classical mechanics, quantum wave functions, or relativity, physics is literally the mathematical description of how states evolve over time.

For me, there is no gap. The dynamic process you are subjectively feeling is the real-time physical processing happening in your neural network.

 

[Dale]

How would you objectively verify that the color we both call "red" looks the same to us both?

In the end, I don't think that we need to do that. If we imagine that we have some perfect "brain-ometer" that can completely characterize the physical state of the brain, then we can provide a stimulus in the form of a display of the RGB color 0xff0000, have the subject think about that color, and call that thought "red". At this point you have the brain state for the subject's "red" thought.

This is no different from a survey. In a survey you get a subject to think about something and then they record a mark representing their thought. Here you get a subject to think about something and then you record their brain state representing their thought. All we are doing, in both cases, is evoking a thought and recording some data to measure that thought.

We can survey multiple people about "red" without needing to know that their experience is the same. Similarly, you could "brain-ometer" multiple people about "red" without needing to know that their experience is the same. In fact, it would be a useful question to answer: is the "red" brain state the same between different people, or is it even the same within the same person under different conditions.

Especially the intra-subject comparison is the big problem. I can pretty reliably ask you survey questions about "red", but the accuracy and reliability of current "brain-ometers" is not as high.

 

[gentzen]

when an AI processes a live video feed, its internal state is continuously updating, flowing, and computing

Looks like "state" and "process" are not the only words that have different meanings for us. For me, an AI updates its internal state discretely, not continuously. And I also don't understand in which sense the internal state of an AI could be "flowing" for you.

The brain is the same—neurons are constantly firing, action potentials are moving, hormones are fluctuating.

For me, the hormone concentration is well described as a quasi-static state evolving in time. I don't know enough about action potentials to say whether describing them as "moving" makes sense for me or not. A neuron firing indeed feels like a dynamics process to me.
Of course, somebody deeply believing in "states evolve continuously over time" could invent some continuous evolution for that firing, and up to a certain level of detail this might even contain some truth.

Whether it's classical mechanics, quantum wave functions, or relativity, physics is literally the mathematical description of how states evolve over time.

In classical mechanics or relativity, yes. For quantum wave functions, you already have to ignore the collapse of the wavefunction. And for QFT, you will have a hard time trying to rescue this "states evolve continuously over time" picture.

I think we are just getting caught up on the word "state". When I say "physical brain state," I don't mean a frozen, static photograph. I mean it in the physical and computational sense.

Yes, you mean "state" in the "computational sense". But I don't see how that sort of state, which is essentially the content of the computer memory at a specific point in time, should give rise to subjective experiences.

The physical state has a part which can be described by such a computational state. But then, you have special relativity, general relativity, and QFT, and each of them makes it hard in its own way to really nail down the physical state exactly to such a computational state.
Which is fine, as long as you don't try to reduce physical processes to the continuous evolution in time of such computational states.

So on the one hand, we indeed seem to disagree about the meaning of words. But on the other hand, there also seems to be a deeper conceptual disagreement.

 

[Herman Trivilino]

We don't even know how life arises.

 

[Motore]

For me, an AI updates its internal state discretely, not continuously. And I also don't understand in which sense the internal state of an AI could be "flowing" for you.

I think you are taking my AI analogy a bit too literally. When I say the brain is an incredibly complex computing machine, I am not talking about a desktop CPU with discrete clock cycles and RAM. The brain is an analog computer. The chemistry and electrical potentials don't update in discrete ticks, they flow continuously.
And you are right that a static snapshot of memory doesn't give rise to subjective experience. It is the process of computation over time that does.

For me, the hormone concentration is well described as a quasi-static state evolving in time. I don't know enough about action potentials to say whether describing them as "moving" makes sense for me or not. A neuron firing indeed feels like a dynamics process to me.
Of course, somebody deeply believing in "states evolve continuously over time" could invent some continuous evolution for that firing, and up to a certain level of detail this might even contain some truth.

It is basic electrophysiology. An action potential is literally the physical flow of sodium and potassium ions across the cell membrane, moving down the axon. It is a continuous physical and chemical process that we can already describe perfectly using continuous differential equations.

In classical mechanics or relativity, yes. For quantum wave functions, you already have to ignore the collapse of the wavefunction. And for QFT, you will have a hard time trying to rescue this "states evolve continuously over time" picture.

Ok sure, but are talking about states evolution in the brain. The brain is a macroscopic, warm, biological environment where classical physics (electromagnetism, thermodynamics, chemistry) governs the mechanics.
At the scale of neural networks, the physical state does evolve continuously over time. The physical flow of ions, action potentials, and changing chemical gradients is exactly the continuous dynamic process I am talking about, and it acts perfectly as the "computational state" of this biological machine.

So you're right, there is a deeper conceptual disagreement. You seem to feel subjective experience cannot be reduced to these classical physical dynamics, whereas I see it strictly as an emergent product of them.

 

[PeroK]

We are quite protective of physics on this site. We don't like it when medical practitioners come here and tell us we are doing physics all wrong.

We should extend the same courtesy to neuro scientists and those genuinely researching the neurological aspects of consciousness. And not dismiss the whole endeavour on the basis of some homespun philosophy about the colour red.

 

[Dale]

My background is medical imaging, specifically MRI. So, my comments are not about protecting physics. My comments are about recognizing the current limitations of the technology:

Just because you have a cool device that can record brain signals doesn't mean that you have something that measures thoughts better than a survey does.

But the corollary is:

Thoughts can be measured, at least to the accuracy of a survey.

 

[Matterwave]

Fair point, my wording was a bit sloppy there. Let me be precise: I am arguing they are exactly the same thing. The subjective experience is the complex emergent pattern of the physical brain.

This is now a definitional argument (which is fine, it's good to get our definitions sharp). But for me, such a definition requires more and better motivation. For example, I would like to know how such a definition bears fruit and why it is more fruitful than a definition which separates the two.

In my post #32 I stated that my conception of physicalism did not require physics to explain everything, only that everything arises from the physical. I think the latter claim requires less and is more defensible.

I apologize if I am corrupting your argument. But I see many of your statements as saying the subjective experience supervenes on the objective state. I think such is a valid argument direction which, for me, does not benefit from the redefinition you are attempting here.

 

[PeroK]

My background is medical imaging, specifically MRI. So, my comments are not about protecting physics. My comments are about recognizing the current limitations of the technology:

Just because you have a cool device that can record brain signals doesn't mean that you have something that measures thoughts better than a survey does.

But the corollary is:

Thoughts can be measured, at least to the accuracy of a survey.

I wasn't thinking of your posts. I was arguing against the point that consciousness cannot be objectively researched.

Clearly, for example, you could correlate what people report about their conscious experiences with objective physiological measures.

 

[renormalize]

We can develop objective proxy measures, and assume they indicate the subjective experience. But won't be able to check that they actually do objectively.

It that's true, isn't discussing the "subjective experience" more akin to debating philosophy or religion instead of science?

 

[Motore]

For example, I would like to know how such a definition bears fruit and why it is more fruitful than a definition which separates the two.

For me, treating "subjective experience" and "objective state" as two separate things is just a red herring. Every experience anyone has ever had stems from processes in the brain. I think it's really up to those who want to separate them to explain why that separation is more logical.

As for why my definition is more fruitful: treating them as the same thing allows us to actually do science on it. It means consciousness can be measured, mathematically modeled, and objectively compared (like we are starting to do with fMRI and AI models). Separating subjective experience into its own isolated philosophical category doesn't add any explanatory power, it just makes it artificially immune to physics.

But I see many of your statements as saying the subjective experience supervenes on the objective state. I think such is a valid argument direction which, for me, does not benefit from the redefinition you are attempting here.

If by "supervenes" you mean the subjective experience can't change without a corresponding change in the physical brain state, then yes, we agree on the mechanics.

But if it perfectly supervenes on the physical state, then functionally, it is that physical state. I don't see the point in keeping a definitional wall between them. It doesn't help us actually figure out how the brain works.

 

[gentzen]

You seem to feel subjective experience cannot be reduced to these classical physical dynamics

What I wrote was:

Yes, you mean "state" in the "computational sense". But I don't see how that sort of state, which is essentially the content of the computer memory at a specific point in time, should give rise to subjective experiences.

You are right that "classical" physical dynamics can be reduced "too much" to the physical state and its evolution in time. And that I therefore find it unplausible that my subjective experiences would arise in that way.

Maybe a computer also has subjective experiences. But for that case, the passive observer model would be correct, because we know that our modern computers essentially operate deterministic.

As I argued before, the passive observer model seems unplausible for human subjective experiences. I have no idea whether the difference between physical states and physical processes is related to that. But I object when you use words or concepts so carelessly as to blur that difference.

My overall perspective is probably similar to the one put forward by Matterwave. But I want to be a bit more explicit about what I believe to be true (and would argue for), and where I am agnostic.

 

[Dale]

I was arguing against the point that consciousness cannot be objectively researched.

Yes. That is a point that is clearly contrary to a very large body of professional scientific research.

 

[Matterwave]

Separating subjective experience into its own isolated philosophical category doesn't add any explanatory power, it just makes it artificially immune to physics.

Thank you for the additional arguments and clarification!

If by "supervenes" you mean the subjective experience can't change without a corresponding change in the physical brain state, then yes, we agree on the mechanics.

Yes, this is the way I meant it.

But if it perfectly supervenes on the physical state, then functionally, it is that physical state.

Supervenience is not symmetric in general. For example, the words on the screen supervene on the pixels but not vice versa since you can move one pixel around without changing the word. But it is reflexive by definition.

I don't see the point in keeping a definitional wall between them.

'A' supervenes on 'A' is a tautology (supervenience is reflexive). So I at least see some benefits on keeping them separate so we can discuss them clearly.

But of course, a definition argument can not be ultimately "resolved" (I don't mean by consensus). If it is your contention that subjective experience should be defined the same as objective brain state, then that is fine. I understand your position better now. :)

 

[Motore]

But I object when you use words or concepts so carelessly as to blur that difference.

I don't see how I'm blurring the difference, I'm just using standard definitions. A single frozen state isn't a process, but the continuous evolution of those states is.

Maybe a computer also has subjective experiences. But for that case, the passive observer model would be correct, because we know that our modern computers essentially operate deterministic.

As I argued before, the passive observer model seems unplausible for human subjective experiences.

Why does being deterministic mean it has to be a "passive observer"? A computer program driving a car isn't passively observing, it's actively processing input and making physical changes in the world.

If the brain is a macroscopic, classical biological machine, it operates largely deterministically too. But that doesn't make us passive observers. The subjective experience isn't a ghost passively watching the brain's computation from the outside, it is the computation. It is the active processing.

 

[gentzen]

Why does being deterministic mean it has to be a "passive observer"?

The specific argument I gave why the passive observer model seems wrong doesn't work for a determinstic computer:

One reason why it seems to be wrong is that we do talk about our subjective experiences, and what we "objectively" do say about them doesn't seem to contradict what we "subjectively" did experience before.

If what our brain thinks about our experiences and what we really experience were really just coincidental, then it would be hard to explain why they wouldn't contradict each other more often.

Of course, I could try to find other arguments, which still work for a determinsitic computer. But if those arguments were totally unrelated to why I don't believe in the passive observer model, then they won't really help me.

 

[Dale]

We can develop objective proxy measures, and assume they indicate the subjective experience. But won't be able to check that they actually do objectively

We can do more than this. We can induce the subjective experience that we want to measure, and see how inducing that experience affects the objective measure.

 

[A.T.]

We can develop objective proxy measures, and assume they indicate the subjective experience. But won't be able to check that they actually do objectively.

If that's true, isn't discussing the "subjective experience" more akin to debating philosophy or religion instead of science?

Yes, but you cannot write that in a grant proposal. Therefore...

... the scientific community is treating it exactly the way I am describing: as an objective, measurable, and structural property