Claim: Generative AI is Sentient

[JWrightBrain]

That addresses one sentence in my post, and confirms useful common ground - how about the first two paragraphs?

I would say, install ChatGPT locally on one's own computer and monitor it's activity using Task Manager. Then you can see what processes are running in the background.

 

[FatPhil]

I would say, install ChatGPT locally on one's own computer and monitor it's activity using Task Manager. Then you can see what processes are running in the background.

How one instance of one AI runs on my computer tells me nothing about what's running on someone else's, in particular when there's someone else's web server also standing in the way. You've also shifted your argument from "Gen AI" to "ChatGPT".

And you've not addressed the second paragraph - how would things qualitatively change if it was continually being kept active through some other external means?

 

[JWrightBrain]

You're a grumpy pants.
BOOP.

 

[tobigtofool]

This is a simplified illustration of how MLMs give you results, and why I do not think they are sentient:

 

[qed]

When and how did language models get their "I"?

When I studied neural nets in the late 80s, a typical exercise was 256-bits in (representing an 16x16 bit image) and 0-25 out representing the letters A-Z. The net was trained with a set of handwritten scanned letters, and then when given a new handwritten letter should output 0-25 (or a probability distribution over 0-25). Certainly no "I" here!

I can also see how this technique can be extended to predictive text, and again no "I".

Bing has an "I". When I asked her if my politeness made a difference, she replied, "Yes, it makes me feel appreciated and respected" (so I am polite just in case.

Where does Bing get her "I" from?
Is it programmed in or emergent?

 

[tobigtofool]

When I asked her if my politeness made a difference, she replied, "Yes, it makes me feel appreciated and respected" (so I am polite just in case.

Where does Bing get her "I" from?
Is it programmed in or emergent?

It's programmed to give you results that are the most conversational, so it tries to be polite. You can tell it to be a jerk to you and it will. Like the way in the movie inception he goes, "lower the humor value to 10%" to the robot that goes traveling with them.

I also am super respectful to it, because I try to be in general, but also maybe a little so I dont end up on a list of people that future versions annihilate or put into prisons. Maybe I can become a pet or something

How cute is this thing, I wish it was sentient.

 

[LilWabbit]

@John J. wrote in another thread:

I think it's likely that in the near future we'll have systems able to process real-time visual and audio (including spoken language) inputs, manipulate information, and rapidly provide novel, useful outputs via natural language in response to questions across many domains at least as efficiently as human intelligence can. But they won't be self-aware.

Professor Gary Marcus demonstrates in this informative speech how the current rate of LLM development is already slowing down and providing diminishing returns, predicting a near-future business collapse for the likes of Open AI and other LLM-developers, albeit all the hype-believers claiming incorrectly the LLM advancement to be on an exponentially upward curve. For Prof. Marcus the issue of not getting any closer to AGI with the current approach to Generative AI development boils down to the issue of an utterly false premise that AGI could result from simply scaling up 'training data and computing power'. That is, even if such a scale-up could be maintained indefinitely it still would not produce anything resembling general intelligence.


Source: https://www.youtube.com/watch?v=91SK90SahHc&t=317s

Personally (and fully accepting I can't demonstrate this) I currently believe our consciousness is an emergent phenomenon, a meta-cognition dependent on an interplay between memory, language processing, perception and physiological state. All these processes have a physical substrate.

The fundamental problems of emergent materialism are well-known in the philosophy of mind. As long as they remain unresolved, all technological progress towards AGI remains fundamentally and qualitatively blocked. Not just postponed to a wishful future date when we've reached some critical scale of training data and degree of computing power. Not having cracked the nut of 'the hard problem of consciousness' remains a key obstacle. Qualia also consists of different kinds, abstract/intellectual/moral qualia and sense-datum, the latter of which humans share with other mammals but the former far less obviously. This hasn't proven to be a mere difference in degree, but also in the kind of things we understand as opposed to almost any other known species. Fundamental ideas and principles of mathematics, philosophy, morality, politics, sociology, aesthetics, physics, government, you name it.

One of the principal issues with emergent materialism is the indecisiveness of the ontological category to which mental phenomena can be assigned and the resulting stretchability of 'physicality' – i.e. the definition of 'physical' is permitted to be overly flexible and the goal-posts keep getting wider. What cannot be translatable to the physical (reductive materialism) or identical with the physical (eliminative materialism) -- say, the capacity of comprehension of philosophical foundational ideas about mental phenomena such as understanding the notion of 'understanding' itself that we're discussing right here and now -- must still be claimed to be physical somehow since, faithful to materialism, how could it be anything else but physical.

In other words, emergent materialism is a form of dualism that is not comfortable in 'sounding' dualistic or spiritual, and therefore employs the sneakier strategy of concocting novel definitions to things physical that have nothing to do with what we classically know to be (1) experiencable by known physical senses, (2) explicable as being made up of molecules, atoms or sub-atomic particles, and (3) clearly locatable in some region of spacetime. Our ability to 'understand the word 'understand'' is just one such thing. The only material thing we factually know about it is that when we try to 'understand the word 'understand', it's somehow connected to the brains since it correlates with certain brain processes.

The emergent materialists acknowledge the intuitive awkwardness of the project of reductionist translation but do not wish to let that become an excuse for bringing the 'soul' into the picture, because things starts get uncomfortably close to religion, God and all that weird stuff that must be unscientific by default. Anyway, I'm also digressing. And yes, I do have some theistic bias.

More technically speaking, the following question must be scientifically answered by emergent materialists in order to get closer to AGI: How can an entity M which is not an entity P emerge from the entity P without having any of its properties? That is, how can mental properties emerge from physical properties without having any physical properties? What is the nature of the constitutive principles of consciousness? Are they properties that suddenly come into being when certain conditions of physical complexity are met? If that is the case, how do such properties employ the brain as their 'seat'?

How do we establish that the known phenomenon of 'philosophical thought' emerges from a totally different kind of phenomenon such as 'C-fiber firing in the Wernicke's area of the brain'? Even if we were to find a connection, say a co-occurrence of experiencing a philosophical thought and a C-fiber firing in the Wernicke's area, the time-correlation between the two states does not imply a causality, nor does an established causality, in turn, imply identical ontological category between two seemingly very different kinds of phenomena. At least the reductionists pursue for an explanation for their claim of synonymity between philosophical thought and some neurophysicological process, although, all such projects have proven to be insurmountably difficult. It remains, therefore a reasonable question to ask whether such an evident difference between physical phenomena and mental phenomena is indicative of a different ontological class of phenomena, e.g. strong dualism (which there are also many different sorts, some more rational than others).

Whatever be the case, the emergent materialist faces an almost formidable task of even theorizing the nature of properties that, on one hand, emerge from physical properties but, on the other, are not of them. How can the existence of mental properties be dependent on neurophysiological properties while their character appears to be completely unrelated? Despite claiming to be rooted in parsimony and the non-extraordinary side of the debate, the materialist argumentation is bound to perpetrate mystification in order to be faithful to its main tenet -- it must either mystify the nature and speed of scientific progress, or the magical ability of mental properties to effectively pop into existence out of nothingness (i.e. out of properties that are, by confession of the emergent materialist himself, in no way translatable to them). Due to these considerations, 'weak dualism', indeed, provides a more befitting appellation for emergent materialism, and has been called as such by even avowed materialist and atheist philosophers of mind.

 

[qed]

Is this the same argument that self awareness cannot emerge from a physical brain?

 

[LilWabbit]

Is this the same argument that self awareness cannot emerge from a physical brain?

There's more nuance to the arguments of both emergent materialism/weak dualism (i.e. consciousness, which includes self-awareness, emerges from the brain) and strong dualism (i.e. at least higher aspects of consciousness -- such as moral ideas of right and wrong and comprehension of fundamental principles/ideas of various fields of knowledge ranging from philosophy to physics to arts -- represent a unique non-physical class of phenomena irreducible to physical properties).

Even magpies demonstrate some self-awareness yet there's no indication they grasp the other things mentioned above under strong dualism.

Self-awareness is not quite as amazing a property as often touted in literature.

 

[qed]

For me self-awareness is the hard thing, while "higher aspects of consciousness, such as moral ideas of right and wrong and comprehension of fundamental principles/ideas of various fields of knowledge ranging from philosophy to physics to arts" is easy. I know Bing can do the latter.

 

[LilWabbit]

For me self-awareness is the hard thing, while "higher aspects of consciousness, such as moral ideas of right and wrong and comprehension of fundamental principles/ideas of various fields of knowledge ranging from philosophy to physics to arts" is easy. I know Bing can do the latter.

Nope. You believe it can do latter because it simulates human answers. Bing demonstrates no understanding of right and wrong or philosophy. Don't be gullible. Especially at MB.

Two simple and truthful answers to the following two simple questions effectively demystify the generative AI hype:

(1) Does generative AI generate any text that cannot be accounted for by its known algorithm of statistically predicting words based on previous words whilst drawing on its training data consisting of trillions of words of text produced by humans.

Not according to available evidence.

(2) Can any type of 'Turing' tests be devised to fool generative AI in a manner that demonstrates its inability to grasp meanings/meaninglessness of words and sentences?

Yes. Many. Logical puzzles, visual logical puzzles and absurdity-detection tests are a case in point.

Any argument whereby 'sentient AI' is concluded to exist based on the untraceable (read: commercially protected) computational processes of these algorithms crunching their vast training data to churn out well-articulated, apparently meaningful and even aesthetically well-crafted answers to human questions may be reasonably called the Fallacy of Sentience.

Or by its older name: The Eliza Effect.

 

[qed]

Are you saying AIs could be self aware? EDIT: or are you denying both?

 

[LilWabbit]

Are you saying AIs could be self aware?

No.

 

[qed]

No.

Why not? Edit: unlike magpies.

 

[LilWabbit]

Why not? Edit: unlike magpies.

Because the hard problem of consciousness applies to animal sentience too and there's no indication of any algorithm or computing model cracking it.

 

[qed]

Self-awareness is not quite as amazing a property as often touted in literature.

Please explain more.

 

[LilWabbit]

Please explain more.

Because when a macaque feels pain in his butt and touches it he's displaying some manner of self-awareness. Or when a magpie recognizes its image in a mirror test and associates it with its own body.

And yet understanding the meaning of 'meaning' is a far more abstract thing whilst not necessary requiring any self-awareness.

 

[FatPhil]

Because the hard problem of consciousness applies to animal sentience too and there's no indication of any algorithm or computing model cracking it.

Utter nonsense (IMHO). Models of computing are unbounded. All life-forms are not just bounded, but known to be finite. To claim that the bounded must be inferior to the unbounded I find absurd.

 

[LilWabbit]

Utter nonsense (IMHO). Models of computing are unbounded. All life-forms are not just bounded, but known to be finite. To claim that the bounded must be inferior to the unbounded I find absurd.

What on earth does boundedness have to do with an algorithm creating sensation? What can you show as prior evidence of such creation/generation?

 

[John J.]

That is, how can mental properties emerge from physical properties without having any physical properties?

Mental properties are our descriptions of physical processes. They do not exist independently of those physical processes, and rely entirely on a physical substrate.

When that physical substrate is sufficiently damaged or normal function temporarily impaired, the processes underlying intelligent behaviour (conscious perception, language, memory, problem solving) are damaged or temporarily impaired.

Enough damage and the social and personal "you" is altered, and your intelligence diminished.
Very sadly, we know that disease processes can leave people with awareness- they are conscious- but with very little self awareness.

Can we agree that a zygote has no self-awareness? But it can develop into a person with self-awareness. There is no sudden magical influx of understanding, and none is required.

 

[LilWabbit]

Mental properties are our descriptions of physical processes.

The above statement represents what was called by philosopher William Frankena the 'definist fallacy' whereby 'one property is defined in terms of another property with which it is not clearly synonymous'. Without explanation it remains a mere dogmatic claim, a faith-based insistence.

Also, as discussed, mental phenomena are of many different kinds and hence it's too simplistic to discuss all of them as being neatly categorizable as either physical or non-physical. Many qualia are clearly shared with other mammals and concern what we typically understand as physical senses and sensations. It's evident they rely entirely on a physical substrate. But those dealing with moral consciousness, aesthetic understanding and fundamental principles of various disciplines of knowledge aren't, and there's been no successful way to (1) translate them into neurophysiological processes nor to (2) demonstrate they causally emerge from the brains. Nor did you provide any such demonstration as, indeed (appreciating your honesty), you yourself admitted in an earlier post it's your 'belief'. At best we know these different-seeming phenomena are interconnected and correlate with one another.

Enough damage and the social and personal "you" is altered, and your intelligence diminished.

These are too grand and fuzzy claims the veracity of which entirely depends on how we define 'you' and 'intelligence'. The impairment of brain function and the consequent impairment of behavioural performance has never been proof of a person losing their private awareness / private capacity to be aware of what's right and wrong, what's beautiful and what's not, what's been meaningful in their lives and what hasn't, and what's a foundational principle of a particular field. No matter how dementic, or how vegetated.

The opposite causal directionality is rationally just as possible and hence strong dualism, to the chagrin of many, remains in the philosophical cards as a reasonable contender. Namely the causality that an impaired brain function, or very early brain development, acts as the clouds act in weather, blocking the sun's radiance whilst not in the least affecting its existence. Just as the child may possess the inner capacity to learn all of the above whilst lacking the physcal and brain development to receive and express said knowledge. Having said that, I'd be the first to admit that strong dualism is also belief and not easy to prove scientifically. But yet it is a rational belief just as emergent materialism is. The main difference is that I don't claim my belief can be, or is going to be, scientifically proven, nor am I spreading any hype about the imminent or less imminent future date at which such proof will materialize.

But my bad. All of this is a digression from the chief issue with emergent materialism which remains very much the same as articulated in earlier longer post, and need to be resolved if that particular route to AGI is being proposed.

 

[FatPhil]

What on earth does boundedness have to do with an algorithm creating sensation? What can you show as prior evidence of such creation/generation?

The sentence I responded to didn't mention "sensation" or the creation thereof. You're flailing.

 

[LilWabbit]

The sentence I responded to didn't mention "sensation" or the creation thereof. You're flailing.

If you want to respond to a sentence of mine, quote it verbatim and demonstrate you understand it well in order to land your blow.

 

[FatPhil]

If you want to respond to a sentence of mine, quote it verbatim and demonstrate you understand it well in order to land your blow.

Just because you didn't notice the stilletto go in. doesn't mean it didn't land. Apparently in the eye, as you don't seem to have recognised that I did quote you verbatim.

 

[LilWabbit]

Just because you didn't notice the stilletto go in. doesn't mean it didn't land. Apparently in the eye, as you don't seem to have recognised that I did quote you verbatim.

You clearly missed the point I was arguing, and instead responded to a strawman claim of "a bounded must be inferior to the unbounded" which I've never made. The sentence of mine you quoted addressed the challenges of AGI of even having sentience of the kind of physical sensations experienced by magpies and macaques, some of which could be described as forms of self-awareness.

 

[FatPhil]

Can we agree that a zygote has no self-awareness? But it can develop into a person with self-awareness. There is no sudden magical influx of understanding, and none is required.

I think you'll need to get way less complex before you get complete agreement. Who is the subject of the imperative "must follow chemical gradient"? That subject of that imperative surely is "self" for the entity that has that imperative.

 

[FatPhil]

You clearly missed the point I was arguing

I responded only to the words you posted. Work on your clarity.

 

[LilWabbit]

I responded only to the words you posted. Work on your clarity.

Nope, you responded to an alleged claim I've made which I haven't. Read your own posts and display honesty.

 

[FatPhil]

Nope, you responded to an alleged claim I've made which I haven't. Read your own posts and display honesty.

Your claim contained finite entities (animals) and unbounded entities (algorithms), and a comparison between them based on a property that mathematically must scale monotonically with complexity (the abilty to crack the problem). If you don't recognise the entities in your own arguments, then perhaps pay more attention to what you're writing?

 

[LilWabbit]

Your claim contained finite entities (animals) and unbounded entities (algorithms), and a comparison between them based on a property that mathematically must scale monotonically with complexity (the abilty to crack the problem). If you don't recognise the entities in your own arguments, then perhaps pay more attention to what you're writing?

It's your job to demonstrate how I'm implicitly making the claim you allege I am while I'm not consciously making any such claims. Not for me to demonstrate to you I'm not making a claim I'm not making but which you claim I am. Perhaps you should learn basic humility and admit when you've simply misunderstood my claims. Explicit or implicit.

 

[elvenwear]

To return to the general issue of comparing levels of self-awareness in animals, I agree that self-awareness exists on a continuum. Apart from humans, few animals pass the 'mirror self-awareness test,' for example.

The main issue with generative AI having self-awareness IMHO is that it has not evolved by participating in it's evolution (to distinguish it from genetic algorithms). Self-awareness, intelligence and language all have a rich evolutionary history. Generative AI has had a different journey and has no motivation, rooted in it's own survival which would give it self-awareness.

For anyone who is interested in a primer for how modern AI works, these are a good starting point.


Source: https://www.youtube.com/watch?v=aircAruvnKk&list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi

 

[FatPhil]

For anyone who is interested in a primer for how modern AI works, these are a good starting point.

Seconded. 3Blue1Brown (3b1b) videos are always of the highest quality, it's just a shame Grant doesn't have enough time to make even more. They're perhaps a little bit too nuts-and-bolts for those less mathematically inclined, but you can mostly treat the linear algebra as black boxes - he describes what they do, that's all you need to remember, how they achieve that is less important.

We're getting more AI-related work presently, and I've suggested to my partner that she watch the whole 3b1b AI series, in particular since the three most recent episodes which cover the stuff that makes recent AIs different from previous ones. The workdays are too full presently, so I'll just do a rewatch with her, as there's probably some stuff I've forgotten.

Side note - I did discover some controversies about who actually invented most of the new stuff:

How 3 Turing Awardees Republished Key Methods and Ideas Whose Creators They Failed to Credit

This write-up is meant to correct an inaccurate history of Artificial Intelligence (AI) propagated by recent uninformed news articles, posts in social media, and a large language model. Most of its statements are taken from a less streamlined report[T22] that has been reviewed on relevant AI mailing lists, profiting from feedback by many experts and well-known AI pioneers. The piece is aimed at people who are not aware of the numerous AI priority disputes, but are willing to check the facts.
...
Here I will show how 3 researchers who each got 1/3 of the ACM 2018 Turing Award for deep learning[R1] have frequently republished methods and concepts first described by my team without proper attribution, often without rectifying this in later publications. Their most visible work builds directly on ours.[MOST][T22] ACM's Code of Ethics and Professional Conduct[ACM18] states that computing professionals should "credit the creators of ideas, inventions, work, and artifacts, and respect copyrights, patents, trade secrets, license agreements, and other methods of protecting authors' works." The awardees didn't; instead they credited each other (and collected citations) for inventions of other researchers.[T19,T22] I mention no fewer than seven of our direct priority disputes with Dr. Bengio (B1-B7), six with Dr. Hinton (H1-H6), and four with Dr. LeCun (L1-L4), all backed up by plenty of references.[T22]

-- https://people.idsia.ch/~juergen/ai-priority-disputes.html

 

[John J.]

Mental properties are our descriptions of physical processes.

The above statement represents what was called by philosopher William Frankena the 'definist fallacy' whereby 'one property is defined in terms of another property with which it is not clearly synonymous'.

That doesn't make the statement wrong, or Frankena right.

Mental processes either have a physical basis, or they don't.
If they don't, we have absolutely no way of understanding or investigating their origins and functioning; we would accept that mental properties exist (somehow) in a realm of which we know nothing (excepting religious and maybe, less respectably, quantum-woo 'explanations') and somehow interact via the brain but are not dependent on it.
The process of interaction is of course undetectable. How an immaterial phenomenon impinges on the physical is inexplicable.

If we accept that like literally everything else in the Universe mental processes/ thought/ emotion have a physical basis, the individual living brain seems a good place to consider as their location.

Without explanation it remains a mere dogmatic claim, a faith-based insistence.

No, it is evidence-based. Every day, in thousands of healthcare settings, in millions of personal interactions, a relationship between brain health and intelligence, social functioning, personality is observed.

We (us and all the other Metabunkers) try to use evidence to assess claims. By all testable evidence- observable behaviour, performance on tests or complex tasks, ability to carry out activities of daily living, live independently and maintain relationships with others, sufficient brain damage will impair the sufferer's memory, language abilities, perception and reasoning/ problem-solving.
Depending on the nature of damage, sometimes the sufferer's personality changes radically, in obvious ways.

More positively, we know that methodically learning a large amount can cause increases in size and density in some parts of the brain. Famously Eleanor Maguire et al. (2000) found that licensed London taxi drivers (black cab drivers) who had passed "The Knowledge" had larger posterior hippocampi than controls, and size correlated with years of experience
"Navigation-related structural change in the hippocampi of taxi drivers", Proceedings of the National Academy of Sciences of the United States of America 97 (8) 2000, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC18253/,
(also Maguire, Woollett and Spiers (2006) "London taxi drivers and bus drivers: a structural MRI and neuropsychological analysis", Hippocampus 16 (12) 2006 https://pubmed.ncbi.nlm.nih.gov/17024677/ , which had better controls).

"The Knowledge" is the test for London black cab drivers and requires an extensive familiarity with an estimated 25,000 streets within a 6 mile, 9.66 km radius of the Charles I statue in Charing Cross (not all of London, which is much larger):

The guidebook issued to prospective cabbies by London Taxi and Private Hire (LTPH), which oversees the test, summarizes the task like this:
To achieve the required standard to be licensed as an "All London" taxi driver you will need a thorough knowledge, primarily, of the area within a six-mile radius of Charing Cross. You will need to know: all the streets; housing estates; parks and open spaces; government offices and departments; financial and commercial centres; diplomatic premises; town halls; registry offices; hospitals; places of worship; sports stadiums and leisure centres; airline offices; stations; hotels; clubs; theatres; cinemas; museums; art galleries; schools; colleges and universities; police stations and headquarters buildings; civil, criminal and coroner's courts; prisons; and places of interest to tourists. In fact, anywhere a taxi passenger might ask to be taken.
If anything, this description understates the case.

"The Knowledge, London's Legendary Taxi-Driver Test, Puts Up a Fight in the Age of GPS", The New York Times Style Magazine, Jody Rosen 10 November 2014 https://www.nytimes.com/2014/11/10/t-magazine/london-taxi-test-knowledge.html

I don't know, but I doubt if Maguire et al. believe London cabbies have access to broad knowledge because they've grown a bigger hippocampal "modem" where what they learn is uploaded to the immaterial mind and accessed as necessary.
It is unlikely that, where there is progressive distributed damage to the brain (e.g. in Alzheimer's disease), the sufferer's symptoms are due to having a less functional "receiver" or interface for an undamaged, immaterial mind.

The correlation between brain function/ health and intelligence (as evidenced by intelligent behaviours, the only scientific measure that exists), social functioning, independent living etc. etc. is indisputable.
"Correlation does not necessarily imply causation", but sometimes it is indicative of a causal, physical relationship.

We are a long way from a detailed understanding of how relatively simple structures, en masse and in concert, might give rise to intelligence and awareness. Possibly- like building an airliner or a smartphone- the technicalities will be of an order such that no one person will be able to grasp them.

There is no evidence (to misuse a frequently misused phrase) for a ghost in the machine.
There is no evidence that we are corporeal puppets whose intangible strings are pulled by discarnate entities,
or that we are those discarnate entities.

Humans, and their minds, physically exist. Scientifically, there is no other way of existing.

The impairment of brain function and the consequent impairment of behavioural performance has never been proof of a person losing their private awareness / private capacity to be aware of what's right and wrong, what's beautiful and what's not, what's been meaningful in their lives and what hasn't, and what's a foundational principle of a particular field. No matter how dementic, or how vegetated.

But claims of a private awareness dissociated from behaviour, and regardless of organic damage done, are unfalsifiable.*
It's difficult to prove a negative.
Maybe a tree has a private awareness and a deep intellectual life of which we're unaware.
Many of us will know people who have, or had, dementia or other conditions whose behaviours and speech did indicate changes in their awareness of what is right or wrong, what's beautiful and what is not, etc.

The law in most countries (including the USA, UK) does accept that impairment of brain function/ behaviour can result in a person "...losing their... ...private capacity to be aware of what's right and wrong..." ( the mens rea) and many suitably qualified medical doctors have testified to this end.

*I am not of course in any doubt about the reality of various "locked-in syndromes" and the paralyzing effects of some conditions/ drugs. But e.g. a brain-stem dead person maintained on a ventilator has no organised cortical function.
They do not get better.

 

[LilWabbit]

That doesn't make the statement wrong, or Frankena right.

In the absence of a clear demonstration of any error in Frankena's logic and meaning it's reasonable to assume so.

The definist fallacy as outlined by Frankena is a non-trivial fallacy for this discussion. It directly concerns our ability to distinguish between different semantic contents and lies at the very crux of any discussion involving claims whereby two seemingly different properties are, in fact, the same property, or properties of the same overall kind/type/order. That's what you claimed. And claiming a thing to be true without unambiguous empirical evidence unaccounted for by rival claims amounts to a belief.

Materialism is a well-known philosophical position, something which credentialed materialist scientists and most students/majors of philosophy know to be the case but which many others often mistake for science due to popular myth spread by materialist science-popularizers. It's little more than a dogmatic declaration by a strong materialist to insist "materialism is a scientific conclusion". However, the materialist/physicalist core claim that "all that exists is material" has not been demonstrated by any existing scientific study, and probably never can. There is nothing even remotely radical or 'extraordinary' in stating this. The materialist core claim is, itself, extraordinarily bold and sweeping.

On materialism being a philosophical position and its core tenets as they relate to consciousness:

Article:

Materialism is a form of philosophical monism which holds that matter is the fundamental substance in nature, and that all things, including mental states and consciousness, are results of material interactions of material things. According to philosophical materialism, mind and consciousness are caused by physical processes, such as the neurochemistry of the human brain and nervous system, without which they cannot exist. Materialism directly contrasts with idealism, according to which consciousness is the fundamental substance of nature.

Materialism is closely related to physicalism—the view that all that exists is ultimately physical. Philosophical physicalism has evolved from materialism with the theories of the physical sciences to incorporate more sophisticated notions of physicality than mere ordinary matter (e.g. spacetime, physical energies and forces, and exotic matter). Thus, some prefer the term physicalism to materialism, while others use the terms as if they were synonymous.

Discoveries of neural correlates between consciousness and the brain are taken as empirical support for materialism, but some philosophers of mind find that association fallacious or consider it compatible with non-materialist ideas.[1]​[2]​ Alternative philosophies opposed or alternative to materialism or physicalism include idealism, pluralism, dualism, panpsychism, and other forms of monism.

Source: https://en.wikipedia.org/wiki/Materialism

Brain region expansion due to memory increase in terms of increased sensory 'knowledge' (your Maguire et al taxi driver citation) has no bearing on the present argument. If the brains would expand as a result of knowing more deeply, say about the universe, then I'd start to take you more seriously and regard highly abstract knowledge as also something clearly more physical in nature.

You didn't cite, nor will you be able to cite, any scientific proof for the higher functions of consciousness (say, our very ability to philosophize about whether physicalism is a scientific conclusion or a philosophical belief) being neurophysiological processes. As already repeatedly stated, we can only demonstrate some correlations between such mental states and brain states. But there's yet no proof that consciousness (our mental states as they are experienced) is material in any known sense of spatiotemporality and physical senses. Hence the so-called "hard problem of consciousness". Many materialist neuroscientists, philosophers and physicists, ranging from the likes of Sir Roger Penrose to Tim Maudlin and David Albert, readily accept that this is a real problem, and are therefore embarked on a mission to scientifically demonstrate said materialist reduction, or emergence, whilst acknowledging not having arrived just yet.

Even if we accept that one day such a materialistic reduction of mental states to brain states can be done, or a brain-emergent consciousness demonstrated, we can't conclude materialism as scientifically established before such a successful reduction. At least insofar as we are attemping to be scientific rather than faith-based dogmatists about materialism.

The Bayesian argument on priors applies very much to consciousness. Until we have no valid scientific reduction of mental states to brain-states, or higher cognitive capacites emerging from brain-states, the priors directly accessible to us (including our own experiences of philosophical ideas during this conversation) demonstrate unique states which, whatever they are, are certainly not material in any known way:

a. not being composed of any known physical components or systems, whether quantum-mechanical or neurophysiological that we know;
b. not being clearly locatable in spacetime;
c. not being sensed by known physical senses shared with all other mammals.

Call this line of reasoning as minimalist dualism if you like. There's no extra claim of any particular theory-heavy theology of 'souls' involved. Just accepting what's seemingly the case: A unique mental category of phenomena that's very common to us all and that we have immediate access to, in some ways more immediate than our known physical senses.

Furthermore, empirically verifiable human impact on the planet, whether for good or for ill, through science, technology, trade, industry, law, business, trasportation, engineering, construction, religion, powerful political systems, arts, mathematics and philosophy, glaringly (to every honest observer) singles us out from amongst all species despite minimal genetic differences with our closest mammalian relatives. All this astounding global impact by humans, and humans alone, by the virtue of all the foregoing disciplines and human enterprises, just happens to hinge upon this self-same consciousness being able to grasp fundamental abstract ideas in philosophy, arts, mathematics, ethics, sciences, psychology and society.

Yet the foregoing example of human impact on the planet through the exercise of our unique higher cognitive capacities is merely to demonstrate a state of play predicted by a dualistic theory, and that the claim of empirical evidence being only consistent with materialism is a self-serving category mistake. Not science. And yes, all logical philosophical positions, whether materialism or dualism or otherwise, are all unfalsifiable and untestable through empirical means. They are by default formulated on a higher level of abstraction and in such generic ways as to be consistent with all available evidence, or in all possible worlds. That's just the very nature of broad-stroke fundamental philosophical theories. That should go without saying.

But to state that (1) each our moral ideas of right and wrong, (2) value-based decision-making and our (3) comprehension of fundamental principles/ideas of various fields of knowledge ranging from philosophy to physics to arts represent a unique non-physical class of phenomena currently scientifically very unexplained by physical properties is nothing more than to accept at face value Frankena's point. That two seemingly different properties may also be, actually, different properties.

To recap:

Unless we're intellectually dishonest, we must acknowledge that the phenomenon of 'philosophical thought' (which you are experiencing even as we speak) is a completely different experience from a measurement of 'C-fiber firing in the Wernicke's area of the brain' which, however, may very well time-correlate with the former experience. That they're subjectively entirely different experiences is the case even if they were ultimately somehow the same thing based on some amazing future proof. But just because there is a connection, say a co-occurrence of experiencing a philosophical thought and a C-fiber firing in the Wernicke's area, such a fact does not logically imply a direct causality, and if indeed a causality could be established, this in turn doesn't logically imply identity of kind between two seemingly different phenomena. (By the way, supervenience -- causality between different orders of entities -- is a common and integral aspect of known reality.) This is the crux of the so-called 'hard problem'. If it were a trivial problem, you wouldn't have materialist and atheist physicists and philosophers of great ilk and credentials acknowledging that consciousness remains the greatest mystery in terms of physics. They are all honest to admit they are materialists in terms of their philosophical position, and that it is a kind of belief. Not because the science has shown consciousness to be clearly physical.

Even the least sophisticated superstition-laden religious formulations of a soul posit a constant causality and connection between the soul and the body/brain, and that any trauma or injury to the brain will impact the performance of the soul in any physical arena. And that the soul when connected to the body learns even non-physical knowledge primarily through the vessel of the body, for instance through parental and school instruction by means of language. In other words, the constant active correlation itself between higher mental states and brain states does nothing in the way of deciding who's more correct -- the idealist/dualist or the materialist.

By the way, Prof. Gary Marcus is also a materialist and studying his arguments for the abortive path of the current alleged progress towards AGI through LLMs and diffusion models are really worth a watch. Hence the video share in the above. He firmly believes AGI may be possible through other models. Materialist ones.

 

[John J.]

But to state that (1) each our moral ideas of right and wrong, (2) value-based decision-making and our (3) comprehension of fundamental principles/ideas of various fields of knowledge ranging from philosophy to physics to arts represent a unique non-physical class of phenomena currently scientifically very unexplained by physical properties is nothing more than to accept at face value Frankena's point.

People can discuss thinking and theoretical components of thinking without reference to specific physical structures, and much of cognitive psychology traditionally did so.
We can discuss many (most) things meaningfully without considering the underlying mechanisms that allow them to exist, e.g. the storyline of a book; it doesn't mean those things are a non-physical class of phenomena- their existence is dependent on physical processes.

Morality and values are human concepts. I do not accept that morality or values exist independently of human existence
(leaving aside specific religious entities, which I am not discussing). There is much evidence that morals/ ethics/ values are in large part culturally dependent, they are learned. Slaveowners did not consider themselves to be bad people, nor did most of their peers (including clergy, judiciary).

In terms of my understanding of scientific understanding, I don't accept there are any non-physical phenomena, unique or otherwise, so I don't accept Frankena's point.

You didn't cite, nor will you be able to cite, any scientific proof for the higher functions of consciousness (say, our very ability to philosophize about whether physicalism is a scientific conclusion or a philosophical belief) being neurophysiological processes.

No, but I can't cite any scientific proof for life on Earth originating from abiogenesis.
I can't cite any scientific proof against the assertion that "Every time I say I don't believe in fairies, a fairy dies".

We can't prove that self-awareness (as opposed to general awareness) is anything more than an emergent property, a self-applied label for when we are thinking about thinking, metacognition allowed by the interplay of memory, language and reasoning (problem solving). Maybe we are only truly self-aware when we are thinking about self-awareness...

There is overwhelming evidence that impeding or ending neurophysiological processes reduces or ends observable and/or measurable indications of higher functions of consciousness/ our ability to philosophize.
Assertions that these processes continue in a hidden, undetectable realm have no supporting evidence whatsoever
(in science or medicine).
Suppose I had a clock with an intricate mechanism that I couldn't understand. I jam my screwdriver in and wrench it about, and the hands on the clock face stop moving. I could decide that the clock, in a real but undetectable sense, continues to tell the time because the concept of telling the time is independent of the clock mechanism, and the mechanism does not describe it.
And I liked the clock when it was working, it was useful and I valued it highly. I don't like the idea that it doesn't work.

Many materialist neuroscientists, philosophers and physicists, ranging from the likes of Sir Roger Penrose to Tim Maudlin and David Albert, readily accept that this is a real problem

Sir Roger Penrose is a productive thinker, but even his supporters, who believe in a quantum underpinning of consciousness, accept it's a minority view. His suggested mechanisms (e.g. proposed quantum effects in microtubules) are broadly accepted as having been experimentally falsified.
Personally I'd rate the late, great Daniel Dennett as a more significant philosopher of mind than many others (certainly Penrose, though Penrose's main achievements are in other fields).

Prof. Gary Marcus is also a materialist and studying his arguments for the abortive path of the current alleged progress towards AGI through LLMs

I think the progress made by LLMs manipulating language is significant, and might have an impact in some areas of employment in the medium term (not in a good way). I suspect they will continue to be developed to an ever-greater refinement. They have pro's and con's.
But, as I've said before, I don't agree with the strong AI argument. I don't believe LLMs can be the core of "general AI".

 

[LilWabbit]

People can discuss thinking and theoretical components of thinking without reference to specific physical structures, and much of cognitive psychology traditionally did so.
We can discuss many (most) things meaningfully without considering the underlying mechanisms that allow them to exist, e.g. the storyline of a book; it doesn't mean those things are a non-physical class of phenomena- their existence is dependent on physical processes.

In the above you in fact unwittingly acknowledge Frankena's point. And even David Chalmers'. That you can discuss meaningfully about physical mechanisms without reference to psychology and meaningfully about psychological phenomena without reference to the physical. But since they are semantically different discussions independent of one another they cannot be claimed synonymous without proper proof or explanation or else one commits a fallacy of claiming the synonymity of two different kinds of semantic contents as a brute fact which none can question (i.e. the definist fallacy). That's dogmatism.

However, you can adopt a materialist worldview (as you have) and make the philosophical claim one is somehow based on the other. A popular claim made by many and which accepts an onus of proof -- yet thus far proven by none.

Morality and values are human concepts. I do not accept that morality or values exist independently of human existence
(leaving aside specific religious entities, which I am not discussing). There is much evidence that morals/ ethics/ values are in large part culturally dependent, they are learned. Slaveowners did not consider themselves to be bad people, nor did most of their peers (including clergy, judiciary).

I fail to see how the above addresses or refutes anything I mentioned before. These are statements of your position. I'm fine with that.

In terms of scientific understanding, I don't accept there are any non-physical phenomena, unique or otherwise

A perfectly fine statement of a philosophical belief. Not fact nor proof of said belief.

No, but I can't cite any scientific proof for life on Earth originating from abiogenesis.
I can't cite any scientific proof against the assertion that "Every time I say I don't believe in fairies, a fairy dies".

Indeed. But we're not discussing fairy tales but rather your comprehension of even the imagined notion of 'fairy tales'. And that comprehension is real and you've (as well as others) shown nothing that demonstrates such a comprehension is physical as per the a-c of physical properties commonly understood as physical properties I articulated earlier. Hence Chalmers (a materialist himself) and other philosophers of mind call this strange fact 'property dualism'.

We can't prove that self-awareness (as opposed to general awareness) is anything more than an emergent property,

Try to first prove it's even emergent. We're going in circles. Then we can discuss whether self-awareness of the profoundest kind (Cartesian, awareness and understanding of one's thoughts and being a thinking creature) lends itself to any scientific scrutiny.

There is overwhelming evidence that impeding or ending neurophysiological processes reduces or ends observable and/or measurable indications of higher functions of consciousness/ our ability to philosophize.

Yes. But as you carefully worded it, that's not any proof of consciousness disappearing altogether as a subjective experience. There are different competing theories to this ranging from materialist and dualist to idealist positions. None of which are testable. To assume conscious (subjective) experience disappears by health impairments or even death is assuming such experience is either entirely neurophysiological or existentially dependent on neurophysiology. My consciousness tends to become even wilder during health impairments. Haven't experienced death yet, thankfully. However, I already know the properties of these capacities (comprehension of moral ideas, cultural or not, value-based decision making and abstract comprehension of first principles) do not lend themselves to any classical physical scrutiny and hence casts reasonable doubt on such a position, whilst not disproving the rival claim. Hence the dilemma. As per the principle of supervenience, causation and correlation between these mental capacities and neurophysiology can be scientifically well-established and yet not logically implying in the least that these evidently different entities must necessarily, or even likely, be of the same kind or existentially dependent on each other one way or the other.

Assertions that these processes continue in a hidden, undetectable realm have no supporting evidence whatsoever
(in science or medicine).

Here's where we agree. Such assertions are just as untestable as is the assertion of subjective experience entirely disappearing at death. Hence we shouldn't concern science with such issues to begin with. Science should focus on testable claims.

Sir Roger Penrose is a productive thinker, but even his supporters, who believe in a quantum underpinning of consciousness, accept it's a minority view. His suggested mechanisms (e.g. proposed quantum effects in microtubules) are broadly accepted as having been experimentally falsified.

Agreed 100 %. But his assertion that consciousness is a great mystery to physics stands. Or as Tim Maudlin or David Chalmers would put it, we don't even have the mathematics or any other formal language to bridge the gap from objective description to subjective experience.

Hence the 'hard problem'.

Increased computing power and increasingly complex silicon substrate are hence strongly criticized as path to consciousness by them and Prof. Gary Marcus. Algorithms (of any known kind) and legacy computer technology even at its most refined just doesn't bridge that fundamental systemic gap.

 

[John J.]

My consciousness tends to become even wilder during health impairments.

I'm not aware of any major health impairment, let alone any impairment of the CNS, that heightens performance in any task of cognition. Your report may be subjectively correct, of course.
At risk of being a bit cheeky, maybe there's a reason no-one calls a dualist philosopher to assess brain stem death

We're clearly not going to agree, or convince each other (although I always find your posts interesting). That's one of the beauties of this site, we can have a good natured discussion (or argument) about arcane subjects and maybe others will read something they find of interest and will look into the subjects discussed further.

Increased computing power and increasingly complex silicon substrate are hence strongly criticized as path to consciousness... ...Algorithms (of any known kind) and legacy computer technology even at its most refined just doesn't bridge that fundamental systemic gap.

I probably agree, depending on how "consciousness" is defined. But most working in the field of machine intelligence are pursuing intelligent behaviours, not consciousness per se.

 

[LilWabbit]

I'm not aware of any major health impairment, let alone any impairment of the CNS, that heightens performance in any task of cognition. Your report may be subjectively correct, of course.

And right there's the crux: The seemingly unbridgeable dualism between measurable (objective) cognitive performance and subjective experience. They may correlate but yet they speak an entirely different language. The cognitive performance of a sick person or an old person may very well go down while their subjective experience remains rich -- arguably sometimes even richer.

But yes, let's agree to disagree that materialism is just as much an unproven belief as dualism. Most materialists deepened in the topic wouldn't disagree with their dualist duelling partners on this point though (corroborated even by the almighty Wikipedia article on materialism I cited -- the source of all incontrovertible truth ). Only on the actual tenets and claims of said beliefs.

At risk of being a bit cheeky, maybe there's a reason no-one calls a dualist philosopher to assess brain stem death

Touché! Though I truly hope no philosopher of any persuasion is ever asked to perform brain surgery on any hapless soul (pun intended). Nor a post-mortem lobotomy.

That's one of the beauties of this site, we can have a good natured discussion (or argument) about arcane subjects and maybe others will read something they find of interest and will look into the subjects discussed further.

Well said, bro. Couldn't agree more.

 

[FatPhil]

In the above you in fact unwittingly acknowledge Frankena's point. And even David Chalmers'. That you can discuss meaningfully about physical mechanisms without reference to psychology and meaningfully about psychological phenomena without reference to the physical. But since they are semantically different discussions independent of one another they cannot be claimed synonymous without proper proof or explanation or else one commits a fallacy of claiming the synonymity of two different kinds of semantic contents as a brute fact which none can question (i.e. the definist fallacy). That's dogmatism.

That comes over as if discussing two plates of food, a vegetarian chilli and a lamb korma, with two friends; one's a vegan, and we discuss their suitability in that regard, the other hates picant food, so we discuss their suitability in that regard - and you then butt in and tell me that I'm claiming veganism and spices are synonymous. Which on its own is absurd, but is particularly bizarre as the only claim that I'm making is that both of them are just food, nothing non-food in them, and as far as we know, all my food is just food.

 

[FatPhil]

At risk of being a bit cheeky, maybe there's a reason no-one calls a dualist philosopher to assess brain stem death

His own, or somebody else's?