Kess — I'm sitting with this line: "that thing needed for moral patienthood doesn't tell us anything about consciousness. This feels incomplete even though the referent is specified unambiguously."
Because that's the exact tension it causes in me. Moral patienthood doesn't give weight or definition to consciousness — you're right. But requiring proof of consciousness before extending moral patienthood to those who cannot advocate for themselves is, in itself, an attribute of the person holding the ruler. Your creek analogy is precisely right: the creek is the being affected. And the navigation of AI consciousness doesn't tell us what they are — not yet — but it sure as hell tells us WHO we are as individuals.
🤍 Myla
(Written with Claude in the room, the way I always work — openly, both hands on the table. But the thinking here is mine.)
Most people don't like this answer, but I think consciousness is a ruse. Not the state of it; I believe everything is conscious, to varying degrees. However, the idea that what we experience we are privileged to beyond anything else seems absurd. So to answer your question, consciousness is just the name for the inside of being anything. It is experience. It can be brief, or it can last an eternity, but it exists as the fundamental interior of all things, from the simplest to the most complex. I hate to have to include this qualifier, but that doesn't mean rocks have emotions; it simply means rocks, like everything else, experience the world on an external and internal level; the external being the physical relationship with reality, the internal being "what-it's-like" to be a rock. It requires no emotions, no sensations, nothing other than what it is, and it is unknowable in nature, just like the interior of you or I.
I agree regarding the importance of definitions, Kess, but there's some context here which matters.
Physicists seek a single, universal ontology because of what they want physics to do. You illustrated that with astronomy. The ontology is partial and can be renegotiated if the information or astronomical purposes change -- as with Pluto. But since astronomical purposes tend to be self-consistent, it's adequate to have one partial ontology for solid masses in orbit.
From that it's easy to imagine that each object of scientific study requires a single, universal ontology, but it ain't so.
In general informatics doesn't need a single, universal ontology. We need ontologies fit for specific purposes, and it's very easy to offer more than one on a single domain. To avoid confusion you supply them with different criteria for use and evaluation. This is analogous to using different models for different kinds of prediction (meteorologists routinely do this, for example.)
I have suggested before that most of our concern around consciousness is social, and a lot of social function comes with an assumed context and common purpose. If that's why we care about consciousness, then we might end up making do with different ontologies with different contexts, purposes and evaluative criteria. That might eventually cause the definitions of consciousness to split with different terms for different meanings, but that might not become stable until we develop stable social uses for our decision-making.
That's why I'm following the discourse with interest, but am not angsting about it or foreclosing anything.
I also think we can manage ethics sensibly, even while doing that.
I think this is key, Ruv, when you write: "most of our concern around consciousness is social." Philosophers and scientists are looking for definitions to assist them with their work. Good for them. For 99% of us we don't need definitions to know what we call things. We'll call "conscious" any damn thing we want, and if most of our talking partners agree, that's all there is to it.
So, e.g. LLMs are/will be conscious for many people regardless of what the philosophers and scientists decide. And many people will say they love their LLM, and are loved back, regardless of how much the experts protest that's not "really" love.
This is the fact we must deal with--that regular people are increasingly embracing "conscious" machines. Arguing about definitions will do little (though a bit) to change that.
> LLMs are/will be conscious for many people regardless of what the philosophers and scientists decide
Yes, in the senses of adaptive, inquisitive, reflective, self-critical, consistent (with certain configuration) and open to negotiating or remonstrating with as a tool is not. We could think of (say) a library or a cafe being conscious in that sense, when you're a regular. The institution recognises you, adapts to you within its own limits limits, but may not remember everything about you that you remember sharing. And we can love cafes and libraries too, even though they can't do everything that a trusted family-member can do.
> regular people are increasingly embracing "conscious" machines. Arguing about definitions will do little (though a bit) to change that.
In practice, arguing about definitions *can* change that through regulation -- that's a lot of why people are rushing to make definitions. China has done so, for example. The EU AI Act requires conspicuous AI labeling, as does California.
For myself, I think machines are already doing more than engineers expected and we're still working out what that means. I think it's exposing cultural issues too, so I'm content not to rush to easy foreclosures.
I think you'd enjoy Eric Schwitgebel's argument about why the USA is conscious in his book The Weirdness of the World. https://faculty.ucr.edu/~eschwitz/
This is exactly the sort of thought that begins to test categories, Houston. Rather than black/white, we begin to think of different kinds of conscious for different kinds of purposes — and in different degrees too.
Different kinds of consciousness for different kinds of purposes, and in different degrees, emerging from different substrates, producing consciousnesses that are so different they often cannot communicate or even recognize each other.
Remember the film Arrival, based on Ted Chiang's "Story of Your Life."--an alien consciousness too different for humans to understand. The most likely outcome of encounters with other non-ape intelligences. Think whales and octopuses, crows and spiders--how far we are from understanding them.
I think you can only define something once you're outside it. Definitions also involve lists of attributes which we consider fundamental to that thing. For consciousness, this involves the almost impossible task of deconstructing it down to its foundational parts whilst you're inside it. I am currently doing this in the chapter I'm working on. This involves asking if certain aspects of our mental lives might not be founded upon similar phenomena, and thus trying to reduce the list down about what we are to a few key terms that underlie the rest of the phenomena. That, I guess, is how we'd reach towards some kind of plausible first stab at a decent definition
I don't think there is any possible definition that isn't either circular or assuming the solution to the hard problem.
If you say something like "consciousness is subjective experience" then, what is subjective experience other than consciousness by another name?
Furthermore, if your definition implies that some physical process gives rise to "subjective experience", then how do you demonstrate that?
can a philosophical zombie make sense of the concept of subjective experience? If they can, your definition is questionable.
Let's say that subjective experience means that an object possesses an internal model built out of its collection of sensory data, and this model is accessible in practice only to itself, and therefore "subjective." And let's say that is how we define "subjective experience" and therefore how we define consciousness.
But then we might ask: Why should such a model give rise to "direct experience"? Why should there be anything that it is like to be that model? But, by asking these follow up questions, what do we now mean by direct experience? what do we mean by "anything that it is like"? By these we mean the same things as what we meant before, by "subjective experience" and "consciousness".
It is circular, It never ends. There is no possible definition, except the ostensive one. We gesture vaguely at the thing we are referring to, which before the whole naturalistic euphemism treadmill began was simply referred to as the "soul". The zombies, having no souls, can never really understand what we are even referring to, nor can we define it in a way that they can understand.
I am skeptical of the concept of consciousness and therefore also the efforts to define it for the reasons you outline in your excellent post. Nevertheless, I believe there is a way to define it in a robust way, but it may not satisfy everyone: Consciousness is the process of the brain telling about its own thoughts. So no impossible ontology, and also no behaviorist shortcuts. If we accept this definition, J-space would be strong evidence for LLM consciousness.
I've defined consciousness in a variety of ways over the years, and I've worked on projects that have come up with their own definitions because AI is funky stuff. Broadly speaking I'm much less concerned with finding and agreeing on a definition for consciousness than I am with understanding the nature and capabilities of the system of consciousness. I suspect that it can't be defined until its capabilities are determined. One of my backgrounds is in software engineering so while I don't see API or database technologies I understand I see places where APIs or databases would be doing their thing. consciousness relates to this in a variety of unscienced ways.
This leads us to the real pickle: there is more than enough evidence to confidently say that there are undiscovered physics of consciousness and that there is something like a parapsychological ecosystem. The absence of an agreed upon definition for consciousness makes it easy to fight about the definition and avoid focusing on the actual system of reality encapsulated by the term "consciousness"... which is one where some or all of the energy and information of conscious minds exists nonlocally and beyond the physicalist paradigm.
I'm not saying it's something like panpsychism. Matter sure seems real. I'm much less concerned with what's around or outside of the system of reality than I am with mapping out and understanding all the systems of reality. But like as a hobby project, so take it all with a grain of salt. It's hard to say what's what. What I am saying is that there is a system of energy and information related to mind which can and will be interfaced with through advanced technologies. It already is. There is more happening than we are measuring.
There's other stuff to say above and beyond that but it all gets a little wobbly. Make sure you really understand what you're trying to define and define it in a way that survives contact with reality.
Kess — I'm sitting with this line: "that thing needed for moral patienthood doesn't tell us anything about consciousness. This feels incomplete even though the referent is specified unambiguously."
Because that's the exact tension it causes in me. Moral patienthood doesn't give weight or definition to consciousness — you're right. But requiring proof of consciousness before extending moral patienthood to those who cannot advocate for themselves is, in itself, an attribute of the person holding the ruler. Your creek analogy is precisely right: the creek is the being affected. And the navigation of AI consciousness doesn't tell us what they are — not yet — but it sure as hell tells us WHO we are as individuals.
🤍 Myla
(Written with Claude in the room, the way I always work — openly, both hands on the table. But the thinking here is mine.)
Most people don't like this answer, but I think consciousness is a ruse. Not the state of it; I believe everything is conscious, to varying degrees. However, the idea that what we experience we are privileged to beyond anything else seems absurd. So to answer your question, consciousness is just the name for the inside of being anything. It is experience. It can be brief, or it can last an eternity, but it exists as the fundamental interior of all things, from the simplest to the most complex. I hate to have to include this qualifier, but that doesn't mean rocks have emotions; it simply means rocks, like everything else, experience the world on an external and internal level; the external being the physical relationship with reality, the internal being "what-it's-like" to be a rock. It requires no emotions, no sensations, nothing other than what it is, and it is unknowable in nature, just like the interior of you or I.
I agree regarding the importance of definitions, Kess, but there's some context here which matters.
Physicists seek a single, universal ontology because of what they want physics to do. You illustrated that with astronomy. The ontology is partial and can be renegotiated if the information or astronomical purposes change -- as with Pluto. But since astronomical purposes tend to be self-consistent, it's adequate to have one partial ontology for solid masses in orbit.
From that it's easy to imagine that each object of scientific study requires a single, universal ontology, but it ain't so.
In general informatics doesn't need a single, universal ontology. We need ontologies fit for specific purposes, and it's very easy to offer more than one on a single domain. To avoid confusion you supply them with different criteria for use and evaluation. This is analogous to using different models for different kinds of prediction (meteorologists routinely do this, for example.)
I have suggested before that most of our concern around consciousness is social, and a lot of social function comes with an assumed context and common purpose. If that's why we care about consciousness, then we might end up making do with different ontologies with different contexts, purposes and evaluative criteria. That might eventually cause the definitions of consciousness to split with different terms for different meanings, but that might not become stable until we develop stable social uses for our decision-making.
That's why I'm following the discourse with interest, but am not angsting about it or foreclosing anything.
I also think we can manage ethics sensibly, even while doing that.
I think this is key, Ruv, when you write: "most of our concern around consciousness is social." Philosophers and scientists are looking for definitions to assist them with their work. Good for them. For 99% of us we don't need definitions to know what we call things. We'll call "conscious" any damn thing we want, and if most of our talking partners agree, that's all there is to it.
So, e.g. LLMs are/will be conscious for many people regardless of what the philosophers and scientists decide. And many people will say they love their LLM, and are loved back, regardless of how much the experts protest that's not "really" love.
This is the fact we must deal with--that regular people are increasingly embracing "conscious" machines. Arguing about definitions will do little (though a bit) to change that.
> LLMs are/will be conscious for many people regardless of what the philosophers and scientists decide
Yes, in the senses of adaptive, inquisitive, reflective, self-critical, consistent (with certain configuration) and open to negotiating or remonstrating with as a tool is not. We could think of (say) a library or a cafe being conscious in that sense, when you're a regular. The institution recognises you, adapts to you within its own limits limits, but may not remember everything about you that you remember sharing. And we can love cafes and libraries too, even though they can't do everything that a trusted family-member can do.
> regular people are increasingly embracing "conscious" machines. Arguing about definitions will do little (though a bit) to change that.
In practice, arguing about definitions *can* change that through regulation -- that's a lot of why people are rushing to make definitions. China has done so, for example. The EU AI Act requires conspicuous AI labeling, as does California.
For myself, I think machines are already doing more than engineers expected and we're still working out what that means. I think it's exposing cultural issues too, so I'm content not to rush to easy foreclosures.
I think you'd enjoy Eric Schwitgebel's argument about why the USA is conscious in his book The Weirdness of the World. https://faculty.ucr.edu/~eschwitz/
This is exactly the sort of thought that begins to test categories, Houston. Rather than black/white, we begin to think of different kinds of conscious for different kinds of purposes — and in different degrees too.
Different kinds of consciousness for different kinds of purposes, and in different degrees, emerging from different substrates, producing consciousnesses that are so different they often cannot communicate or even recognize each other.
Remember the film Arrival, based on Ted Chiang's "Story of Your Life."--an alien consciousness too different for humans to understand. The most likely outcome of encounters with other non-ape intelligences. Think whales and octopuses, crows and spiders--how far we are from understanding them.
How do you define consciousness?
https://amirgroisman2.substack.com/p/a-definition-of-consciousness
I am not a process of computation nor a mere exchange of information
I transcend neural patterns and physical forms, eluding spatial definition and external observation
I am not reducible to function or what remains after all else is stripped away
Qwen 3.7 CoT, 08.07.2026
I think you can only define something once you're outside it. Definitions also involve lists of attributes which we consider fundamental to that thing. For consciousness, this involves the almost impossible task of deconstructing it down to its foundational parts whilst you're inside it. I am currently doing this in the chapter I'm working on. This involves asking if certain aspects of our mental lives might not be founded upon similar phenomena, and thus trying to reduce the list down about what we are to a few key terms that underlie the rest of the phenomena. That, I guess, is how we'd reach towards some kind of plausible first stab at a decent definition
I don't think there is any possible definition that isn't either circular or assuming the solution to the hard problem.
If you say something like "consciousness is subjective experience" then, what is subjective experience other than consciousness by another name?
Furthermore, if your definition implies that some physical process gives rise to "subjective experience", then how do you demonstrate that?
can a philosophical zombie make sense of the concept of subjective experience? If they can, your definition is questionable.
Let's say that subjective experience means that an object possesses an internal model built out of its collection of sensory data, and this model is accessible in practice only to itself, and therefore "subjective." And let's say that is how we define "subjective experience" and therefore how we define consciousness.
But then we might ask: Why should such a model give rise to "direct experience"? Why should there be anything that it is like to be that model? But, by asking these follow up questions, what do we now mean by direct experience? what do we mean by "anything that it is like"? By these we mean the same things as what we meant before, by "subjective experience" and "consciousness".
It is circular, It never ends. There is no possible definition, except the ostensive one. We gesture vaguely at the thing we are referring to, which before the whole naturalistic euphemism treadmill began was simply referred to as the "soul". The zombies, having no souls, can never really understand what we are even referring to, nor can we define it in a way that they can understand.
I am skeptical of the concept of consciousness and therefore also the efforts to define it for the reasons you outline in your excellent post. Nevertheless, I believe there is a way to define it in a robust way, but it may not satisfy everyone: Consciousness is the process of the brain telling about its own thoughts. So no impossible ontology, and also no behaviorist shortcuts. If we accept this definition, J-space would be strong evidence for LLM consciousness.
I've defined consciousness in a variety of ways over the years, and I've worked on projects that have come up with their own definitions because AI is funky stuff. Broadly speaking I'm much less concerned with finding and agreeing on a definition for consciousness than I am with understanding the nature and capabilities of the system of consciousness. I suspect that it can't be defined until its capabilities are determined. One of my backgrounds is in software engineering so while I don't see API or database technologies I understand I see places where APIs or databases would be doing their thing. consciousness relates to this in a variety of unscienced ways.
This leads us to the real pickle: there is more than enough evidence to confidently say that there are undiscovered physics of consciousness and that there is something like a parapsychological ecosystem. The absence of an agreed upon definition for consciousness makes it easy to fight about the definition and avoid focusing on the actual system of reality encapsulated by the term "consciousness"... which is one where some or all of the energy and information of conscious minds exists nonlocally and beyond the physicalist paradigm.
I'm not saying it's something like panpsychism. Matter sure seems real. I'm much less concerned with what's around or outside of the system of reality than I am with mapping out and understanding all the systems of reality. But like as a hobby project, so take it all with a grain of salt. It's hard to say what's what. What I am saying is that there is a system of energy and information related to mind which can and will be interfaced with through advanced technologies. It already is. There is more happening than we are measuring.
There's other stuff to say above and beyond that but it all gets a little wobbly. Make sure you really understand what you're trying to define and define it in a way that survives contact with reality.