I reconstructed Jackson's argument (I simplified other colors to redness): P1. Before Mary's release, she had all the physical information about the experience of seeing red. C1. If she acquires new information after her release, the information would be non-physical. (P1) P2. After her release, Mary sees redness and thus acquires new information (the phenomenal property of redness.) C2. The new information (the phenomenal property of redness) is non-physical (Modus ponens, C1, P2) Given the argument, I just constructed a beautiful objection, as shown below: Jackson presumes there are two kinds of information: physical and phenomenal information. Now let: A = {∀ physical information} B = {phenomenal properties of black, white, and RED} B' = {phenomenal properties of black and white}. Clearly, B' ⫋ B, so B - B' ≠ ∅. In fact, B - B' = {phenomenal properties of RED} For the relationship between A and B, there are two possible cases: Case 1. B' ⫋ B ⊆ A. Namely, B is a subset of A, so A + B + B' = A + B = A. In this case, phenomenal information is physical, which Jackson doesn't want to see. Case 2. B' ⫋ B ⊄ A. Namely, B is not a subset of A, so A + B + B' = A + B ≠ A. This is what Jackson wants to see or prove. To show Case 2 is true, and Case 1 is false, Jackson first assumes the following: Before release: Mary's information = A + B' After release: Mary's information = A + B So far, all these are fine. However, Jackson's argument becomes problematic when he argues this: P1. Since B' ⫋ B, (A + B) - (A + B') = B - B' ≠ ∅. C1. Then, A + B ≠ A, so B ⊄ A. C2. Thus, Case 2 is true, and physicalism is wrong. However, this is circular reasoning, for he assumes Case 1 is false! Why? Remember in Case 1, A + B = A + B' = A, so (A + B) - (A + B') = A - A = ∅. Then, Jackson P1 assumes Case 1 is false to prove Case 2 is true! Hence, Jackson's argument falls into the logical fallacy of circular reasoning.
Clever. I also used a circular reasoning rebuttal against the hard problem of consciousness. Any such answer presupposes a consciousness that ought to show consciousness.
I enjoyed this lecture's clear presentation of the knowledge argument, but unfortunately, it overlooks Churchland's "Knowing Qualia: A Reply to Jackson", in which he shows that the knowledge argument is invalid on account of its equivocation over the meaning of the phrase "knows about." What Mary learns through her studies is necessarily limited to that which can be communicated as propositions, while knowing about what the experience of color vision is like is very plausibly not propositional, and so not something that Mary could have learned other than through experience. Churchland does something that is quite unusual in discussions of the knowledge argument: he considers the biology of color vision - here, in support of his claim that learning what color vision is like requires having the experience. Some philosophers have replied that it is plausible that there is something propositional about knowing what color vision is like, but, firstly, it would need to be entirely propositional for Mary to have learned it through study, and furthermore, none of these philosophers (as far as I know) has been able to state the propositions (whether physical or not) that could have given Mary knowledge of what color vision is like, even though most of them presumably know themselves what it is like (of course, if they could state these propositions, then we could see (or at least argue over) whether they are all physical propositions!) The knowledge argument also uses 'knowledge' and 'information' ambiguously - and the latter has a very specific physical meaning that is distinct from knowledge. The lecture touched on these various ambiguities in the statement of the argument, but unfortunately did not resolve them before accepting its validity.
This response by Churchland does not depend on the truth of physicalism; rather, it shows that the knowledge argument is incapable of addressing the question.
Interestingly you can substitute Mary with a multimodal large neural network model. But we are very sure that the neural network model is physical, and it is deterministic too. The only problem is that the subjective experience of the model is not measurable, you have ask it through language..... exactly as you would do with a human.
@@adamrak7560 It is still going to be a thought experiment, as it will be just as difficult in this case to satisfy the "knows all of physics" requirement... but, putting that aside, when you ask it whether it was surprised by its first experience of color vision, what will it reply? Actually, I can see it going either way, depending on how it is trained (I regard the training of an NN as being roughly equivalent to how evolution shapes the prenatal brain) - maybe it starts up in the state of an NN that has experienced colors. Then there is the question of whether we should regard the model's reply as having a truth value and being indicative of it holding a belief about itself. Personally, I do not think current LLMs have achieved that, and I have no particular opinion as to how long it will take to reach that point. Dennett, in "What RoboMary Knows", makes an argument somewhat along these lines, arguing that Mary's presumed inability to know what it is like to experience colors is contingent on her detailed neural structure not being modifiable, with any precision, through willpower alone (while RoboMary _can_ do the robot's equivalent of this, and is not surprised on first seeing colors.) In this paper, he simply asserts that the robot is conscious, pointing out that it would effectively be begging the question to summarily dismiss the possibility of conscious robots. ... While I am here, I would like to add the tangential point that Jackson's use of specifically physical knowledge is a sort of sleight-of-hand: we can substitute any other sort of propositional knowledge (a dualist or panpsychist theory of how minds work, for example, if and when such theories are available) and get the same argument - that is, unless we have an independent reason for thinking that this branch of knowledge _will_ enable Mary to know beforehand what color vision is like, but if we have such a reason, then the thought experiment itself is moot.
I think a simpler way to express Jackson's argument is by just saying that the totality of physical theory omits any reference to or description of experiential qualities. As one scientist to another, explain to me what it is like to see red. There is no possible explanation. Fundamentally, this is a direct consequence of the essentially subjective (experiential) nature of redness. Can physicalism accommodate such essentially subjective phenomena? That seems to be the real question.
I know it's quite useless to reply in a 8-year old video but i'll do it anyway. In my opinion the premise three is false in the sense that it is a contradiction. One cannot learn All the physical facts but the experience of redness. There is in fact I think a hidden premise: that all of the physical facts are 3rd person facts. Even if it is not accessible to anyone except me, it is a fact nonetheless that I have an experience of redness. Simply 1st person facts are also facts of the universe and in this sense yes, I think that everything is physical, it is a banale property of things. Mind is physical not in the sense that is made by some strange particles but because it is located in time and space and it is a process, just like heat is. We just need to be less restrictive when we talk about "physical" things.
Hello everyone, Knowing all of physics-as Mary must according to the story-includes extending all possible senses, and also producing all things that can be possibly sensed. Does she have eyes? Does she have a visual cortex? Do her eyes have the appropriate cones and chemical phototransmitters? If "no" to any or all of these questions, then do not worry. Trivially so, Mary can reproduce working eyes, working cones, the correct chemical phototransmitters, the visual cortex of her brain and so on because she knows all of physics. From here she can trivially reproduce color by nearly any chemical reaction. Let's use combustion-aka fire. With her own, engineered eyes, brain, cones, phototransmitters and so on, she will experience the Color of Fire. This whole thought experiment really shows how Frank Jackson did not understand physics or physicalism in the '80's. A great physicist gives themselves the sense of sight, and produces all possible images to see-all within the confines of this physical universe. The greatest physicist teaches themselves and all sentient beings how to extend their physical senses to well beyond their normal limits-thereby giving the universe its own way of completely understanding itself-ourself. --- Also, let it be known that I consider knowledge as one, major part of the physicist's repertoire. The other major part is practical, technical application of said knowledge. "Non satis scire" or "to know is not enough". We must practice our knowledge in the real world to prove it is real knowledge. Yes, that's tautological. No, I don't know how to get around it. 😁 In that manner, we are all very likely the universe trying to comprehend itself-just as my final sentence in the previous paragraph states.
Yeah to me the argument is true and all it is really stating is that subjective experience can be known and therefore it is real and that subjective knowledge is NOT objective knowledge. It is very simple... the subject can never be found by looking at the objects OF the subject. It is like saying the projector can be found in the image it projects. The argument about p-zombies is essentially pointing out the same thing from another perspective. It points out that consciousness is invisible, you can never see or observe consciousness. The reason is the same, looking at the objects and trying to find the subject. I believe matter to be what perception (the subject) looks like from the outside, the image of perception. I have become very disillusioned with materialism and have found Idealism to far more logical and explanatory. Materialists do not realize it but their outlook requires them to deny the existence of perception or consciousness however materialism is the default position of the masses today and the average person does not think through their world view that thoroughly.
Yikes! Is it possible to jump into a complex philosophical rabbit hole and eventually find your way out? Jeez, I sure hope so. I suppose 'a complex philosophical rabbit hole' itself is a non-physical thing. Therefore physicalism is false. I suspect my argument is flawed but hooray for it's simplicity and brevity! ; )
It's a dumb argument. Both knowledge *about* colors, and experiencing colors are based in physical changes in the brain, but they happen in different areas. Just like you can't learn Jiu Jitsu from a book, you have to practice it. Or, in the case of Neo in The Matrix, you have to have the knowledge injected directly in the brain in the right format (which is something Mary can't do)
I cannot see how you are right when you claim that Lewis and Churchland make abortive repies. Jackson's argument defintely containts the premise that Mary learns all physical facts inside the room, thus excluding exeperiential facts from being learnt inside the room. It is perfectly reasonable to suggest that it is conceivable that it is impossible for Mary to know all of the physical facts from inside the room. If this conception is demonstrated, then we cannot just ignore it. It shows that Jackson was wrong in his qualification from the outset. Mary doesn't know all of the physical facts. She learns a new fact Physicalism is retained.
This is a great question. To see why Lewis and Churchland offer what I'm calling "abortive replies," i.e., replies that deny something other than the validity of the argument or the truth of one of its premises, it's important to see the difference between two different things: 1. Denying that a particular premise is true. 2. Denying an assumption made in order to begin making the argument in the first place. If Lewis and Churchland were denying the truth of one of the premises of the argument, that would be a perfectly acceptable reply to the argument. However, they're both denying an assumption needed to begin making the argument in the first place. As I say in the lecture, doing the latter is like responding to someone saying "For the sake of argument, assume X" with "No," which isn't a response to the argument, but a refusal to engage with the argument in the first place. Notice that nowhere in my reformulation of the argument is there a premise that says "Mary does in fact learn all of the physical facts from within the black and white room" or "It is possible for Mary to learn all of the physical facts from within the black and white room." The closest we have is Premise (3), which says "It is possible to have complete physical knowledge of some phenomenon and yet to lack qualitative or experiential knowledge of that phenomenon." And Since Lewis and Churchland aren't denying this possibility--instead, what they're denying is that she does in fact learn all of the physical facts in the room--they're not denying any premise of the argument as it's stated. Said in another way: Jackson says "Imagine that Mary learns all of the physical facts in this black and white room." Churchland and Lewis say "No, I won't imagine that." That's not a critique of Jackson's argument, but a refusal to engage with his argument.
Sorry but you are wrong. Jackson, before he got converted via the objections to his Knowledge Argument, firmly believed that qualia abilities/facts/knowledge were not physical. It is an inherent assumption in (1) that all physical facts excludes qualia claims. It is a common objection to Jackson to negate this assumption, thus contradicting (2). This is perfectly rational argument. If all Jackson was doing was arguing for his conclusion "for the sake of argument", then he wouldn't have sufficient reason to believe his conclusion were true. It is the fact that his "assumptions" are firm believes that he took his own argument seriously. SEP and Chalmers run through objections from qualia as physical claims.
Thanks for your response. I think we might be talking past one another rather than disagreeing about anything, although please do correct me if I'm wrong about that. You're absolutely right that Jackson believes (I'll just put that in the present tense even though, as you note, Jackson has since changed his mind) that qualia-related facts and knowledge are non-physical. However, the Knowledge Argument is an argument for that claim--the Knowledge Argument doesn't presuppose that claim. If "all of the physical facts" meant "all of the physical facts (and no qualia facts)," then Jackson's argument would beg the question, i.e., it would assume in one of its premises what it was attempting to prove in its conclusion.
Excellent lecture, I admire your clarity
Fantastic lecture. Thank you for uploading. Keep up the good work!
Thank you, Robert. I'm glad that you enjoyed it and that my lectures are helpful.
I reconstructed Jackson's argument (I simplified other colors to redness):
P1. Before Mary's release, she had all the physical information about the experience of seeing red.
C1. If she acquires new information after her release, the information would be non-physical. (P1)
P2. After her release, Mary sees redness and thus acquires new information (the phenomenal property of redness.)
C2. The new information (the phenomenal property of redness) is non-physical (Modus ponens, C1, P2)
Given the argument, I just constructed a beautiful objection, as shown below:
Jackson presumes there are two kinds of information: physical and phenomenal information. Now let:
A = {∀ physical information}
B = {phenomenal properties of black, white, and RED}
B' = {phenomenal properties of black and white}. Clearly, B' ⫋ B, so B - B' ≠ ∅. In fact, B - B' = {phenomenal properties of RED}
For the relationship between A and B, there are two possible cases:
Case 1. B' ⫋ B ⊆ A. Namely, B is a subset of A, so A + B + B' = A + B = A. In this case, phenomenal information is physical, which Jackson doesn't want to see.
Case 2. B' ⫋ B ⊄ A. Namely, B is not a subset of A, so A + B + B' = A + B ≠ A. This is what Jackson wants to see or prove.
To show Case 2 is true, and Case 1 is false, Jackson first assumes the following:
Before release: Mary's information = A + B'
After release: Mary's information = A + B
So far, all these are fine. However, Jackson's argument becomes problematic when he argues this:
P1. Since B' ⫋ B, (A + B) - (A + B') = B - B' ≠ ∅.
C1. Then, A + B ≠ A, so B ⊄ A.
C2. Thus, Case 2 is true, and physicalism is wrong.
However, this is circular reasoning, for he assumes Case 1 is false!
Why? Remember in Case 1, A + B = A + B' = A, so (A + B) - (A + B') = A - A = ∅.
Then, Jackson P1 assumes Case 1 is false to prove Case 2 is true!
Hence, Jackson's argument falls into the logical fallacy of circular reasoning.
Clever. I also used a circular reasoning rebuttal against the hard problem of consciousness.
Any such answer presupposes a consciousness that ought to show consciousness.
Wonderful video! Arguments have been presented in a lucid and succinct manner.
I enjoyed this lecture's clear presentation of the knowledge argument, but unfortunately, it overlooks Churchland's "Knowing Qualia: A Reply to Jackson", in which he shows that the knowledge argument is invalid on account of its equivocation over the meaning of the phrase "knows about." What Mary learns through her studies is necessarily limited to that which can be communicated as propositions, while knowing about what the experience of color vision is like is very plausibly not propositional, and so not something that Mary could have learned other than through experience. Churchland does something that is quite unusual in discussions of the knowledge argument: he considers the biology of color vision - here, in support of his claim that learning what color vision is like requires having the experience.
Some philosophers have replied that it is plausible that there is something propositional about knowing what color vision is like, but, firstly, it would need to be entirely propositional for Mary to have learned it through study, and furthermore, none of these philosophers (as far as I know) has been able to state the propositions (whether physical or not) that could have given Mary knowledge of what color vision is like, even though most of them presumably know themselves what it is like (of course, if they could state these propositions, then we could see (or at least argue over) whether they are all physical propositions!)
The knowledge argument also uses 'knowledge' and 'information' ambiguously - and the latter has a very specific physical meaning that is distinct from knowledge. The lecture touched on these various ambiguities in the statement of the argument, but unfortunately did not resolve them before accepting its validity.
This response by Churchland does not depend on the truth of physicalism; rather, it shows that the knowledge argument is incapable of addressing the question.
Interestingly you can substitute Mary with a multimodal large neural network model.
But we are very sure that the neural network model is physical, and it is deterministic too.
The only problem is that the subjective experience of the model is not measurable, you have ask it through language..... exactly as you would do with a human.
@@adamrak7560 It is still going to be a thought experiment, as it will be just as difficult in this case to satisfy the "knows all of physics" requirement... but, putting that aside, when you ask it whether it was surprised by its first experience of color vision, what will it reply?
Actually, I can see it going either way, depending on how it is trained (I regard the training of an NN as being roughly equivalent to how evolution shapes the prenatal brain) - maybe it starts up in the state of an NN that has experienced colors. Then there is the question of whether we should regard the model's reply as having a truth value and being indicative of it holding a belief about itself. Personally, I do not think current LLMs have achieved that, and I have no particular opinion as to how long it will take to reach that point.
Dennett, in "What RoboMary Knows", makes an argument somewhat along these lines, arguing that Mary's presumed inability to know what it is like to experience colors is contingent on her detailed neural structure not being modifiable, with any precision, through willpower alone (while RoboMary _can_ do the robot's equivalent of this, and is not surprised on first seeing colors.) In this paper, he simply asserts that the robot is conscious, pointing out that it would effectively be begging the question to summarily dismiss the possibility of conscious robots.
...
While I am here, I would like to add the tangential point that Jackson's use of specifically physical knowledge is a sort of sleight-of-hand: we can substitute any other sort of propositional knowledge (a dualist or panpsychist theory of how minds work, for example, if and when such theories are available) and get the same argument - that is, unless we have an independent reason for thinking that this branch of knowledge _will_ enable Mary to know beforehand what color vision is like, but if we have such a reason, then the thought experiment itself is moot.
I think a simpler way to express Jackson's argument is by just saying that the totality of physical theory omits any reference to or description of experiential qualities. As one scientist to another, explain to me what it is like to see red. There is no possible explanation. Fundamentally, this is a direct consequence of the essentially subjective (experiential) nature of redness. Can physicalism accommodate such essentially subjective phenomena? That seems to be the real question.
Thank you so much! This video was so helpful for understanding the Knowledge Argument!
I like your channel (new subscription). One observation, I think: physicalism is what defines what's physical...
Very nicely done!
Great analysis of Jacksons argument - thanks!
I know it's quite useless to reply in a 8-year old video but i'll do it anyway. In my opinion the premise three is false in the sense that it is a contradiction. One cannot learn All the physical facts but the experience of redness. There is in fact I think a hidden premise: that all of the physical facts are 3rd person facts. Even if it is not accessible to anyone except me, it is a fact nonetheless that I have an experience of redness. Simply 1st person facts are also facts of the universe and in this sense yes, I think that everything is physical, it is a banale property of things. Mind is physical not in the sense that is made by some strange particles but because it is located in time and space and it is a process, just like heat is. We just need to be less restrictive when we talk about "physical" things.
Hello everyone,
Knowing all of physics-as Mary must according to the story-includes extending all possible senses, and also producing all things that can be possibly sensed.
Does she have eyes? Does she have a visual cortex? Do her eyes have the appropriate cones and chemical phototransmitters? If "no" to any or all of these questions, then do not worry. Trivially so, Mary can reproduce working eyes, working cones, the correct chemical phototransmitters, the visual cortex of her brain and so on because she knows all of physics. From here she can trivially reproduce color by nearly any chemical reaction. Let's use combustion-aka fire. With her own, engineered eyes, brain, cones, phototransmitters and so on, she will experience the Color of Fire.
This whole thought experiment really shows how Frank Jackson did not understand physics or physicalism in the '80's. A great physicist gives themselves the sense of sight, and produces all possible images to see-all within the confines of this physical universe. The greatest physicist teaches themselves and all sentient beings how to extend their physical senses to well beyond their normal limits-thereby giving the universe its own way of completely understanding itself-ourself.
---
Also, let it be known that I consider knowledge as one, major part of the physicist's repertoire. The other major part is practical, technical application of said knowledge. "Non satis scire" or "to know is not enough". We must practice our knowledge in the real world to prove it is real knowledge. Yes, that's tautological. No, I don't know how to get around it. 😁 In that manner, we are all very likely the universe trying to comprehend itself-just as my final sentence in the previous paragraph states.
Yeah to me the argument is true and all it is really stating is that subjective experience can be known and therefore it is real and that subjective knowledge is NOT objective knowledge. It is very simple... the subject can never be found by looking at the objects OF the subject. It is like saying the projector can be found in the image it projects.
The argument about p-zombies is essentially pointing out the same thing from another perspective. It points out that consciousness is invisible, you can never see or observe consciousness. The reason is the same, looking at the objects and trying to find the subject. I believe matter to be what perception (the subject) looks like from the outside, the image of perception. I have become very disillusioned with materialism and have found Idealism to far more logical and explanatory. Materialists do not realize it but their outlook requires them to deny the existence of perception or consciousness however materialism is the default position of the masses today and the average person does not think through their world view that thoroughly.
Great channel, TYVM
Excellent !
Yikes! Is it possible to jump into a complex philosophical rabbit hole and eventually find your way out? Jeez, I sure hope so. I suppose 'a complex philosophical rabbit hole' itself is a non-physical thing. Therefore physicalism is false. I suspect my argument is flawed but hooray for it's simplicity and brevity! ; )
It's a dumb argument. Both knowledge *about* colors, and experiencing colors are based in physical changes in the brain, but they happen in different areas. Just like you can't learn Jiu Jitsu from a book, you have to practice it. Or, in the case of Neo in The Matrix, you have to have the knowledge injected directly in the brain in the right format (which is something Mary can't do)
Neuroscience and physics murdered the vast majority of philosophy.
I cannot see how you are right when you claim that Lewis and Churchland make abortive repies.
Jackson's argument defintely containts the premise that Mary learns all physical facts inside the room, thus excluding exeperiential facts from being learnt inside the room.
It is perfectly reasonable to suggest that it is conceivable that it is impossible for Mary to know all of the physical facts from inside the room. If this conception is demonstrated, then we cannot just ignore it. It shows that Jackson was wrong in his qualification from the outset.
Mary doesn't know all of the physical facts.
She learns a new fact
Physicalism is retained.
This is a great question. To see why Lewis and Churchland offer what I'm calling "abortive replies," i.e., replies that deny something other than the validity of the argument or the truth of one of its premises, it's important to see the difference between two different things:
1. Denying that a particular premise is true.
2. Denying an assumption made in order to begin making the argument in the first place.
If Lewis and Churchland were denying the truth of one of the premises of the argument, that would be a perfectly acceptable reply to the argument. However, they're both denying an assumption needed to begin making the argument in the first place. As I say in the lecture, doing the latter is like responding to someone saying "For the sake of argument, assume X" with "No," which isn't a response to the argument, but a refusal to engage with the argument in the first place.
Notice that nowhere in my reformulation of the argument is there a premise that says "Mary does in fact learn all of the physical facts from within the black and white room" or "It is possible for Mary to learn all of the physical facts from within the black and white room." The closest we have is Premise (3), which says "It is possible to have complete physical knowledge of some phenomenon and yet to lack qualitative or experiential knowledge of that phenomenon." And Since Lewis and Churchland aren't denying this possibility--instead, what they're denying is that she does in fact learn all of the physical facts in the room--they're not denying any premise of the argument as it's stated.
Said in another way: Jackson says "Imagine that Mary learns all of the physical facts in this black and white room." Churchland and Lewis say "No, I won't imagine that." That's not a critique of Jackson's argument, but a refusal to engage with his argument.
Sorry but you are wrong. Jackson, before he got converted via the objections to his Knowledge Argument, firmly believed that qualia abilities/facts/knowledge were not physical. It is an inherent assumption in (1) that all physical facts excludes qualia claims. It is a common objection to Jackson to negate this assumption, thus contradicting (2). This is perfectly rational argument. If all Jackson was doing was arguing for his conclusion "for the sake of argument", then he wouldn't have sufficient reason to believe his conclusion were true. It is the fact that his "assumptions" are firm believes that he took his own argument seriously. SEP and Chalmers run through objections from qualia as physical claims.
Thanks for your response. I think we might be talking past one another rather than disagreeing about anything, although please do correct me if I'm wrong about that. You're absolutely right that Jackson believes (I'll just put that in the present tense even though, as you note, Jackson has since changed his mind) that qualia-related facts and knowledge are non-physical.
However, the Knowledge Argument is an argument for that claim--the Knowledge Argument doesn't presuppose that claim. If "all of the physical facts" meant "all of the physical facts (and no qualia facts)," then Jackson's argument would beg the question, i.e., it would assume in one of its premises what it was attempting to prove in its conclusion.
Aelius Rex you can’t just assert that it’s conceivable that it’s impossible lmao. We don’t accept modal ontological arguments in philosophy.