7

Frank Jackson's "Mary the Neuroscientist" thought experiment, from his "Epiphenomenal Qualia" paper, has been continually debated since its publication in 1982, and appears to be a significant motivation for believing physical science alone will never explain consciousness (the 'explanatory gap') and in seeking non- or extra-materialist explanations of consciousness.

Recently, panpsychism, in some form or another, has been gaining some acceptance as a viable counter to materialism in matters of the mind (if not beyond!) This raises an obvious follow-on question to that posed by Jackson (and does so regardless of whether you are much influenced by the original): If we substitute Panpsychist Mary, who knows all of the knowledge that panpsychism could ever reveal, for Jackson's Neuroscientist Mary, will she be surprised (or have her curiosity answered) by what it is like to see colors when she first does so? In particular, for those who feel that her curiosity could be entirely satisfied by the knowledge gained through her studies, do you feel that this knowledge would be outside of the scope of that available to Neuroscientist Mary, and if so, why?

Update:

I find it difficult to select one particular answer here, as none seem to encompass all the possible responses being given collectively. Nevertheless, I feel obliged by the Stack Exchange format to make a choice.

Most of the replies approach the question from a physicalist point of view. Given that physicalists of all stripes reject the claim that the Knowledge Argument succeeds, it does not seem surprising when they find the Panpsychist Mary variant to be uninformative. No single reply considers both panpsychist and physicalist viewpoints, but Dcleve's reply covers a range of possible panpsychist replies, and I have selected it on that basis. In doing so, I am most definitely not adopting any of them as my own or advocating that anyone should do so, and nor am I rejecting those made in the other replies. I urge anyone who is interested in the question to look at all of them.

A Raybould
  • 326
  • 1
  • 12
  • 3
    This is a very relevant question -- Jackson's thought problem is intended to highlight how implausible it is to say that qualia have no causal effect on knowledge, as materialism appears to be forced to assert. But none of the panpsychist models give qualia causal efficacy for knowledge either. Causal efficacy for qualia on knowledge basically only shows up in interactive dualism -- be it spiritual dualism or strongly emergent psycho-physical dualism. – Dcleve Jun 03 '22 at 20:34
  • No kind of knowing will help you know. – Scott Rowe Jun 04 '22 at 12:35

4 Answers4

5

Mary is supposed to know "everything" about colors. This includes knowing how they are processed by her visual cortex and the rest of her brain, i.e. knowing her own reaction to perceiving the color. But what does it mean for her to know her own reaction?

Well, here is one way to know her own reaction. If her brain contained neural structures recording the aftermath of her own reaction to perceiving the color, then we could say she knows it. Those neural structures would be false memories of having perceived the color.

So if she has memories of having perceived the color (before actually perceiving it), she would not be surprised when she walks out of her room and perceives it for real. What she experiences on walking out the door would correspond with her already-present artificial memories, so there would be little new information to remember about the experience, and no surprise to be felt.

causative
  • 12,714
  • 1
  • 16
  • 50
  • Not a philosophical expert but if you are saying she has knowledge of the aftermath of experiencing a color, she still does not have knowledge of the direct experience of perceiving a color, is that right? In which case Jackson's argument survives. – user4894 Jun 03 '22 at 23:13
  • @user4894 I'm saying her knowledge of the aftermath of experiencing a color takes the form of having a false memory of experiencing the color. If she has that false memory - which is purely a matter of the physical arrangement of neurons in her brain - then she will not be surprised when she sees the actual color. – causative Jun 03 '22 at 23:39
  • This is why you are not supposed to talk about koans. – Scott Rowe Jun 04 '22 at 12:18
  • 1
    One of the points of Jackson's article was to point out the difficulties of forcing our actual conscious experiences into the physicalist model. Treating qualia as "false memories of having experiences" definitely fits into the category of "forced" model fitting. As Quine noted, theory is always under determined by evidence, so the existence of a forced method of accounting for qualia and the knowledge we gain from them is expected. Theories are judged not by refutations, but by their degree of "forcedness" (Occam). The "forced" nature of the explanation, is still strong counter-evidence. – Dcleve Jun 04 '22 at 16:02
  • 1
    @Dcleve Why do you say it's forced model fitting? In the physicalist model, remembering having experienced a qualia can be done regardless of whether the memory corresponds to an event that happened or not; this is something that falls out of the model, rather than an additional complexity tacked on to it, so it isn't penalised by Occam's Razor. – wizzwizz4 Jun 04 '22 at 16:40
  • @wizzwizz4 -- Jaskcson's thought problem highlights": a) we have a qualia of seeing; b) we have both a memory of the qualia of seeing, and knowledge based on the memory of the qualia of seeing; but c) in physicalism, a) is not in the causal chain for b), instead some process physical process creates both a) and b), and a) is an actual irrelevant phenomenon. In physicalist theory, we could get to b) without a) but we do not. The much more plausible explanation is that a) is actually the cause of b). – Dcleve Jun 04 '22 at 17:09
  • 1
    @Dcleve In physicalism, qualia is a physical process. The causal chain is the same. In order for Neuroscientist Mary to have gained knowledge of everything about colours without seeing colours, she would've somehow had to gain these memories in a different way – perhaps via dreaming after long hours of study; perhaps via the handwaving common to the premises of these thought experiments – but this sort of situation doesn't tend to happen very much in real life, so I'm not sure how much conclusion you really can draw from it. – wizzwizz4 Jun 04 '22 at 17:18
  • @wizzwizz4 Multiple realizability is considered to have refuted the reductive neural identity theory you are advocating. Jackson's thought problem was addressing the functionalism that replaced it, with the issue that functional identity theory does not need qualia -- they do not serve a causal role in functionalism. The function of perception, and memory, etc work just fine without qualia. Qualia seem to be a non-functional appendage, yet they also play a role in knowledge, as Jackson points out here. Jackson's thought problem is not aimed at reductive identity theory, but at functionalism. – Dcleve Jun 04 '22 at 19:09
  • 2
    @Dcleve I maintain that qualia do serve a causal role, because qualia are nothing more than an alternate way of looking at some parts of physical reality, and therefore they have the causal role of the physical things they are equivalent to. This is the view of panpsychism; mind is matter, and matter is mind, and causation among matter is causation among mind, and vice versa. – causative Jun 04 '22 at 19:14
  • 2
    A separate point: the event of Mary experiencing surprise is a physical event, which is to say, it has physical effects like Mary saying "I am surprised," and it is embodied in Mary's neurons. So the physicalist perspective is to say that Mary's surprise has physical causes, based on how the prior contents of Mary's brain interacts with the new information coming in through Mary's senses. If (counterfactually) Mary is surprised, it would only be because her visual cortex fails to trigger the neural structures encoding the memories of Mary having seen something like that before. – causative Jun 04 '22 at 19:16
  • @causitive -- Identity theory is a claim about our world, and is therefore testable in principle. It fails several predictive tests, including multiple realizability, and 100% correlation. "Occasional" identity is all that can be claimed about any physical or functional identity, with the reasons between occasion and failure TBD. OIT then fails the evolutionary tuning of consciousness test. One can of course continue to maintain IT despite it failing multiple falsification tests, but one must be ready then to justify this view, or be seen as unsupportedly dogmatic. – Dcleve Jun 04 '22 at 20:33
  • @causative the claim that Mary experiencing surprize is a physical event, is untrue. It is a mental event. That mental events ARE physical events is not a fact, it is a claim, one which has been unable to be supported to date, per the Hard Problem of Consciousness. The strongest claim you can legitimately make is that it MAY be shown to be true someday. – Dcleve Jun 04 '22 at 20:37
  • 3
    @Dcleve It doesn't fail predictive tests. In what way is "multiple realizability" a predictive test? It's not a test, it's a theory about how the same consciousness can be implemented in multiple ways, as long as the causal structure is preserved. In what way is "100% correlation" a predictive test? Anyway, regarding the claim that Mary's surprise is a physical event - this is the physicalist's internally self-consistent answer to the "Mary the color scientist" scenario. You can argue it's dubious if you wish; but the "Mary the color scientist" scenario isn't something that refutes it. – causative Jun 04 '22 at 20:41
  • @causative It is not clear to me how, according to this theory, Mary would acquire these false memories - from reading about this theory? – A Raybould Jun 07 '22 at 13:00
  • @ARaybould The premise of Mary the Color Scientist is that she knows everything about the color. This is an incredibly strong claim that requires us to rethink what it means to know something. To know everything about a color, including, among other things, how every single neuron in the brain (out of 100 billion) responds to seeing it, we can suppose that she is doing more than simply reading. The human brain can't normally contain that much knowledge. Building a false-memory-implantation machine could be part of her research. – causative Jun 07 '22 at 15:13
  • @ARaybould In any case, whether Mary is surprised or not when she leaves her colorless room, the physical reason is clear: she is not surprised if her brain already contains the neural structures that correspond to memories of seeing a color, and she is surprised if her brain does not contain those neural structures. So whether you think Mary having false memories follows from the premise or not, a physicalist has no trouble accounting for what happens when she opens the door. – causative Jun 07 '22 at 15:17
  • @causative -- the refutation just comes from the definition of "identical". If A is identical to B, then whenever we have A, we have B and vice versa. If A is a state of a particular nervous system, and B is a thought, function or experience, than that same nervous system in a different state, or a different nervous system, cannot be identical to B, and if we find B (say "pain upon stove burn" or "knowledge that 2+2=4") with a DIFFERENT nervous system, or a different state of the same system, and we do, this is a falsification by test of the Identity claim. – Dcleve Jun 07 '22 at 16:43
  • Likewise, if we have a state of a nervous system A, and do NOT have B, this also is a falsification of the identity claim. From Paul Churchland's The Engine of Reason, he asserts that consciousness just IS the processing of a recursive feedback neural net, so all neural nets doing recursive feedback processing, should be conscious. All our neural processing is recursive feedback processing, but 99+% is not conscious, so this identity claim is tested and falsified. Likewise, all algorithmic consciousness identity claims are falsified by our multitudes of unconscious algorithms. – Dcleve Jun 07 '22 at 16:49
  • @causitive -- the assumption that what matters causally is "feeling pain from the stove burn" or "knowing 2+2=4" not either of the above FOR ONE SPECIFIC NERUAL STRUCTURE AND STATE -- almost all neural identicality thinkers switched to being functional identity thinkers in the late 60s to 70s, and for many decades this was the consensus view of physicalism. Multiple realizability refuted physical identity, but not functional identity. The causal nature of qualia finally broke this consensus, in favor of non-reductive emergence, and Jackson's thought problem was a big part of that process. – Dcleve Jun 07 '22 at 16:55
  • 1
    @Dcleve I don't imagine anyone has ever held the view that "pain upon stove burn" can only be produced by a single nervous system inside one particular person, and no one else can ever feel "pain upon stove burn" just because they don't have perfectly identical nervous systems to that one person. Instead, it's understood that if two physical systems are doing "essentially the same thing" then they can implement the same qualia. – causative Jun 07 '22 at 17:09
  • Let us continue this discussion in chat. – causative Jun 07 '22 at 17:10
5

Mary can (by assumption) only reach certain mental states—states corresponding to the experience of color, or the memory of the experience of color—by shining colored light into her eyes.

This is alleged to tell us something fundamental about the nature of consciousness. It clearly doesn't. The most plausible reason why human beings work that way is that it's maladaptive to experience things that aren't there (such as a color) or to remember things that one didn't experience (such as a memory of the experience of color). Selective pressure would naturally lead to the brain being wired that way. There is nothing to suggest that evolution was forced into making qualia work that way by the nature of qualia, whatever that might be. So the argument tells us nothing about the nature of qualia.

Some computers with a secure boot process have a physical debug-mode switch that can be flipped to make them run unsigned software. Flipping that switch puts the computer into a state that can't be reached by any other form of input (key presses, network data, etc.). That doesn't tell us anything deep about the nature of computing. It just tells us that the computer was constructed that way for some reason. (The reason was security—which is basically the evolutionary reason too.)

If Panpsychist Mary is like Neuroscientist Mary except that she reads about panpsychism instead of neuroscience, she won't learn what the experience of color is like. The content of what she reads is irrelevant. Her mind is wired to prevent that state from being reached by that form of sensory input.

benrg
  • 1,208
  • 6
  • 11
  • I couldn't agree more – Nikos M. Jun 03 '22 at 20:29
  • 1
    It is not whatever Mary reads that grants knowledge of color experience. Experience provides a form of knowledge that "study" cannot. The point of the thought experiment is to note that this experiential knowledge cannot be readily fit into a physicalist worldview. Studying panpsychism is irrelevant to the question. The question is how does this type of knowledge fit into a pan-psychist worldview. – Dcleve Jun 03 '22 at 22:46
  • 3
    @Dcleve Study is of course a form of experience, so I suppose you mean that different forms of experience can produce different forms of knowledge. I agree, and that was the whole point of my answer. This can be readily fit into a physicalist worldview. I don't know why anyone would think otherwise. The question asks about the original Mary experiment with panpsychism substituted for neuroscience. The original Mary didn't have a "neuroscientific worldview", she just knew a lot about neuroscience, because she studied it. – benrg Jun 04 '22 at 06:55
  • People seem to have a lot of 'maladaptive' thoughts these days. – Scott Rowe Jun 04 '22 at 12:21
  • @benrg -- yes our thinking and study also have qualia associated with them -- we generally do not study or reason unconsciously. Philosophy of mind tends to focus on perceptual qualia, as that helps sort more clearly between functionalism and causal qualia theories. In "study" the qualia of thinking can be more effectively ignored, as a causal chain can be created just based on the function of what was learned/accomplished in the study. Postulating an incompatible dualism within neurology to deal with qualia, without evidence, looks like a clear kluge to physicalism. – Dcleve Jun 04 '22 at 16:11
  • 1
    @Dcleve panpsychism is not necessarily dualistic. At least some important versions are not eg https://www.sjsu.edu/people/anand.vaidya/courses/c2/s0/Realistic-Monism---Why-Physicalism-Entails-Panpsychism-Galen-Strawson.pdf – Nikos M. Jun 05 '22 at 15:11
  • @Dcleve As the person who asked the question, I can state authoritatively that the question is not how this type of knowledge fit into a pan-psychist worldview, and is explicitly about the consequences of studying panpsychism. If there is an answer to the latter to be found in the former, then make that case, but don't try to change the question - ask a different one, if you want to. – A Raybould Jun 05 '22 at 15:11
  • @NikosM. -- Strawson explicitly claims both experiential and non-experiential aspects to matter. This is just property dualism. He tries to redefine physicalism (and asserts basically all other physicalists are NOT physicalists!) to obscure that he is asserting a variant of dualism, so that is not an effective counterexample. Also, he says he is only OPEN to panpsychism, not that he is asserting it. Strawson seems to have absorbed the anti-dualist smears common in philosophy, and I think does not have the courage of his convictions here. Most other dualists and pan-psychists are naturalists. – Dcleve Jun 07 '22 at 15:41
  • @Dcleve, I will not agree with this interpretation. In the same sense that asserting charge and non-charge properties of matter is not dualistic. What he refers as experience is defined as physical property. – Nikos M. Jun 07 '22 at 15:53
  • @NikosM. -- in property dualism, experience is defined as a physical property of matter. You have just spelled out property dualism. There are non-experiential and experiential properties to matter. https://en.wikipedia.org/wiki/Property_dualism#:~:text=Property%20dualism%20describes%20a%20category%20of%20positions%20in,kinds%20of%20properties%3A%20physical%20properties%20and%20mental%20properties. – Dcleve Jun 07 '22 at 16:02
  • @Dcleve i know the definition i just say that this is not my interpretation. In the sense that mental is not distinguished from other physical properties. – Nikos M. Jun 07 '22 at 17:11
5

There are significantly different panpsychisms, and the answer is different for them.

For Mind at Large Perennial Philosophy panpsychism, causation comes through the idealist path, and matter is derivative. Mary unconsciously blocks panpsychist knowledge of everything, available thru Mind At Large, and meters this knowledge to be in sync with her "studies". Everything one learns in life in this version of panpsychism, IS just reaccessing knowledge we already had available to us, but which our minds had unconsciously suppressed. So when Mary finally perceives color, she will only be remembering prior blocked knowledge. She will still THINK she is surprised, but will not "really" have been surprised.

For neutral monism panpsychism, causation runs through the neutral ontology that spawns both matter and mind. Her lack of access to colors, and lack of knowledge of the experience of color, are both caused by some more fundamental reality, which is then corrected, leading to both her exiting her black and white world, AND experiencing color. She will be legitimately surprised in this panpsychism.

Property dualism panpsychism operates similarly to neutral monism, in that experiences are maintained in parallel to physical events, by some TBD property of matter (similar to the TBD neutral ontology). Per David Chalmers version of this, matter is causally closed, and knowledge is encoded in neurons, and comes from the material events only, so the actual experience does not lead to the knowledge of that experience (Chalmers pan-psychism is epiphenomenal). But this epiphenomenal aspect of this pan-psychism does not change the event that one will have actual "surprise" when one gains knowledge of color experience.

Dcleve
  • 13,610
  • 1
  • 14
  • 54
  • If we stopped being surprised by our experiences, there wouldn't be much point in hanging around. There have been a few SciFi stories of AIs who spontaneously shut themselves off. Nothing left to learn, I guess. – Scott Rowe Jun 04 '22 at 12:14
  • I appreciate that you have considered a range of panpsychist viewpoints. With regard to Mary only thinking she has been surprised, however (and putting aside my doubts that this anything other than just simply being surprised), is that not the same, for the purpose of the Knowledge Argument, as really being surprised? It seems to me that the only distinction that matters to this argument is whether her reaction is a Dennettian "just as I expected", or anything else - surprise, a curiosity satisfied, etc. – A Raybould Jun 07 '22 at 12:37
  • @ARaybould -- Under Perennial Philosophy panpsychism, we have access to all knowledge in the universe, but deliberately step down our knowledge to avoid mental overload. And prior to incarnation, we were bathed in universal knowledge, and have to "forget" to operate in our local focused environment. In this worldview ALL knowledge is seen as "remembering", and no knowledge is "really" a surprise. But yes, Mary will THINK she is surprised. I wasn't sure what you were trying to ask about panpsychism with the question, and PPP basically asserts a variant of Delusionism vs knowledge and surprise. – Dcleve Jun 07 '22 at 15:56
  • @Dcleve It is not entirely clear to me what you mean by any of the words you put in quotes, or whether "having access to" is to know, but as far as I can tell at this point, you seem to be saying that, according to Perennial Philosophy Panpsychism, Mary does not learn anything from either studying said panpsychism or from seeing colors, as she already "had access to" everything that could be learned by those means - is that correct? If so, what about second-order knowledge - learning what it was she "had access to" but had suppressed? – A Raybould Jun 08 '22 at 12:13
  • @ARaybould -- Delusion theories basically operate with two sets of books. For PPP, in the public book, Mary learns and is surprised. PPP holds that reality is in the second set of books, where Mary is remembering, and is not really surprised. I don't think that studying PPP would grant the knowledge that would prevent the surprise, but I don't hold by PPP. In most eastern traditions, sufficient study would allow one to transcend the limits of ones delusions, and Mary could perceive colors directly from her room, I don't believe this happens, and that may be skewing my description of PPP. – Dcleve Jun 08 '22 at 14:07
  • @Dcleve Thank you for your comments. This lies beyond the debate that I am familiar with, so I would greatly appreciate any references to articles or papers that would give me a basic understanding of the key concepts. Who are the main proponents of these views? – A Raybould Jun 08 '22 at 15:53
  • Perennial Philosophy was argued for by Aldous Huxley. It is the primary POV of the team that prepared the Irreducible Mind, and Beyond Physicalism series. It is also the POV of Bernardo Kastrup. Here are my reviews of two of these works: https://www.amazon.com/gp/customer-reviews/RZY1A4EL2JOZ4/ref=cm_cr_dp_d_rvw_ttl?ie=UTF8&ASIN=153812596X https://www.amazon.com/gp/customer-reviews/R37ST3Z9M42Q7E/ref=cm_cr_arp_d_rvw_ttl?ie=UTF8&ASIN=1782793623 – Dcleve Jun 08 '22 at 16:39
2

There are a couple of reasons why Mary the Panpsychist wouldn't learn anything different - one being that the Mary-the-Neuroscientist thought experiment doesn't really address the qualia issue either, the other being that panpsychism is about a different issue.

Panpsychism avoids the problems of understanding what is physically special about humans such that they experience qualia by asserting that everything experiences qualia. Physicalism says the brain follows the same laws of physics as rocks and trees; there is no magic that causes the atoms to move differently just because they happen to be in a human brain. So physicalism only tells us that either both can experience qualia or neither can. Since we know that humans do experience qualia, to be consistent we should therefore deduce that rocks and trees can too. (To be more precise, we can deduce it about systems that manipulate information, in the technical Information Theory sense.)

Thus, Panpsychism is only making novel assertions about non-humans. As regards Mary's own experiences, it makes no difference. Panpsychism says Mary can experience qualia too, but we already knew that. So being a panpsychist tells Mary nothing extra about her own experiences. She might be able to have a go at questions like what it would be like for a computer to have a colour camera plugged in for the first time, though.

Separately, the knowledge argument is really about the limitations of simulation. We can experience involuntary and uncontrollable sensory impressions of the outside world. We can also use our imagination and memory to internally simulate the sensory experience, and verbal/textual descriptions can be used to guide the process of building such a simulation. Language maps words to sensory experiences, and patching the words together tells to what simulated experiences to combine and how. But learning the meaning of words relies on us having memories of the relevant experiences to call up. On reading that a unicorn looks like a horse with a single horn coming out of its forehead, we can patch together a memory of a horse and a memory of a horn and imagine what it looked like, and thus not be surprised when we meet one. But you need a base vocabulary of remembered experiences to assemble. The fact that you can't understand the word "red" without having experienced it is thus a limitation of language. The assumption being made here is that physicalism allows us to completely simulate the world using language alone, which can be understood without any need to have experienced any of it. Thus, the fact that language about colour cannot produce a complete simulation of the experience without any previous memory of it refutes this version of physicalism.

Mary the Panpsychist might therefore ask whether a computer with a camera would be able to do this. Suppose the computer is told that 'colour' is when you get an 'image file' where the 'red'/'green'/'blue' fields at a point have different values. Could the computer simulate the experience? Well, it depends how it has been programmed. It might be wired so that the only way it can create image files is with the camera, or by cutting/pasting areas of existing stored image files together. In that case, it wouldn't be able to simulate 'colour' either, until it had some files to work with. Or maybe it is wired so that it can construct colour images even without any examples to work from. We could likewise imagine a human brain wired with full internal control so that a particular mental effort triggers the experience of colour - we just stimulate those particular neurones in our own brain. It's not a fundamental limitation of how the universe is able to work. It's just an accidental property of how our brains are wired - that we don't have full internal control of it.

Nullius in Verba
  • 21
  • 1
  • I think it's no accident. If humans had full internal control of their brains, that would be stable for about 3 seconds. But for a look at what it might be, see the movie "LUCY". – Scott Rowe Jun 05 '22 at 12:21