The mind/body problem is the feeling/function problem (Harnad 2001). The only way to "solve" it is to provide a causal/functional explanation of how and why we feel...
Click here to view --> HARNAD VIDEO
Note: Use Safari or Firefox to view;
does not work on Chrome
But... "Claustrum Nostrum"
"To suppose that something that is molecule-for-molecule identical to me could fail to have feelings sounds about as sensible as to suppose that something that was molecule-for-molecule identical to the moon could fail to have gravity." So T5 passing T4/T5 would solve the hard problem?
ReplyDeleteAh! But you say that the hard problem is insoluble later in the comments. Is it because T5 is not enough or because T5 is impossible?
DeleteEmma, please combine your commentaries rather than give several in a row. See reply below.
DeleteIn Chalmers' replies to you, "It's a manifest fact about our minds that there is something it is like to be us", I found myself asking if there are such terminological debates in other cultures. A lot of false assumptions about cognition seem to be rooted in "individualism" or more kid-sibly, an importance attributed to who I am as a person, and the value of persons in general. I wonder if less individualistic societies have better answers to the hard problem ?
ReplyDeleteEmma,
Delete1. "something it is like" is weaselly: What is meant (and should be said) is "something it feels like" -- or, more simply, and kid-siblier -- "something is felt". And, yes, it's individuals that feel, not crowds, or cultures.
2. The notion of a T5 zombie is metaphysical nonsense (apart from the OMP or normal scientific underdetemination). And it is the explanation of the causal mechanism of a T3/T4 robot that would solve the Easy Problem. Would the mechanism of a T5 biorobot solve the HP (even if we set asidee the OMP)?
My main takeaway from this reading is that when it comes to ‘consciousness’, the main question is the hard problem: how/why do we feel? Beliefs and experiences are also feelings as we experience them as such. As mentioned above, T3/T4 can be used to solve the EP (how/why we do what we do) because by definition T3/T4 does what we (humans) do. However, to solve the HP, something that feels as we feel must be reverse engineered. Without the OMP, a T5 that feels could be reverse engineered to solve the HP. A T3 or T4 could be a better candidate for reverse-engineering as they may have more understandable causal mechanisms, but they can only be used to solve the HP if we can guarantee that they feel.
DeleteEmma, this is a super interesting idea! In more collectivistic cultures, I believe feelings are more thought of as atmospheric, environmental. If it were the case here, the hard problem would be a question less for cognitive science and more for social psychology. While this is getting abstract, trying to dig deeper and deeper into the individual's experience has lead to so many dead ends (the EP, the HP, the OMP) that one starts to wonder if the questions are better asked elsewhere.
DeleteI found it interesting how Dennett separated the questions posed by Descartes and Kant, while professor Harnad has combined them. Dennett believed that the questions represent fundamentally different approaches to understanding the nature of consciousness. While Harnad believes both are important for answering the hard/easy problem. Descartes' question is concerned with the causal relationship between the mind and body, while Kant's question is concerned with the epistemological relationship between the mind and the world. Harnad argues that both questions can be answered by providing a causal/functional explanation of how and why we feel.
ReplyDeleteMarie-Elise, why is it a "Hard Problem" (HP) to explain how and why we feel? That's what is at issue here, not Kant or Descartes.
DeleteAnd what is Dennet's "heterophenomonolgy"? And is it a solution to the HP or does it who that there is no HP?
I will attempt to answer your questions professor Harnad but I still have some misunderstanding about its connection to the hard problem.
DeleteHP is considered "hard" because while science can explain various brain functions (the "easy problems"), like memory or perception, it can’t explain why these functions come with subjective experience - When you remember an event, you don't just recall the facts but also how you felt during that event. The emotional quality of the memory is the subjective experience. It’s the same thing with perception; when you see a color for instance, it’s not just about the mental process of recognizing the wavelength of light, it's also about how you experience the color – its vividness or the emotions it evokes. These personal experiences are unique to each person and are more than just the sum of neural activities. They are what make our inner lives rich and complex, and this is why explaining them purely in terms of brain function remains a major challenge, hence the “hard” problem.
Heterophenomenology brought by Dennett, is an approach to studying consciousness that combines third-person data with first-person reports of experience. By observing and interpreting people's reports of their subjective experiences without making assumptions about the reality or validity of those experiences, his method aims to create a comprehensive, external description of human consciousness.
However, heterophenomenology may not fully solve the Hard Problem because the Hard Problem centers on why and how subjective experiences (qualia) arise from physical processes in the brain. It focuses more on describing and categorizing these experiences rather than explaining their fundamental nature or origin.
Could Dennett's concept of heterophenomenology suggest that there is no such a thing as a "Hard Problem" of consciousness? By analyzing both the objective data about brain states and subjective reports of experience, the apparent mystery of subjective experience (the essence of the Hard Problem) is more about language and conceptual understanding than a fundamental gap in scientific knowledge. Does this mean that from Dennett's perspective, the Hard Problem is not be an insurmountable mystery but rather a challenge of proper interpretation and understanding of consciousness?
Heterophenemology focuses on connecting subjective verbal accounts with their objective physical states. This method does not solve the hard problem because it doesn’t answer the how and why that we are after. Dennet aims to use heterophenemology as a branch between objective scientific measurement and phenemological accounts, but does not actually asses feelings. Even if the interpreted verbal accounts of “convictions, beliefs, attitudes, emotional reactions” along with the objective physical data were enough to draw a causal relationship, this method would not be able to solve the hard problem.
DeleteAmelie, good synthesis, but:
Delete(1) Both "subjective" and "experience" are, in the context of the HP, weasel-words. Here is a sample of the quasi-synonyms that are usually played off against one another as if more than one thing were at issue in the HP. (Taken from one of the PPTs for Week 10):
WEASEL-WORDS for feeling: consciousness, awareness, qualia, subjective states, conscious states, mental states, phenomenal states, qualitative states, intentional states, intentionality, subjectivity, mentality, private states, 1st-person view, 3rd person view, contentful states, reflexive states, representational states, sentient states, felt states experiential states, reflexivity, self-awareness, self-consciousness, , raw feels, experience, soul, spirit, mind...
If you stick to the words feel/felt/feeling you will confuse yourself and others much less about the HP (which is hard enough without weasel-words). You can show how weaselly the weasel-words are if you substitute the f-words in their place: An example is "subjective experience", which becomes the absurd and redundant "felt feeling."
(2) "Feeling" does not refer only or even mainly to "emotions": yes, all emotions are feelings, but most other feelings are not emotions: tiredness, warmth, coldness, wetness, darkness, brightness, hearing, seeing, tasting, touching, moving, wanting, believing, understanding, meaning, thinking -- these are all felt states: states that it feels like something to be in. Some of them are emotionally positive (pleasant) or negative (unpleasant), but most have no particular emotional valence at all (emotionally neutral). Yet they all feel like something.
(3) Please read the other replies (10a and 10b). DD does not solve the HP: He denies that there is an HP, because feelings are really just "beliefs" about feelings.
Megan, felt states are correlated with physical states, but that does not explain how or why the states are FELT rather than just DONE ("zombily").
First, I am confused with some terms. Since a belief, as an example of cognition, is a felt state, cognition in general refer to the activities happening in the brain only during the felt state. On the other hand, activities happenng in the unfelft state inside the brain should not be consider as cognition. How does "unfelt state" comparing to vegetative state? Are they the same?
ReplyDeletePersonally, I think Emma made a really good point. It seems that a T5 candidate is able to have feelings because "something that is molecule-for-molecule identical to humans must to have feelings". And based on that, the moment that humans are able to create a perfect T5 candidate, through reverse engineering process (not any other ways since humans can always create T5 candidates through laboring), I think HP will be solved or the answer will be at least hinted at that moment.
I agree in thinking that a molecule-for-molecule replica of a human being would have feelings since they're literally the same as us, but I think that even if we made such a T5 passing entity, we wouldn't know what it is that makes it feel. I don't think I understand what you mean by "laboring" for the process of making a T5 candidate, but I'm not sure we would have an answer via reverse engineering either. If adding or removing one type of molecule or one brain structure leads to the candidate having feelings, would that be a sufficient answer to the Hard Problem?
DeleteEugene, yes, believing something is a feeling, but not all feeling is believing: it can be seeing, hearing, wanting, touching, etc. The boundary between "cognitive" states/capacities and vegetative states/capacities is fuzzy, or even arbitrary. And as the example of Ohrie trying to recall her 3rd grade school teacher showed, even during the felt cognitive state of recalling, we only feel the output of the ongoing brain processes; we have no idea of what our brains are atually doing, nor how they produce their output. Cognition is thinking, but HOW the brain produces thinking, or any of its cognitive (or vegetative) capacities is both unfelt and -- till cogsci figures out what they are and explains them causally -- unknown.
DeleteIf cogsci successfully reverse-engineers T3 capacity, it has shown how to produce T3 capacity; if cogsci successfully reverse-engineers T4, thenit has shown how to produce T3 capacity with synthetic T4 function; if T5, then with natural biological T5 function. But in all these cases, the explanation is in how the causal mechanism works. Producing a T5 (or T4 or T3) without having reverse-engineered it and explained how it can DO what it can do is as useless to cogsci as cloning a T5, or simply producing it by the usually means of human reproduction. (I think you understood this, but I'm saying it explicitly, just in case.)
Omar, I think what Eugene meant by "laboring" is explaining causally how a successful T3, T4 or T5 can DO what it can do. What it can do, outside and inside, is all observable, right down to the molecule. That's all just EP. But because of the OMP, whether and what a T3 or a T4 or a T5 (or a stone) feels is observable only to the T3, T4, T5 (or stone). It's absurd to doubt that a T5 feels.
But even a T3 together with a divine certification that it feels would only explain how and why it can DO what it can do, and not the fact that it feels. If you removed or altered some widget that did not alter anything the candidate can DO, externally, and now you got a divine certification that it does NOT feel, that still would not explain how or why it feels. It would just show that TT could be passed without that T4/T5 widget. That's why the HP is so hard: because the EP seems to cover all of the available empirical, causal, explanatory territory. This is what makes the HP, because of the special nature of feeling, even more underdetermined than ordinary scientific (and engineering) explanation.
This comment has been removed by the author.
DeleteFrom this reading, I understood that one of Professor Harnad's core points regarding the insolvability of the Hard Problem is the insufficiency of correlation as an explanation for feelings. I have trouble understanding that if, for example, someone with damage to an area in the brain stops feeling, it would be incorrect to think that this is valuable information in addressing the Hard Problem. I understand how it would not be a causal explanation for why feeling emerged in the first place, but would it not nonetheless be informative by indicating that this brain area is responsible for feeling, as damage to it causally resulted in a lack of feeling?
ReplyDeleteI remember this was a common theme in skywritings for Fodor's 4.b reading (why localizing certain "doing" functions in the brain isn't important), and I kind of agree that knowing what areas are responsible for certain functions/feelings seems important, but I think the point is that just because this brain area is "responsible for" that feeling, or in other words, activity in this area predicts that feeling, we're no closer to the explanation of how this area causes these feelings by simply finding their neural correlates. Not sure if this adequately answers your question though!
DeleteHi! I must agree with Jess, yes it makes sense that localization would help get some information on feelings - which "area" do they come from. However, it does not explain how this particular area of the brain itself generates feelings. Also, if you were wondering about somatic responses and how sometimes it is a bottom-up process, it still does not bridge the gap between the feeling and the cause. "There is a lot of noise, so my ears tell my brain there is a lot of noise, I become stressed because my brain tells me to be stressed". But how does my brain tell me this? It is starting to be an infuriating question but remains unanswered.
DeleteOmar, Jess, Garance, the reason the HP is so hard (hence frustrating) is that for the EP cause/effect is "transparent": remove a widget and you lose an observable (DOing) capacity. Then you can wrk out how and why produces or contributes to the EP doing capacity. But when you pull out the widget and the EP doing-capacity remains intact, and only the feeling disappears (certifiable only by an omniscient authority) you have not solved the HP, just confirmed how hard it is.
DeleteThe only hope is that, with the loss of feeling, some doing capacity is lost too, and that someone can figure out how and why the feeling is needed for the doing. Till then the HP is as unsolved with the discovery of the widget as mirror-capacity is unexplained by the discovery of "mirror-neurons."
Exercise: What would the referent of the content-word "feel" be for a T5 (or T4 or T3) zombie? In what would that word be grounded? What would it mean when it says (like ChatGPT) either that "I feel" or "I don't feel"? (That's a connection between the HP and the SGP.)
To attempt to answer the exercise question: A T5 zombie has the ability to ground symbols, as well as every other doing capacity a human has. For one, when this zombie says, “I feel,” it is lying, but this is beside the point, since it still means something (even as a trick). The zombie’s grounding for the word “feeling,” rather than being based on feelings of its own, would probably be built on data it has gathered (and corresponding categories it has built) by interaction with people or animals that do feel. The zombie could understand what feeling is, and what kinds of things people do because o feelings, without knowing what it feels like to feel.
DeleteI’ve assumed that the zombie would know that it is a zombie, and that the people it is interacting with are not. Though this is sci-fi, I wonder what would happen if the zombie thought that it has feeling when it in fact does not, since there would be nothing to distinguish it from everyone else in terms of doing, and admittedly everyone else feels. Such a zombie would have an incorrect understanding of feeling, so the referent of their version of ‘feeling’ might be something correlated with feeling that the zombie does.
In response to the exercise, I hesitate to think that the word “feel” (and maybe any word at all) could truly be grounded and that a zombie could remain a zombie. If you (or me or a machine or a zombie) know what a word refers to, it feels like something to know that… I think. Because words are arbitrary symbols, there must be some sort of feeling related to the referent to actually do the meaning-grounding. This might get into why we could need feeling for doing, but I’m really not sure on that front. So Adam, I am not sure I agree that a zombie could understand what feeling is without knowing what it feels like to feel. As we’ve seen in class, it feels like something to understand (or not understand) Hungarian. It’s possible I’m committing myself to the position that truly reverse engineering the things we do as thinking beings in a T3 or a T4 would produce something that it feels like to be that machine… for the moment I think I’m OK with that.
DeleteFor me, the idea that a T5 zombie could ground the word "feel" based on observed data and interactions with feeling beings seems plausible. But the grounding wouldn't be through personal experience but through the collection of external data. Isn't it a bit like the Searle's chinese room? Where you feel like you understand because you memorised the data?
DeleteI agree with Adam when he says that a T5 zombie might lie when saying "I feel". The lie would be that he understands the concept of feeling but does not actually feel.
Elliot, that’s a really interesting thought. I look at the SGP as a possible avenue towards solving the EP. What you are suggesting is that we may need feeling (HP) to solve the SGP (EP), definitely thought-provoking. Though I am going to push back a bit.
DeleteI am not so convinced that knowing what a word refers to necessarily entails a feeling. Like you mention, it feels like something to understand, not that “understanding” itself is a feeling. I think its perfectly plausible that a T3 could “understand” in the sense that it could do everything that we can do with a particular object, describe it, use it, pick out the physical referent - all purely through sensorimotor grounding. Because of the OMP, we can never really say if the T3 is incapable of feeling, but its certainly possible that its doing everything its doing (including describing to us that it feels) purely in a sensorimotor fashion. This only works toward solving the EP, as long as its possible that the solution to the SGP is purely sensorimotor, we’ve still made no headway on why or how we feel.
By damaging or removing the Amygdala, which is critically involved in processing feelings, particularly those related to fear and aggression, it does not result in a complete absence of feelings (a reduced feeling of fear would be shown after instead). In fact, this is also OMP because theoretically, it could also become a new feeling of "having no feeling", but we never know. We can receive some hints from the causal relation between neural structures and certain specific changes of feeling, such as decrease/ increase of certain feelings due to certain structural damages; but we cannot solve the HP through this process.
ReplyDelete"The gap, in order to solve HP, is between non-feeling and feeling, not between this-feeling and that-feeling".
DeleteEugene, that's right.
DeleteIf you think of this hypothetical situation in which you manage to remove feelings from someone completely, that person would also lose all their memories as well, because memories are grounded in the feelings that a person recalls experiencing during that time in the past. If you remove the capacity for feeling, would you effectively remove all the person's memories as well? When you think of it in terms of neural correlates, would it be possible that feelings would re-emerge naturally, because that's what the brain is programmed to do, or au contraire, would the individual just be vegetative functions?
DeleteYes! The causal relationship between neural structures and specific changes in feelings doesn't necessarily solve the broader issue of the Hard Problem. Linking various sensorimotor events and mapping the brain does not explain how or why we are feeling. Knowing that there is an Amygdala still does not explain WHY.
DeleteMy question is could feelings potentially re-emerge naturally, given that the brain is programmed to experience feelings? Or, on the contrary, would the individual exhibit vegetative functions?
To me, the question is not really whether or not feelings could re-emerge naturally after having been removed from the brain, because even if they did so, the HP would still be unresolved. Indeed, even if so, it would only prove that feelings are a natural product of the brain that can be reproduced over and over. However, it would not offer an explanation for the how and why these feelings constantly occur.
Delete“I don't care if every nook and cranny, every last JND of my feeling life is correlated with and hence detectable and predictable from something you can pick up on your polygraph screen or can infer from my behavior. That's not the question! The question is: How/why does anyone/anything feel at all?”
ReplyDeleteThis quote, and this reading in general, emphasizes the fact that just because we can map all of the activity going in the brain, and link it to various sensorimotor events as well as felt states, does not mean we are closer to an explanation of how or why we are feeling during these events (ie. solving the Hard Problem). This relates back to Fodor’s "Why the Brain" reading (4.b), where he emphasized that all of these brain imaging efforts don’t really provide us with our required causal explanations of how the brain can do what it does. Although, if I’m recalling correctly, Fodor's piece focused more on the Easy Problem rather than the Hard Problem.
Jessica, you're right, on both points.
DeleteI share a similar thought to Jessica's: while science and technology may provide us with crude correlations between know-how and biological states in the brain, it's also possible that we'll never be able to casually explain those correlations in the same way that a doctor can explain the correlation between the heart's blood pumping and the observed performance of the heart.
DeleteHarnad emphasizes in this reading that heterophenomenology doesn't tackle the fundamental questions of causality necessary for understanding the “how and why” of the Hard Problem. It mainly offers correlations of feelings without addressing their underlying causes. Additionally, Dennett's argument about neutralizing different "types" of feelings is deemed irrelevant in solving the Hard Problem. Consequently, heterophenomenology, at best, equips us to address the Easy Problem, but falls short of providing a solution for the Hard Problem.
ReplyDeleteMelika, correct; HTRPHENO is just T3/T4 reverse-engineering + mind-reading (meteorology) based on correlations. That may predict whether or what someone is feeling, but not how or why.
DeleteI think that an interesting aspect of Professor Harnad’s paper is the back and forth on terminology between himself and Chalmers. Namely the distinction that Chalmers “might know it’s consciousness” but Harnad alternatively asserts that it is in fact “feelings”. This leads into why the idea of heterophenomenology does not allow us insights into the hard problem. This is because the issue is that Chalmers does not address the how and why questions surrounding feelings. He talks instead of lots of peripheral cases that do not get at the heart of the issue. Specifically the fact that beliefs about feeling are not what we should be referencing, rather the feelings themselves. Thus heterophenomenology will never solve the hard problem because it is fundamentally dealing with the wrong data and evidence.
ReplyDeleteJenny, correct; but Chalmers is not trying to solve the HP; he's only explaining that it's hard. And DD is not trying to solve it either, but to deny it's a problem.
DeleteThis reading clarified a lot of the concepts presented in the 10a reading. In fact, the reading revealed the holes and mistakes in Dennett’s argument, showing that what we should really be after is “how/why we feel like anything at all”. Through this reading professor Harnad shows how Dennett failed to understand the hard problem of cognition. Indeed, Dennett correlated emotions/feeling with physical states and responses to develop and justify his heterophenomenological approach. Dennett states in his paper that by using this approach we will be able to not only infer and describe feelings, but also to explain them. But this does not make much sense as a way to solve the hard problem of cognition which is asking how and why we can feel at all. To me it seems like Dennett’s approach is more suited and it could help to solve the easy problem of cognition which is how and why we can do what we can do, since his approach correlates physical states to mental states by collecting raw data and internal conditions and drawing interpretations from this. Thus, this could help to shed light on the internal processes and mechanisms that allow us to think, reason, have feelings.
ReplyDeleteValentina, that's right. We already know the brain causes both doing-capacity and feeling; we're only asking HOW? and WHY? Replying that feeling is just a belief is not an answer.
DeleteI find it interesting that many of the authors of our readings (Chalmers, Dennett) can't seem to narrow down the hard problem to feelings. There is talk about consciousness, internal states, experience etc., but these make the topic seem more complicated than it is, which is that many explanations of cognition don't account for the fact that it FEELS like something to think. To begin solving the hard problem, we need causal and functional explanations for feeling, not for "consciousness" or "experience".
ReplyDeleteAdrienne, yes, especially because "consciousness" and "experience" are just weasel-words (for feeling). See the other replies.
DeleteProfessor Harnard argues for incorporating subjective feelings into the study of consciousness, challenging theories solely relying on behavioral or functional criteria. This emphasizes the complexity of the hard problem and the importance of explaining not just behaviors or functions but also the subjective awareness that accompanies them. Moreover, his criticism of the term "belief" reflects a broader skepticism toward concepts lacking a direct tie to subjective experience. According to him, belief is a weasel word. He proposes that beliefs gain meaning only when connected to a subjective, experiential aspect. This places the discussion within a framework that prioritizes the first-person perspective, even when approached through external, behaviorally observable methods. In essence, Harnard shows that a comprehensive understanding of consciousness requires considering its subjective dimension, even when examining it from an external, behavioral standpoint.
ReplyDeleteJulide, you're more or less right, but try saying it all again replacing all the weasel words by the f/f/f words.
DeleteAya, you're more or less right too, but it doesn't solve a problem to say it's can't be solved. Feeling is a genetically involved brain caapacity. Lazy evolution does not encode traits -- especially traits as widespread as that one -- if they have no adaptive (hence causal) function. But explaining what that causal function seems to be harder to explain than any other evolved trait (even UG). (I certainly admit I'm stumped.)
In this reading, professor Harnad argues that Chalmers’ paper fails to address the Hard Problem and that correlation is insufficient to explain feelings, because it does not explain how or why we feel. If the brain is capable of doing all the same things without feeling, why do we feel? As professor Harnad says, “Why/how does any of [the] ‘drawing’ [of attention] have to be a felt drawing, rather than just a drawing? Why/how does any of that attention have to be a felt attention, not just a ‘selective processing.’” I think this notion is very interesting. My first inclination was that feelings are important as motivators, which drive us to do certain things. For instance, a mother feeling love for her child must to some capacity cause her to protect it more than if she just simply was following an instinct. However, the more I thought about it, the more I realized that this may not be true: evolutionary instincts can be powerful and motivate people to do things through chemical signalling without the accompanying ‘feeling’ of fear or love. My second thought was that somehow ‘feeling’ indicates intuitive information about internal or external states. For instance, I may feel ‘uneasy’ walking alone at night, which is an indication that I need to be aware of my surroundings. But that too didn’t fully explain the role of ‘feelings’, because I can simply experience alertness without feeling uneasy.
DeleteZoe, you've found all the rejoinders to your own questions.
DeleteI'm going to try to replace the weasel words.
DeleteProfessor Harnad urges us to explore how people feel, not just their actions, in the study of consciousness. This shows the complexity of the hard problem and the importance of explaining not just behaviors/functions but also the subjective awareness that accompanies them. Harnad criticizes using the term "belief,” considering it a weasel word. According to him, beliefs only make sense when connected to a subjective, experiential aspect. In studying consciousness, Harnad emphasizes considering how individuals feel, even when observing external behaviors. He shows that a comprehensive understanding of consciousness requires considering its subjective dimension, even when examining it from an external, behavioral standpoint.
I agree. I am also a bit confused, because it sounds like Heterophenemology cannot work from the beginning because it involves introspection. To combine first-person verbal description of feelings and third-person data, sounds like the first person part requires introspection on one’s own feelings (time to ask homunculi), and the third-person data (as it is observable) comes back to touch on the easy problem.
ReplyDeleteI’m starting to feel like the Hard Problem is impossible to answer. It seems our field is stuck using science to correlate behaviour to physical states (e.g., neural activity, neuroanatomy) or mental states (e.g., verbal reports). But the causal explanation for why it feels like something to be alive, alludes us. If our technology only allows us to acquire psychophysical data that is measurable, and this data is NOT enough to explain a causal mechanism for felt experience, what is left? Is there even an answer or is this entire class a rhetorical question? Is that the curse of being human, that we will forever live with enough awareness of our felt experience, but not enough understanding or knowledge to figure it out?
ReplyDeleteKristi, slow down. The HP is hard; possibly unsolvable. But definitely an empirical problem, not a "rhetorical question" or a "curse."
DeleteHi Prof, yes, I agree an empirical problem that needs pondering and testing…
DeleteProfessor Harnad argues that Dennett’s paper ignores the hard problem, asking how and why there are any feelings at all. For example — how and why is attention or learning felt, what is the function of feelings in those contexts? Harnad also argues against Dennett’s rejection of 1st person methods (ie: self-report of feelings), since only the subject themself can be certain of what they feel. Experimenters can measure overt reactions from a 3rd-person point-of-view, which may have correlations with certain feelings (ie: hissing after an electric shock may indicate the subject feels pain). However, there is always a possibility that the experimenter is wrong (ie: hissing because the subject feels startled by the unexpected shock). Since only the subject can report whether they feel the pain, it’s important to use 1st-person scientific methods, especially if one hopes to address the hard problem.
ReplyDeleteIf I'm not mistaken you're asking here about the evolutionary benefits that the feeling of deliberateness confers on those that possess it. To be honest I'm not totally sure the answer to that. It seems to me that many species get along just fine without it (some might argue all species other than human, I'm not totally convinced). Clearly the feeling of deliberateness is not inhibitory for survival given the relative success of the human kind in this endeavour however, couldn't it just be a non-inhibitory trait? I'm not sure I see any specific advantages being conferred by feeling as though one is the agent of their own actions. Maybe it allows further creativity in behaviour and thus led to many of the other developments (technological, behaviour, cultural, etc.)?
DeleteStephen, it's just a hunch, and not worked out. Many people, when they write about categorization, make a big fuss about the need to distinguish own-body from "the rest of the world": what's me and what's not-me. Of course feeling correlates with that -- I feel what touches me and not what touches something else. But feeling itself is unexplained, and a toy robot can easily be built to detect (or to learn to detect) and respond to that difference without feeling a thing. So no HP solution there.
DeleteMaybe the same is true of "agency" and volition: Tthe distinction between what I do "because I feel like it", and things that just happen, independent of my volition (including the reflexive response of my leg to a patellar tap by a doctor) could be detected, zombily, rather than felt. The difference is already there in the well-known visual effect that the world moves if the doctor (or even my own finger) moves my eye, but it stays still when I move my eye voluntarily with my eye-muscles (even though the image on my retina changes in both cases). The hunch is that the connection between FEELing and DOing is somehow highlighted there, making "detecting" some kind (non-motor) doing too. (Same is true when you just shift your attention, or do something, like long-division, "in your head".) This is also some sort of interface between unsupervised and supervised learning. But it's not clear why it would call for something as radical as feeling, rather than just some sort of perfectly zombyish "executive meta-detection". So it's probably just another false-start on a wannabe HP-solution.
And of course all sentient species have both voluntary and involuntary (e.g., reflexive) behavior.
So I doubt that armchair phenomenology has much of a chance with the HP...
It seems to me that one of the major bones of contention between Dennett and Harnad is in understanding what is the best (or at least a passable) way of grounding the feelings of others (OMP) on a common and coherent basis across individuals. Dennett says that we don't know if one person's experience of an emotional 10 is equivalent to another person's, but what we do believe is that the physiological response will be a consistent basis upon which we can compare. Harnad challenges this and says that while we can collect all this neural and behavioural data, we don't actually know what to do with the data once collected, and that interpreting feeling from these is impossible. It seems to come from an implicit disagreement of whether these physiological variables provide any insight into feeling. Team A says yes, while Team B says no. Debating between the two on the merits of different minutiae is a nonstarter since there is a fundamental disagreement (and one that can't be definitively answered) of the hard problem.
ReplyDeleteMadeleine, cogsci is not about interpreting (hermeneutics) but explaining. See replies in 10a and 10b.
DeleteFrom my understanding, the main frustration comes from the idea that "solving" a problem means showing how things cause other things. However, when it comes to feelings, the regular ways of explaining cause and effect seem to fall short. The discussion hints at the difficulty of breaking down consciousness into observable actions or brain structures, likening it to taking out a metaphorical "widget." Emphasis was put on how the standard causal mechanisms can't quite bridge the gap between non-feeling and feelings, making the Hard Problem "insoluble", though it exists. (we cannot know what's in the black hole, yet it doesn't mean the black hole is not there).
DeleteKristie, kid-sib says metaphors and enigmas don't answer the simple question: how and why does the brain produce feeling rather than just doing what needs to be done to survive and succeed? How and why is feeling not causally superfluous (even though it does not feel superfluous)?
DeleteMiriam, that sounds right, but it's much too close to my own words. Make sure you explain things in your own words, and address them to kid-sib.
DeleteBTW, when you touch a stove, you actually feel the pain after you've reflexively pulled back your hand. That (along with the Libet experiments on the timing of willing and executing voluntary movement [what do they show?]) makes the HP even harder. [Why?]
Fiona and Tina, predicting whether and what people are thinking, based on neural/behavioral correlations (including verbal ones) is a part of cogsci (OMP), but OMP is part of cogsci's much bigger mandate to explain how and why people (and other organisms) can DO what they can do (EP) (of which learning, categorizing, and language are the main components). But the HP is neither of these. and part of cogsci's mandate too. Is it solvable?
ReplyDeleteSelin, feelings are not "processed", they are felt -- sometimes as a reaction to an external input or in response to an endogenous stimulus or state. "Function" is somewhat vague; but no one would suggest that the mapping between feelings and doings is one-to-one.
ReplyDeleteProfessor Harnad’s response to Dennett’s argument and methodology clarified and summarised the reactions I had to Dennett’s argument in the 10a reading this week. As mentioned in previous skywritings, I struggled with Dennett’s claim that an individual’s subjective experience (is this equivalent to saying ‘what it feels like to be them?’) can be causally explained by neural and physiological correlates (which reminded me of Fodor’s 4b article). Further, through Harnad’s responses I was struck by how often Dennett seemed to be aiming to answer the easy problem, rather than the hard problem—as Dennett’s heterophenomenology seems to only be able to answer what processes are correlated with a feeling/felt state, but not why these feelings occur at all. Finally, I found Harnad’s observation that what must be solved by any answer to the hard problem is the “the gap. . .between non-feeling and feeling, not between this-feeling and that-feeling” was helpful to guide and orient my own thinking towards what exactly the hard problem asks, and what a solution to this problem would need to consider.
ReplyDeleteHi Shona! I felt the same way about these readings. I was also reminded of our discussion of mirror neurons, and the futility of relying on brain structure alone in answering questions of consciousness. Also, Professor Harnad will probably have a more informed answer but to me "subjective experience" seems apt for what you're trying to describe.
DeleteLillian, let me count the ways "subjective experience" is a W-W: Can a feeling be anything other than "subjective" and "experienced"? That's what de-weaseling makes obvious: "subject experience" is just "felt feeling."
DeleteFrom my perspective, HP may never be fully solvable because it involves subjective experiences. In contrast, we still have the chance to solve EP, as the DOING something is observable. Especially it is questionable whether feeling is intelligible for human being. While neuroscience would be considered as the way to solve HP, the problem should be underdetermination of theory and mind-framing within our knowledge. For some reason, I think HP is more like a philosophical problem, which is too abstract to find the solving way.
ReplyDeleteEvelyn, I felt the same way about HP, it is not very likely that HP is completely solvable. Therefore, instead of thinking about how to solve the HP problem (it feels like something that is unique to the individual that can't be perceived or tested by others, and is therefore the least likely to be solved), maybe HP is better viewed as something like an infinity in math, where we can only get close enough to it to make some assumptions about it, but can't really get to it to test it in terms of its magnitude and specifics.
DeleteI found the concept of zombies in this reading quite intriguing. David Chalmers utilizes the concept of philosophical zombies in consciousness studies to illustrate the "hard problem" of consciousness, which concerns understanding subjective experience. Zombies, in his thought experiment, are beings indistinguishable from humans in behavior and physical makeup but lack conscious experience. Chalmers argues that the mere possibility of zombies suggests that physical and functional processes cannot fully explain consciousness. This highlights a fundamental gap in physicalist theories of mind. By using zombies, Chalmers underscores that conscious experience, or qualia, has unique properties that go beyond physical explanations, advocating for new approaches to understand consciousness. I have a hard time seeing this as a very authoritative argument… The possibility of conceiving something does directly translate to it being a possibility.
ReplyDeleteAimee, you're right that the philosopher's armchair conjectures ("maybe there could be 'flying spaghetti monsters' " was Richard Dawkins's spoof on this kind of thinking, in religion, superstition, and the supernatural) are not a very compelling way of thinking. Chalmers should have said that an explanation of how and why there cannot be "zombies" would be a solution to the HP.
DeleteBut please re-write your comment replacing all the W-Ws by the f/f/f words to see how vacuous, distracting and redundant the W-Ws are, and how they only clutter our thinking.
I’ve read Eugene’s post and the professor’s reply to him, however, I’m still confused about the relation between feeling and “belief” (weasel word, so replaced with “feels-like.”). How do “feeling” and “feels-like,” then, relate? Would it be correct to think of it like “taste” and “tastes-like,” where the latter is an experience of the former?
ReplyDeleteIn that case, would it be appropriate to draw an analogy as follows:
Suppose Person A is someone whose sweet taste receptors have been temporarily disabled.
Suppose Person A is given a sugarcube. Because (1), it doesn’t “taste-like” sweet to him.
Then, would it be correct to understand Dennett’s question be something along the lines of “If the sugarcube doesn’t “taste-like” sweet to Person A, would that be an experience of tastes-like at all?”
And, would Harnad’s response to that be that we’re focusing on the wrong issue - if we taste sweetness, we taste it, if we don’t taste it, we don’t taste it. We want to know HOW and WHY we taste at all?
Would this analogy be a correct interpretation?
My understanding is that the difference between feeling and belief is a little like asking the difference between fingers and thumbs. A belief is a feeling (though we certainly feel more things than beliefs). It's a felt state, we can rephrase I believe xyz as I feel that xyz is true. Dennett is saying that if we can reverse engineer, to use your example, taste then there is no reason to believe we don't also feel that sweetness. I think you are generally right about what Prof. Harnad would say. Though he agrees that feeling would co-exist with function the hard problem is showing how it is that feeling is generated through this process and why, if we can get function without it, is it needed in the first place.
DeleteOhrie and Marie, I feel that you both get it,
DeleteThis reading greatly clarified some confusion I had with Dennett’s argument in the previous paper. I had been confused about his heterophenomenological approach to the Hard Problem, namely its reliance on introspection and internal markers in studying feelings. Introspection has been pretty widely critiqued as a reliable method of studying aspects of psychology, and I was confused throughout Dennett’s paper as to how he thought it would solve the Hard Problem. To me it seemed similar to asking someone about what specific neurons are active in their brain at a given time, or to provide their blood iron levels, for example. These biological factors are important to their output, however not identifiable through internal measures alone. As pointed out by many others in this weeks thread, this is because Dennett was oriented towards the Easy Problem, and claiming to solve the HP.
ReplyDeleteLillian, DD thinks there is only the EP. No HP.
DeleteI must admit, this reading left me completely hopeless in regards of the hard problem ever being solved. Especially after reading the excerpt: “Suppose you have a successful causal mechanism for predicting feelings, including all the functional conditions and states in which they occur, right down to the last reportable JND, in every conceivable situation. You will still not be able to give even a hint of a hint as to how it is that mechanism feels at all. . .nor of why . . . it feels”. So, my question is, are we doomed? In a previous reply (to another dejected student), prof Harnad wrote that HP is not rhetorical nor a curse, and it is still an empirical problem. But I am stumped as to how it can be so, if every empirical explanation for it has so far hit a dead end? As well, how can we ever escape the OMP? How far can we go until we know, for certain, whether HP is in fact unsolvable?
ReplyDeleteI commented on this idea in my post for 10A because I’m also entertaining the possibility that the Hard Problem is actually just the Impossible Problem. You comment that “every empirical explanation for it has so far hit a dead end”, and that is very true. We see in this very paper, through Professor’s commentary on Dennett’s arguments, that the heterophenomenological approach is one such explanation at a dead end. In summary, simply mapping brain activity and functional organization, and then linking that to feelings and felt states, is not enough to explain why the feelings are felt at all. Why can’t we just DO without the FEEL? That is the question, which is so far unanswered, and will remain unanswered even if we solve the EP. This is because, supposing we solve the EP with a mechanism X, we can remove an element from X and see a demonstrable effect on X’s ability to do a specific function Y. However, supposing we remove an element from X and see no change in any function Y, aside from a loss of feeling, then we are still no closer to explaining why that element gives rise to feeling. So, clearly, the heterophenomenological method does not suffice, and neither does solving the EP. Back to the drawing board. Well, that’s the scientific method at work! Just because one, or ten, or one hundred hypotheses or empirical methods fail to solve an empirical problem does not guarantee that the problem cannot be solved at all, that the solution is not somewhere out there. Just because the solution to the HP and the empirical method to go about finding the solution have thus far eluded us, and just because the HP may be impossible, does not mean that its impossibility is a guarantee or a certainty. This is precisely why the HP is still an empirical problem. So, ultimately, no reason for concern (it took 100 years for evidence to emerge for Einstein’s gravitational waves, and ~350 years to prove Fermat’s Last Theorem!). I should say, more accurately, that there is no reason for concern…yet. I say yet, because I want to answer your question of “how far can we go until we know, for certain, whether HP is in fact unsolvable?”. The intimidating fact here is that, if the HP is unsolvable, we will NEVER know. It would take too long to explain why this is true, but if you’re interested to know the reasoning behind it, take a look at Godel’s Incompleteness Theorems and the notion of Undecidable Problems raised by Turing. In brief, no system of math or science can ever prove every true fact about the universe (incompleteness), and we can never know from the outset whether a given theory is provable or not (undecidability). If the HP is unsolvable, we will forever flounder about in the dark vainly looking for a solution because (1) our systems are incomplete and (2) we can never know if it is provable or not until it is proved, and in this hypothetical it cannot be proved.
DeleteI was hoping to build off what the both of you have said and comment that maybe the pure empirical study of consciousness itself is why we continuously hit dead ends. It is inherently impossible to observe something that isn’t material. And unfortunately for the hard problem, it seems to deal with the immaterial. Whether or not the Hard Problem will ever be solved is a question of time like Stevan brings up. Currently, we don’t have the technology or just general methods to even begin to reverse engineer something that we can’t even be sure everyone has (OMP). For now I believe that it is a fruitless endeavour to continuously try to solve something through empirical means, when that thing is abstract and immaterial.
DeleteHi Stevan, I think we have to be careful about applying Godel's Incompleteness Theorem (GIT) here. In class we learned of things that we can be certain of, which exhaustively include: a subset of arithmetic if we accept Peano's axioms, and that I am feeling therefore something is felt. If you think about it, we cannot be certain about anything else, so for that reason alone everything other than those two things are unprovable. In terms of Incompleteness, GIT shows that in any consistent formal axiom system in which arithmetic can be applied there must be statements of the language that can neither be proved nor disproved. This is why we can only be certain of *some* of arithmetic. This doesn't necessarily tell us much about the HP's provability, because again, nothing except those two certainties are really provable. But unprovable does not mean "unsolvable" in the way that we normally think of "solutions" in the sciences. There are many things that we say we "know" about the world thanks to the sciences and the HP may very well turn out in the future to be something that we understand to the extent that we understand, say, quantum physics, or maybe to an even greater extent.
DeleteEven though I agree that the Hard Problem feels like entirely unsolvable or a dead end, I feel like if reverse-engineering a TT in terms of the Easy Problem (how and why of DOING capacity), is ever achieved, there is not much we can do in terms of the Hard Problem. If a successful TT can do everything we can do, we might just accept that it feels as well since it entirely indistinguishable than us. We can never know for sure, due to OMP, but if we know we feel, then there is nothing much to do than accept that an indistinguishable TT would be feeling as well.
DeleteThis paper points out fatal flaws in Dennet’s hetero phenomenological method. Harvard critiques that Dennett does not fundamentally understand the hard problem at all. He points out how Dennett’s method does not address how and why, which would express the actual causality. Additionally, it only really discusses the concept of feelings in relation to the topic. While the idea of different ways of feelings is interesting, it derails from the question at hand: how and why feelings exist. This reading at its heart is that it is not an issue to describe feelings, we are having issues explaining them and figuring out how they work. This critique has reminded me of the necessity to take away weasel words when describing concepts, such as feeling, to unveil the true meaning of what we are discussing, especially when it comes to feeling in relation to cognition.
ReplyDeleteI think that trying to sort feelings into different types, like Dennett suggests, isn't really helping us figure out how or why they exist, as Harnad explains with the example about how whether or not the toothache was real, it still feels like a toothache and we are only interested in how/why it feels like anything at all. Harnad's argument against this part of Dennett's idea reminds me of why we don't need to worry about individual differences in behavior when we're trying to understand how things work. We just want to understand the basic principles that make them work.
ReplyDeleteEvelyn, surely feeling is the most concrete, least abstract thing of all! And also the thing we are the surest about (when we are feeling): Even if I can't describe what something feels like, I can feel it! That we can't explain how ir why we feel does not mean we don't know exactly what it feels like...
ReplyDeleteThis paper on Dennett demonstrates a gap in Dennett’s approach on the Hard Problem. Dennett's reliance on heterophenomenology may illuminate parts of the Easy Problem, but this approach ignores the how and why feelings arise at all–bypassing the core of the Hard Problem. Moreover, there are limitations of the A and B team framework. Dennett's reduction of the debate to two sides shouldn’t be confined to a binary. It's imperative to embrace diversities of thought, acknowledging that the journey to understand consciousness involves exploring multiple, nuanced perspectives, not just two or groups of people who possibly all agree with each other. I also appreciate the explanations and concerns of using words like ‘certain experiences’, consciousness, and beliefs, which are often too vague and are thrown around to make problems sound ‘complex’ to be explained.
ReplyDeleteBoth the unpublished article and the video from Professor Harnad mentioned the “why question” from Darwin, which is described by Mr. Harnad as being an adaptive question. This reflects on the role that consciousness has to play in every single one of our behaviors, emotions, thoughts and refers directly to the hard problem. This distinction between the how and why made me think that neurosciences are only scratching the surface of consciousness, and as mentioned at the beginning of the video by our professor, we might never be able to answer this why question. But even further, by reverse-engineering our brain we should be able to understand how an input can lead to a certain output, so answering the how question of Turing, but can we even just think about a solution, an experiment or anything else to explain the “why”?
ReplyDeleteHarnad's response to Dennett was clear and easily understood, which immediately contrasts it to Dennett's paper. While reading Dennett's "the fantasy of first-person science" I felt confused and bombarded with distracting and frankly irrelevant information(eg. blindsight, qualia etc.). In Harnad's response by continuously bringing it back to the question of how/why we feel, I felt the hard problem(and why it is so hard) was clarified. Importantly, the response highlighted that functional correlations are not the same as understanding the causal mechanism. Yes, Just Noticeable Differences can be studied to no end but psychophysics won't give us any explanations for why we feel, as opposed to just incorporating necessary sensory information into our actions. I also particularly resonated with what Harnad said about the zombie hunch. To me it is hard to imagine something functionally equivalent to a human but non-feeling(I honestly think this is impossible), and equally to have something molecularly identical that doesn't feel doesn't make sense(I think we can agree that it is something in our make-up that allows us to feel!).
ReplyDeleteOne thing I was curious about was how there can be different ways of addressing our own felt states. Felt states will always feel like something, however based on how we explicitly label (or dont label) our own state, what we do in it or why we do something may be affected; for instance in different types of mindfulness meditations, people learn to pay attention to what they are feeling, particularly when they are doing different things. That may then allow them to gain insight into what it is specifically in their actions or environment that is causally related to making them feel some state. They can then learn to purposefully enter a more pleasant feeling state, or avoid an unpleasant one with things they do or environments they surround themselves with. Could such mindfulness methods help us understand how feeling things is necessary for doing certain things? Or is this still only an understanding of the output of a brain process and not an understanding of the process itself?
ReplyDeleteHarnad comments on Dennett’s approach and how he fails to address the issue of explaining how and why we feel. Unlike Dennett, he believes that there is no way to account for feelings functionally. Even if we understood and were able to explain all the neurological processes and relate it to what is felt, it still wouldn’t solve what we call the hard problem of consciousness, which Dennett denies exists.
ReplyDeleteI wanted to start this discussion post by commenting on Harnad’s breakdown of studying cognition and the hard problem (although Dennett doesn’t believe in the hard problem) in the 1st vs 3rd person. Harnad essentially suggests that the 1st and 3rd person perspective on studying feeling are interchangeable terms (what he also calls weasel words above) for feeling. This is because even attempting to study 3rd person correlates of feeling cannot sufficiently explain what the initial person (the 1st person) is feeling. Using objective 3rd person data like heart rate, hormone changes etc is just non-introspected first-person data. Also, these 3rd person correlates are integrated into our feeling and experience of the world whether we acknowledge it or not. As pointed out by others, Dennet doesn’t address the how or why we feel (he tries to doge the question altogether), but even his methodology cannot fully separate itself from our subjective feelings.
ReplyDeleteJoann, you got most of that right, but "physicalism" is a W-W. Of course everything is physical: What's the alternative? Metaphysical? Supernatural? The question is just causal: How and Why. Not what kind of "stuff" is "it" made out of.
ReplyDeleteWe know that the brain (not the moon or the stars) must cause feeling: The HP is explaining how and why. Brain correlates are surely also causes: but HOW? and WHY?
The HP remains cogsci's problem, not ontology's. And it remains unsolved. (So does feeling not matter? Tune in in Week 11, where we will find out that not only does it matter, but it is the only thing that matters.)
Rosalie, it still sounds like you're talking about explaining why we feel what we are feeling rather than (how and) why any sentient organism feels anything at all... And HP is not exactly looking for an algorithm but a causal explanation.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteProf Harnad points out the obvious correlational problem with Dennett's heterophenomenology where a complete correlational account of feelings is totally possibly and imaginable yet is still obviously missing a causational account, and not just trivially in a sort of Hume-like way but with a lack of substantive explanatory power for how and why we feel: """We feel because this collection of neurons sends this pattern of signals"" … why does THAT cause feeling.
ReplyDeleteI found the part where Dennett says Zombies will still be able to report their beliefs and will presumably do so sincerely, and then Prof Harnad points out that attributing sincerity of beliefs to an unfeeling zombie is equivelent to attributing sincerity to a book, to be very clarifying as previously I didn’t catch Dennett on that. But yes, the weasel word that is ‘believe’ is can be unpacked into ‘the feeling of believing’ and so when we say a zombie ‘sincerely believes’ we are saying that ‘it really has the feeling of believing’ right after we asserted that the zombie is unfeeling! → a nonsensical contridiction. A zombie would report its “belief” but not sincerely, the concept of sincereity is dependent on an assumption of feeling which we have just assumed the contrary.
Equivocation is a synonym for weasel word, with a more official connotation.
“I believe in the Turing Test, though I recognize that neither Turing Equivalence nor Strong Equivalence is as firmly founded as molecular/functional identity when it comes to the probability of feelings.” : Is all that this says something like, we know well the laws of physics and causality but we are not as certain in a principle of Strong Equivalence and so we cannot be certain of feeling capacity with T4+ ? (and not from a trivial place of skepticism of course). This sentence confused me a little bit is all.
This comment has been removed by the author.
ReplyDeleteAlthough Harnad is against the invention of Zombie doppelgangers (because whether they are possible is irrelevant to the HP), I found them useful in clarifying my own understanding of what we should focus on (i.e., a how/why explanation of feelings) and why that is. I admit that I have been bogged down by questions on whether zombies are possible; this is because I assumed that in order to do the things we do, feelings must arise. The widget explanation helped in reminding me that even a successful causal mechanism for predicting feelings does not bring us closer to how and why it is that it feels like anything at all (rather than just doing the things we do), further emphasising how hard the hard problem is.
ReplyDeleteWhen explaining why he thinks there needs to be a C team, Harnad states that he does not think “that there could be a functionally equivalent but molecularly different pair of entities, and one of them feels and the other doesn't.” Does this presuppose that to be the feelings are part of the entity’s function, and therefore to be functionally equivalent they must either both feel or not feel? On a different train of thought, isn’t it possible to have a functionally equivalent but molecularly different pair of entities due to underdetermination?
Professor, you have made it clear both in your writing and in today's class that "feeling" is the word that we should use to describe the culprit of the HP. I am curious, though, as in our last class I interpreted you saying that there is no utility in further delineating different kinds of conscious states (felt states). This brought up a question for me. Taking today’s example of reaction-time to pain (fire); is there no utility in supposing, primarily from an evolutionary perspective, that the feeling of pain and the subsequent reaction is a different kind of feeling than the feeling of, say, love or wonder? I mean kind here as a more of a type I suppose, as obviously every feeling is a different kind of feeling to another, but it may be a different type in the sense that it seems much more plausible to see the “why” in relation to feeling certain psychosomatic states (like being burnt) than others. Maybe I’m missing the point, and I understand it is all just feeling, but I suppose it felt reductionist at first glance to group these two seemingly very different (both causally and experientially) states together and dismiss a distinction.
ReplyDeleteThe article 'Scientists discover the on-off switch for human consciousness deep within the brain' describes the findings of the ‘claustrum’ which is a region in the brain that supposedly activates/deactivates our consciousness. It mentions that when it is stimulated, the body loses its ability to do normal things and basically ‘zones out’ but returns back to normal when the stimulation is removed. However, it seems like the way they define consciousness is more of the state of being awake and being able to do things, they don’t really touch on the feeling.
ReplyDeleteProfessor Harnad explains why Dennett’s paper is not a suitable response to the hard problem, as heterophenomenology still lacks a causal mechanism for feeling — the crucial aspect we should be addressing.
ReplyDeleteI found Professor Harnad’s view on the subconscious interesting; it’s usually such a mystified aspect in popular psychology. However, his comment trivializes it; if a feeling is deemed 'unconscious,' meaning it was not consciously felt, then it was not felt at all. Consequently, it can’t be considered a feeling. Phenomena like semantic priming or blindsight would be explained as purely physiological mechanisms, without the need for a subconscious intermediate — cerebral mechanisms carrying out their functions without the need for feelings. In blindsight, for instance, visual input bypasses damaged visual cortex to subcortical areas, resulting in appropriate actions without the need for it to 'feel like' something to see. In the case of semantic priming, the memory of having seen a word affects subsequent performance for another word through activation patterns in the brain and connection strengths without it “feeling like” something to remember that word. Would we thus characterize them as vegetative functions?
I enjoy these summaries/ annotations as they help put things into perspective. As Professor Harnad says, this does not provide us with any information on why/how we feel. Dan Dennett rather focuses on the prediction/correlation of feelings. However, even if we can decipher the causal mechanism for feelings, we still do not know how or why, making heterophenomenology somewhat pointless…
ReplyDeleteHey Maria, I wouldn't call it pointless but you are right in that heterophenomenology doesn't answer why and how we feel. It is important to understand what feeling is. I think that's what's creating very polarized opinions. The feeling action that Dennett is describing, I think, is more on the simple side. Feelings that are basically reactions to stimuli. Hence the extremely scientific explanations that we learn in neuro courses. The feelings that we discuss in this course are deeper. It's more meta (I'm not sure if that's the right word to use but it should be fine hopefully). Like feeling the feelings. I hope that makes sense.
DeleteResearchers at George Washington University, led by Mohamad Koubeissi, have discovered that electrical stimulation of the claustrum, a small brain region, can reversibly disrupt consciousness in humans. This finding was observed in a patient with epilepsy, where stimulation of the claustrum caused her to lose consciousness and stop her activities, resuming them with no memory of the lapse once stimulation ceased. This discovery suggests the claustrum may play a crucial role in integrating sensory and cognitive experiences into a cohesive conscious experience. However, the study's limitations, including its focus on a single patient with an abnormal brain due to epilepsy and hippocampus removal, mean more research is necessary. This finding has significant implications for understanding consciousness in humans and other mammals, potentially influencing the study of anesthesia, comas, and the consciousness of other animals. The relationship between consciousness and wakefulness is also brought into question, as the patient remained awake despite losing consciousness.
ReplyDeleteHarnad's argument about consciousness really makes you think about the big question, the 'Hard Problem', in studying the mind. He's critical of Dennett's approach, which tries to understand feelings by looking at physical processes. But science hasn't quite figured out the 'why' behind our feelings. As I try to make sense of thinkers like Chalmers and Dennett, I can't help but beg the question: Can we really get to the heart of our feelings using logic and scientific methods alone, or is there something more to it, something that science can't quite capture yet?
ReplyDeleteEven though science can (to some extent) map and predict feelings, Professor Harnad is right in implying that it struggles to explain their intrinsic nature (the how and the why). This suggests that there is more to feelings than what current scientific methods can capture, maybe computational and dynamic approaches are not enough for now. Do we have good reasons to think that there are approaches that will ever be enough to reverse engineer feeling? Is the HP solvable at all?
DeleteThe text gets deep into the tricky stuff of the "hard problem" of understanding consciousness. It's all about this subtle dance between feelings and functions, captured in the "feeling/function" puzzle. Basically, it's asking why and how certain brain activities come with feelings. They throw in an analogy with gravity to show the difference: in gravity, pulling is just what it does, but in consciousness, it's a head-scratcher figuring out why specific brain jobs always bring along feelings.
ReplyDeleteSo, unlike gravity, where pulling is just part of the deal, consciousness throws us a ball. The text is all about wrapping our heads around why some brain tasks in consciousness always bring along feelings.
This might lead to further implications that the explanation and discussion of the mind-body problem, which is based on the separation between feeling and doing, leads to philosophical pondering about consciousness and how subjective experiences are relevant to the observable mechanisms.
I think the difficulty of Zombies was illustrated very well by Harnad when talking about the claustrum. While some consider the claustrum to be an on-off switch for consciousness, Harnad clarifies that it really cuts off awakeness. For there to be an on-off switch for consciousness alone, it would turn us into Zombies. Flicking this switch would let us continue to behave normally, but there would no longer be anything it is like to be us. This really drives home for me Harnad’s assertion that the hard problem involves explaining why we are not Zombies - an explanation that Chalmers continually side steps. If we were to find a true on-off switch for consciousness (in the sense that Harnad makes clear), we would in turn be able to solve the hard problem, because we could determine how and why feeling could dissociate from action. Interestingly, with the consciousness switch either off or on, the heterophenomenologist would be none the wiser, another reason why heterophenomenology is an incomplete approach to understanding consciousness.
ReplyDeleteI like how you put that heterophenomenology is just another "descriptive device". I think there was an interesting bit in this paper about where Dennett tried to show the power of heterophenomenology, by showing how it could be used to predict the motion capture phenomenon, in the same way that chemists can make predictions on the basis of their models. I think Harnad's response to this is very apt - all heterophenomenology is capable of producing is a "prediction of functions from functions", rather than predicting and explaining the connection between feelings and functions. In other words, here too, we see that even with some predictive power, heterophenomenology remains a descriptive device, never explaining the correlation between feeling and function...
ReplyDeleteIn this text, Harnad analyses and dissects Dennett’s paper. Harnad explains that throughout Dennett’s entire text, he fails to acknowledge the how and why questions related to feeling, which is the hard problem and remains unsolved. I enjoyed when Harnad said:
ReplyDelete“Unlike you, Dan, I stand ready to admit that neither I nor anyone else has even a clue-of-a-clue about how one could cash in that "somehow" functionally. Hand-waving -- emergence, giant cooperative entities consisting of dumb homunculi that "add up" to feeling agents "somehow" -- just won't cut it.”
I think this is important because there is nothing wrong with admitting to not having a solution, and trying to brush it off by saying that it isn't really a problem at all only brings us farther from a solution altogether.
I am confused about feelings, as discussed in the paper. In my perspective, feelings are complex and comprise emotional and cognitive states. I doubt whether it is possible to truly understand or replicate feelings, even if we were to create a biologically identical entity like molecule-molecule Robot to solve T5 TT; it may not be able to address the 'Hard problem' of cog sci.
ReplyDelete
DeleteThe robot may have behaviors associated with emotions or feelings, like crying or shaking, that could occur without the experience of those emotions. I then could use a feeling version of the feeling Room Argument (FRA); a robot is programmed to display similar behaviors to react to something as a t2 and above level suggests but did not have an equivalent mental state. It implies that an entity might exhibit behaviors of feeling without the inner subjective experience, leading to the conclusion that the Turing Test might not be sufficient to answer the hard problem.
Hi Jiajun,
DeleteYour perspective is quite interesting, i would like to add and modify a few bits though! Now I believe that what feeling truly consists of is rooted within the HP, we don't know how we feel, and if others even feel (OMP). It helped me understand that feelings are not limited to emotional and cognitive states. Yes we can feel emotional pain that, and yes we experience cognitive states, but I would also argue that we can feel physical pain, we can sense our environment around us and we can feel other people's pain (sort of like mind-reading, recall why the OMP doesn't impact us much). I think to further elaborate on your T5 argument, I also believe that even if we create a robot that is indistinguishable in every form, I don't think it can address the hard problem since we haven't ourselves! I think the subjective nature of feeling hinders us from developing an objective method to observe what people feel, and what other species feel!
Hi Jiajun and Mallak, I would like to add on to your points. While it is true that the T3, T4, and T5 would primarily help us answer the EP (behaviour is easily observable), I believe they would still offer some insight into the HP. They would show us if behaviour leads to feeling. If a Turing machine can be a zombie, this means feeling is not essential. If it cannot, this means feeling is essential to behaviour, and hence we have part of the answer to why we feel. This is of course limited as we would have to find a way to tell if the machine is feeling (OMP), but otherwise it can offer some insight into the HP.
DeleteHi Nicole,
DeleteSpot on! I wanted to touch upon that too, but I wanted to read a bit further before I assured anything :) I think it would be beneficial if we could develop sort of like an effective protocol to research, analyze and understand our feelings before we can ever start solving the HP and develop a FULLY autonomous and functioning T5 robot. We need to first find a way to know what and how and why we feel objectively without relying on introspection, so that we can begin to address the OMP, it is sort of why we don't truly, with certainty, know if fish can feel pain or not, it is subjective in nature. I think once we achieve that, we could employ the T5 to help us gain some insight into the HP too, if we know whether feeling is essential or not for behavior, maybe that could help us observe the mechanism by which we feel, so it would be easier to address the HP hands on, but I still believe we are wayyy far from coming close to this possibility, not until we can objectively understand how other species feel first!
One notion that made me understand a bit better the problem that was at stake in the reading was the problem of the Zombic Hunch. While Chalmers argues that the Zombie believes it has feelings when it doesn’t, Harnad argues that we aren’t concerned with how the Zombie has false beliefs (although it does not have any felt state at all, including beliefs). Rather, IF a Zombie, identical in function and organization, were to exist, how come it doesn’t feel and why doesn’t it feel while we do. Chalmer, by thinking that Turing answered his question about veracity of beliefs, completely sets aside the issue of feelings, thereby just addressing the EP. A robot that would be able to do everything that we are able to do would pass the Turing test, reflecting our success in generating thoughts. But these thoughts wouldn’t be felt thoughts: disregard of the HP.
ReplyDeleteThis paper made me think about many things we’ve discussed this semester from a different angle. First, Harnad presents the 3 questions established by Dennett as 1 that boils down to the hard problem.
ReplyDeleteI also had a similar question as Harnard presents after reading the last skyreading, which was why/how would a Zombie molecularly identical to us not have the same thoughts and feelings, as in what could possibly then explain the difference? His argument is simply not sound.
Also, and I’m not sure this was the goal, but it made me rethink how separate the EP and the HP really are: how and why do we think versus how and why do we feel? Inevitably as humans our feelings influence our thoughts and vice versa. In order to reverse-engineer our functional thoughts would you also have to reverse-engineer feelings?
This comment has been removed by the author.
ReplyDeleteThis reading begins with presenting the mind body problem and a few philosophers (descartes, kant, and turing) who have given interesting questions to prod at this problem. Harnad points out the difference between thoughts and feelings, thoughts can be just input output from going ons of the head but what about felt thoughts, or things we call feelings? Those in camp A believe that the hard problem has already been solved, the other is B which believes that there may be cognitive zombies who are compositionally identical to us but just do not feel. Harnad wants his own camp which is focused on our ability to describe feelings, explain why and how we feel and are not zombies and the gap between this feeling and our just noticeable differences (JNDs)
ReplyDeleteReading Professor Harnad’s paper made me realize that we have categorized ‘feeling’ into many different types. Emotions–sadness, anger, frustration, happiness–are certain kinds of feelings. The sensations that we detect–seeing the color blue–is a feeling, and feeling pain is a hybrid of emotion and sensation (I never thought about it this way): we detect with our receptors a stimulus that surpasses the threshold for pain, causing us to respond with a negative emotion, such as distress. Or desire-states and feeling/ knowing-states are just reducible to simply, feeling. I think that since we have a lot of feelings, we must categorize them into different types so that we can navigate in the world–adapt to the environment, communicate with others, learning and memory. But in terms of the hard problem, it encompasses all feeling–why and how we feel them, no matter the type.
ReplyDeleteI have a question about a specific part of the paper; Professor Harnad said: “Consciousness, being half-epistemic, like thought, is equivocal.” I do not understand what “half-epistemic” means in terms of ‘consciousness,’ which is just feelings. This was his response to Dennet saying that “David Chalmers is the captain of the B team,...He insists that he just knows that the A team leaves out consciousness.” Was consciousness described by the A team to be consisting of half-thought/knowledge (aboutness)–and hence the “half-epistemic” description? How does it correlate with the fact that thoughts are equivocal (things happening in the brain when given an input to produce an output)?
ReplyDeleteProfessor Harnad's paper made it much easier to understand how heterophenomology is not sufficient to answer any part of the Hard Problem. Heterophenomology seems to be stuck at viewing feelings as “beliefs about feelings”, but this doesn’t take us any closer to explaining the how and why of felt states as Professor Harnad emphasizes in the reading. The Hard Problem will always exist due to the OMP, neural or physiological correlates of felt states do not provide valuable information about how and why we are sentient (we feel). Dennett appears to be talking about the Easy problem when discussing his third-person approach, however the causal mechanisms underlying how and why we feel (not the same as believing that we feel) is still left unsolved (and seems like it will always be).
ReplyDeleteIn this reading, Professor Harnad argues that Dennett and his heterophenomenology do not address the necessary questions in answering the Hard Problem (how and why we are able to feel). He claims that heterophenomenology only concerns easy questions like how good experimenters can be at mind-reading, or how good subjects can be at mind-describing. It can, using correlations, predict whether or what someone is feeling, but it is still unable to answer how/why.
ReplyDeleteHarnad, in this paper, argues against Dennet, as Dennet essentially denies the hard problem of cognitive science and advocates for a method of scientific investigation that looks at things from the third person. This is tricky because, sure we can study feelings and gather data points, but at the end of the day, the feeler is the one feeling. Dennet’s approach’s contingencies were described as “how good experimenters/theoreticians can be at mind-reading, how good subjects can be at mind-describing,” but that still doesn’t explain the hard problem of cogsci—which we are well privy to— how and why one feels anything.
ReplyDeleteIn my response to Dennett’s article, I stated that I was leaning towards team A. I still am but I’m a bit more skeptical of it now because I misinterpreted what the teams arguments were about. Reading Professor Harnad’s response was definitely fun and engaging because I felt like I was Dennett since I agree with his heterophenomenology stance. I still do but like I said I have some doubts because I have a bit of pride in being human. What Dennett describes makes it sound like we are robots who can think freely. Feelings are just responses to chemicals in our body being released due to a stimulus. Sentience, feeling those feelings are deeper than that is what I’m understanding from this article. As much as I want to agree because I want to feel somewhat more complex of a being, I can’t. It doesn’t sit right with my logic. Maybe I don’t fully understand consciousness (to be fair who does), to me feelings are just responses and feeling those feelings are also responses to bodily functions. Would that solve the hard problem or render it unsolvable? Please don’t hate me.
ReplyDeleteI see zombies as dead beings still able to function. An analogy for this is how frog legs still move when salted shortly after their death if the nerves are still intact. The system remains functional and it looks like it is alive, but the movement is not felt by the frog.
ReplyDeleteThough I tend to find Dennett’s ideas quite similar to my own, and to justify them well, I appreciated this critique by Harnad. The insistence on using clear, non-weasel words really makes it clear that Dennett is equivocating different the correlates of feeling, with the actual feeling. I mirror Dennett in thinking that we will likely have, in the future, a scientific, complete, and causal and mechanical explanation for feelings. However, I admit that the claim of this belief should be, in no way, convincing to someone who does not share it. The intuition, to me, comes from both the fact that computation is incredibly general and powerful, and that physics must be implementing our wetware in a framework that is largely analogous to computation. A mistake Dennett makes, that Harnad points out, is that the existence of this intuition that a solution exists is not a reason to claim we have found the definitive method for researching this phenomenon; in fact the hard problem seems to be quite beyond the reach of modern scientific inquiry, and correlates of consciousness seem to miss the ball entirely. It's a worthwhile area of research, most likely, but seemingly not for the reasons Dennett defends?
ReplyDelete(*I originally wrote this for Johann's comment, but I cannot seem to find it anymore, so I'll post it here)
ReplyDeleteHi! I don’t think Cognitive Science “collapses” everything to feeling because feeling is the only thing that the Hard Problem of Cognitive Science is about. I do think the case where some people are not able to feel certain feelings but are still able to somehow react to it is interesting because it might give us insights to how we can know unfelt feelings. However, “unfelt feelings” are still feelings, and it is still a target of the Hard Problem.
After reading Harnad's criticism of Dennett, I realized that the crucial question of the hard problem was never adequately answered. Furthermore, I understood that Chalmers' zombie could be conceptualized using the Turing test. Essentially, Chalmers' zombie would be equivalent to at least a T3 robot. To rephrase Dennett's objection, he would then fail to see how a T3 robot does not have qualia. However, as seen with Turing's argument, it is not evident that this would be the case, and that is one of the crucial questions we aim to answer.
ReplyDelete