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"
I really enjoyed how Professor Harnad’s paper broke down Dennett’s arguments as it allowed me to really see the bones of Dennett’s argument and why I had hesitations upon reading it. Harnad’s point is that heterophenomenology isn’t going to provide us with the way to solve the hard problem of how and why we feel. Heterophenomenology is only describing feelings. Further, the problem with the “conscious but unaware” argument is a bit clearer. The issue at hand is that we are trying to solve the hard problem… the how/why of feeling. Thus, if this is the problem we are trying to solve, then a conscious experience that you are unaware of is not relevant; you are not feeling anything. Ultimately, Dennett uses this example to support heterophenomenology, but it is not at all applicable.
ReplyDeleteDe-weaselling weasel-words:
DeleteTo be "conscious" of something means to feel something.
And to "experience" something also means to feel something.
So just as an "unfelt feeling" makes no sense, an "unconscious experience" makes no sense.
An unfelt experience is not an experience, it's just an event.
(Maybe zombies are the only ones that have them...)
There are lots and lots of zombies, by the way. They include rocks, rockets, (toy) robots, and rutabagas (I hope, otherwise there would be nothing left for vegans to eat!). As to Eric, the jury is still out (because of the OMP) but Turing reminds us that the jury is still out for everyone else except me.)
As an addition to my comment, I had a difficult time grasping how feeling is implicated in beliefs, but after this week's lecture I think it is clearer. What clarified this was considering an example of a belief, like in class we talked about the belief that it is Wednesday. What didn't click before, makes sense when you think about the fact that if you believe it is Wednesday, it also feels like something to believe that. Clarifying this distinction really allowed me to better understand Harnad's response to Dennett's paper.
DeleteMy reflection after reading this exchange- the hard problem of consciousness is...hard!
ReplyDeleteIn the debate, one side treats it as a non-problem (there is no residual X to explain) and the other as a still obstinate problem (there is residual X), so it looks impossible to reconcile the two views.
Prof. Harnad's point is that whether some processes occur unconsciously or not is irrelevant to the question, why we feel at all. To be charitable, though, a research "program" on consciousness could be to find the difference between correlates of conscious and unconscious activity (which has to be mined for in the way Dennett elaborated) and in the long run examine the conditions of consciousness, not rejecting the possibility that the conditions may be necessary but insufficient. I assume functionalism has made, and will make attempts in this direction.
As Harnad mentioned, the heterophenomenological method (which combines the "subjective" first-person feeling to the "objective" third-person observations of functionality) results in a "hybrid data-set" -- One's feeling capacity and doing/functional capacity are two different capacities. Simply speaking, observation of one's doing capacities does not explain how and why they are feeling -- the distinction between the easy problem and the hard problem. Even addressing the easy problem by understanding the causal mechanisms behind our doing capacities does not help us understand the hard problem of how and why we feel what we feel -- just like successfully building a T4 robot does not help us better understand if they feel at all because of the other mind problem. The hard problem seems to be never soluble then.
ReplyDeleteI had a similar thought pattern, Xingti! As I mentioned in my skywriting for 10a, heterophenomenology might be the most appropriate method we have to studying feelings at the moment. It’s not perfect, but unless we find a better way, that method seems to be the only way we have. Otherwise, I can only think of introspection as a way to study feelings, but Professor Harnad mentioned from the beginning of the course that it was not a valid way of investigating cognitive science. So yeah, the hard problem is hard, and even if we do happen to be able to build a T4 robot, I feel like the problem will still be there. (We could cut open the robot’s brain or do fMRI scans and see if certain areas ‘light up’ when it feels certain things, but that brings us back to the ‘mirror neuron’ problem, where we would be jumping to conclusions about causation just from correlates.)
DeleteIndeed, just like I mentioned in my 10a skywriting, I believe that even through building a T5 robot that would essentially be a copy of oneself, we will still encounter the other minds problem. This is because even though one can imagine that they are exactly a replica of oneself, in reality, it is all just another robot who may or may not have all the feelings that one would assume it have. This to me seems like a pretty close way to solving the problem, yet again it just seems as though it remains and we cannot be certain that another is feeling unless we were to experience what they experience.
DeleteThis is probably because I don't understand some part of this debate, but I feel like here the two interlocutors are talking past each other - or are not talking about the same thing.
ReplyDeleteDenett seems like he's talking about studying consciousness in the concrete sense of the word: among the processes that goes on in our brain, which one are we aware of and not aware of? What are the instances when we think we are aware of something (for example, we think we know the level of detail of our visual field) when data can prove that we are not? The experiences he cites (change blindness, blindsight, etc) seem to revolve around a more descriptive account of consciousness.
Harnad seem to be more concern with the "why" of consciousness. Why do we need to have feeling? To remove our hand from a flame, couldn't we just have evolved a mechanism to move our hand when our nociceptor, or pain detector, activate? Why do we need to *feel* a burning sensation? And so to answer that question, of course many things that Denett bring up are irrelevant. Denett doesn't even seem concerned about the why, or the evolutionary explanation behind feelings.
It seems like Denett and Harnad are just not trying to answer the same question here. I may be totally in the wrong here, but that's just the feeling I got from reading this commentary.
Louise, I agree that it feels like these two first papers aren’t operating on the same wavelength. As you said, it seems like Dennett just isn’t concerned with the “how and why feelings” question that Harnad is pressing him on.
DeleteI think though, that Dennett thinks he is answering the the "how and why feelings?" question because he thinks that if we come up with a complete functional understanding of mental behaviour, the answer to “how and why feelings?” will become self-evident. But Harnad doesn’t see there to be a possible causal/functional explanation for feelings (as he says on page 17, "A simple methodological (hence also epistemic) point about the constraints on causal/functional explanation: It works for everything else, but it doesn't work for feelings…”).
So we get this weird situation where Dennett thinks he is answering a question that Harnad doesn’t think he is answering, and a sense of talking at cross purposes comes from the disagreement about whether there can be a functional explanation for feelings (or, consciousness). Dennet’s response to “your method can’t explain feeling” is to say “no, look, it’s a very all-inclusive functional assessment, it’s not just looking at one thing like verbal judgements, but also 'every blush, hesitation, and frown’(Dennett, page 458)”, which misses the entire point of a Harnad-type criticism, which is that functional assessment itself, in any form, cannot explain feelings. Harnad’s response, though, might not be giving Dennett enough credit. Anything Dennett says will seem unhelpful to Harnad, because of this core belief that functional explanation is incapable of explanation feeling. But if this belief is wrong (while I intuitively agree with Harnad, I’m not sure I understand the total justification for this belief), then Dennett’s methodology might have some merit.
Just to follow up, I just watched the lecture from last week which I had missed initially, so can now say why there can't be a functional explanation for feeling. If I understood correctly, it's that everything we do (all functional activity) can be explained without feeling, which doesn't leave room for a causal explanation for feeling. If all the causal forces we know of are used up in explaining the easy problem, there's nothing left to explain the hard problem. So, there can be no functional explanation for feeling because it is not necessary to explain the complete causal picture of our functional activity.
DeleteI am in agreement with professor Harnard that the zombie hunch isn’t particularly useful and doesn’t really get us anywhere. Obviously, anybody (animal, mineral or vegetable) could be a zombie (or sentient for that matter!) and we do not have a causal explanation to prove zombie-ness or lack thereof. I don’t necessarily disagree with professor Harnard’s defence of the importance of finding an answer to the hard problem, or how and why we have feelings. However, I don’t really understand how this is attainable. I have no idea how one could even begin to go about finding how and why we have feelings. The only thing that occurs to me is asking millions of people to describe feelings in detail - a bit like for learning more about universal grammar or phenomenological descriptions in herterophenomenology. However, animals cannot describe to us their feelings. Moreover, descriptions (even highly detailed descriptions) cannot tell us why. They most likely cannot even tell us how. I don’t know how human beings could ever answer the hard problem! Ah!
ReplyDeleteAlong this line, another interesting thing to consider regarding feelings is whether the feelings arise first and then the conscious recognition of that happens, or the opposite — our cognition determines our feelings. This is a hotly debated question in psychology and I believe it may unlock some questions of cognition, namely whether cognition is the controller of our emotions, or if our emotion arise from external input and in a sense, control our cognition. Additionally, your comment about asking millions of people to describe feelings could lead to some interesting conclusions, however, I fear that a similar problem to gauging pain might happen. For example, when we put our hand in a bucket of ice water we will surely feel pain, however to what extent for every person is unknown. Although they have developed a method to overcome this (cross-modality labelling, I think its called), where the level of pain is compared to a more concrete entity such as the sound of a train whistle (to indicate high pain) or to a dishwasher (to indicate low pain). Something similar to this may be helpful with emotions, but because they are innately subjective it is difficult to compare “love” felt by 2 different people, let alone the global population. So in this sense, I think a lot of progress must be made regarding the process that emotions arise, and how to objectively compare different human emotions, before we are able to determine how and why we have emotions.
DeleteSame as Bronwen, I don't see an answer to the hard problem ever really being attained. Many feelings, subjective experiences, qualia, whatever you want to call them, are ineffable in nature. If you asked a million people to describe what it feels like to believe that it is Tuesday, would they even be able to do so? The very basis of feelings is that they can only be felt by oneself so how will we ever come to a universal understanding of why they exist? *this is reminding me of the song "unwritten".
DeleteAD, (I might be wrong) but I feel like in this debate I would be careful to conflate emotions with feelings, at least in cognitive psychology. Emotional responses can be explained in terms of neural mechanisms and biological processes, and often quantified with measures like galvanic skin responses. I believe the general understanding of the emotion "happiness" is different from the personal experience of feeling happy, which is much harder to explain.
It is definitely overwhelming to think about the actual solvability of the hard problem, but I've found it really interesting reading these responses and thinking about what we've discussed in weeks prior, and how all these different elements of "how and why" that are still unclear like UG and understanding all overlap. There is not just one element of the hard problem that is unclear, it's really an entire network of our mental capabilities that we are trying to untangle from one another. I think in reading these examples, it reminds me of how our language diversity and our ability symbol grounding gives us an incredible range of language abilities, but the subjectivity of this also limits us. Linguistically, even if we asked everyone in the world to describe feelings, these words don't really mean much the more you zoom out from individual, to community, to globally. What these feeling words are grounded in become far too abstracted to really give us any meaningful understanding of what the essence of feeling is.
DeleteI think you all bring up some very relevant and interesting arguments regarding both the hard problem and the distinction between feelings and emotions. Like Lucie says, many subjective experiences are ineffable in nature — they cannot be described in words. I wonder if the hard problem similarly exists in a way that makes it unsolvable. After all, if we could solve it, why haven’t we by now? I guess that is inherently what makes the hard problem /hard/ after all. So, Bronwen, don’t fret about not being able to see how we could find an answer, you are most definitely not alone in feeling that!
DeleteAs for the distinction between emotional responses and feelings, I would tend to agree with Lucie more. We understand the concept of happiness, for instance, but describing how happiness feels is a more difficult task that is not guaranteed to yield similar responses across everyone. This also ties into what Leah says regarding grounding — we have been able to generally understand what happiness is like because of all of the experiences we have had that have led us to be happy, but everyone’s experiences are different. Thus, no two people will have the same definition of “feeling happy,” and it will be harder to define in simple terms.
This reading was a nice transition from the last, particularly in discussing Dennet’s stance on heterophenomenology failing to be THE method that solves the hard problem. While the hard problem has been defined and is interesting to explore and read debates about, even the easy problem in itself is hard. However, at least the easy problem is within reach of solving due to the Turing Test and all the research that has supported reverse-engineering of human cognition. As with Bronwen, I don’t necessarily disagree with professor Harnad’s stance on finding an answer to the hard problem. Also like Bronwen, I fail to see how it’s possible to find the answer to the hard problem. In the last skywriting I said that we can use TT and reverse-engineering to discover the causal mechanisms behind T3, however it feels impossible to discern whether or not they are actually feeling because we cannot find the right metrics/instruments etc. Although when you boil it down, both a T3 robot and humans are all just atoms and matter, why would ours give rise to sentience and not theirs? Is the problem in the scope of the definition of sentience?
DeleteI really enjoyed this reading and agree with prof Harnad.
ReplyDeleteIn Dennett’s writing, he says: “Ramachandran and Gregory predicted this motion capture phenomenon, an entirely novel and artificial subjective experience, on the basis of their knowledge of how the brain processes vision.” Professor Harnad responds that this is just predicting one function from another, with a causal explanation for the association between the two, and hence fails to convey anything about the feeling. In fact, Ramachandran and Gregory only anticipated how individuals would react to their stimuli, not why viewing their stimulus feels like something rather than nothing. Why we feel a certain way in response to certain things is not what the hard problem is asking, but rather why we feel anything at all. Dennett is attempting to offer a causal/functional explanation of feeling using the motion capture research, but he can't. If the participants in their research were zombies, they would still report seeing the same objects (i.e., the same functional output, and Ramachandran's prediction would still be right), but it wouldn’t feel like anything to see them. What's missing here is an explanation for why viewing a motion capture stimuli feels like something rather than nothing.
“"Thoughts" is 100% equivocal. If it just means "internal goings-on that generate certain outputs in response to certain inputs," then no problem (and no problem solved!). But if "thoughts" means "felt thoughts," then you might as well call them "feelings" (what it feels-like to think and reason is just one instance of the multiqualitative world of feelings; there's also what it feels-like to see, touch, want, will, etc.)”
ReplyDeleteThis passage really pushed my reflection on the relationship between thoughts and feelings. Indeed, there is a clear distinction between thoughts that are merely the result of computations (which current-day computers can do) and thoughts that arise from feelings (“felt thoughts”). For example, in Searle’s Chinese room, the person performing computations does not need to understand the language in order to create grammatical sentences. Because they lack a complete understanding of Chinese, it would be impossible for them to formulate an "original" thought in that language and, therefore, to “feel” what accompanies the mastery of a language (using it as a vehicle for expressing emotions, etc...). Feelings are often associated to physical reactions, but outlandish and unique thoughts almost always accompany them. In order to reverse-engineer a human, it would be necessary to replicate feelings and their associated thoughts. Indeed, if we reverse-engineer the thought processes of humans, it follows that feelings are also being reversed-engineered (and vice-versa).
Hi Camille, I think this comparison to Searle’s Chinese Room and the distinction between thoughts and feelings is very interesting! I agree with you that reverse-engineering human cognition would include feelings, as it seems to be impossible to separate thoughts from feelings. Moreover, another important difference is between experience and felt experience. An experience that creates any sort of feeling, whether it is a simple pinch or a large feeling like anger, is a felt experience. In fact, it is hard to think of many experiences that would evoke no feelings at all. Even the experience of not understanding something is still a feeling, since we are able to distinguish what we understand and what we don’t understand based on how it feels. Thus, the “hard” problem seeks to understand how and why we have feelings, since it is inevitable that we do have them and that they do influence our thoughts and experiences.
DeleteHere is one phrase I picked out that I thought was an interesting rebuke to the zombie hunch:
ReplyDelete“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.”
But, doesn’t this leave out an important point? The moon is just a big chunk of rock, and its gravity exists as a function of the position it has -- in relation to other massive objects, in movement, and most notably, in space. But some giant theoretical hunk of rock, with no relation to other objects, with no place in space or time, would not have gravity in the same way that our moon does. A molecule-for-molecule identical to the moon does not have to exist in relation to anything else! With nothing acting against it (other bodies, various forces, things I don’t understand because I am not a physicist), that moon does not have gravity!
It might be interesting to attempt to redeem the zombie hunch on this basis. We have attributed consciousness to those around us based on our past experiences with them, based on knowing where they come from, where and when they were born, who their mothers are, and many other things. We exist by their confirmation of our existence, and they exist as we confirm them. This is relationality ! A robot, even if it is essentially or behaviourally indistinguishable from a human, does not have a birth chart. It doesn’t even have a mother! It does not exist in relation to other objects, nor forces -- say, like a parent’s love and judgement, which structure consciousness much as gravity structures space. Much as we can argue that a moon outside of relationality has no gravity, can we not argue that a robot, given that its coming-into-existence is so fundamentally different from our own, cannot have consciousness?
Hi Sophia! I found your point to be incredibly compelling and convincing. However, I interpreted what Harnad was saying a little differently. Wouldn’t it be impossible for something to be molecule-for-molecule identical without identical conditions creating and shaping it? In my mind, it is not sensible to believe that there could be an identical copy of the moon without the environment that created the moon to be as it is. In a vacuum, the moon would not exist. Similarly, it could not be that there is an identical molecule-for-molecule copy of Harnad without the environmental factors that shaped him. Harnad could not exist as he does without the context surrounding him. Furthermore, he is as tied to his feelings of the world as the moon is to gravity. They are fundamental parts of each, and to imagine them without these key features is to imagine an entirely different and implausible entity, which may not be helpful in this case.
DeleteI thought this was a really cool point and is a really interesting way to think about how environmental conditions and relational properties do drastically impact the fundamental "thingness" of something, like a thinking person or the moon. But as Madelaine points out, those contexts and relationships are fundamental properties of the thing, and without them they couldn't possibly be the thing. Gravity is a fundamental component of "moon-ness", such that an object that looks like a moon in some non-gravitational vacuum would not be a moon. I think this goes back to the levels of Turing Test. OK, maybe this moon can deceieve us at the level of visual properties - it looks just like a moon. But as we complexify the test of moon-ness, it fails it more, like the robot that can only answer emails.
DeleteI thought it was really interesting that you brought up relationality for this discussion on the zombie hunch. It got me thinking and I came to the conclusion that it all boils down to how much of our experiences (such as "having mothers" or "knowing where we are from") are physically stored in the neuroanatomy of our brain. If all of these memories and experiences are physically stored in our brain as synaptic connections, then once an event has passed, all we are left with are the structural changes it has had on our brain. This mean that despite not having felt these past experiences, a zombie would still have the exact same neuroanatomy as it's sentient twin with a sentient past and it would still be impossible to tell them (and their behaviours) appart.
DeleteI enjoyed reading the comments in this thread, particularly because this thread, as well as the entire paper, reminded me of an earlier skywriting from this semester. Back when we were discussing TT and T2 vs T3, Professor Harnad replied to one of my skies asking me if not kicking Eric the robot was really based on social or linguistic convention and whether or not I considered indistinguishability as an option. Now, as we discuss zombies and a new type of indistinguishabilty to consider, the previous point Professor Harnad was alluding to becomes clearer. If for molecule-for-molecule I wasn’t able to discern the difference between a zombie and a human, I definitely would not justify kicking a zombie. Even if I were to find out it’s a zombie, who has had no life experiences similar to mine and/or was “unconscious” I still wouldn’t. I think this is all made with an added layer of difficulty due to the other minds problem.
DeleteThe most illuminating part of this paper was: "So we have this hetero-soup, with the psychophysicist measuring all manner of behavioral and neural and functional correlate (including predictive functional/computational modeling) of feeling, and then we have the feelers who are actually doing the correlated feeling. Then what? What do we do with this heteromix (or is it miscegenation?)" as it showed one of the points of tension between team A and B.
ReplyDeleteIt seems that the main problem with heterophenomenology is that it makes no progress on the hard problem, namely how and why we feel.
I agree with what April said above that after reading this exchange, it really clarified how 'hard' the hard problem of consciousness was. However, Harnad goes as far to say that the hard problem is impossible to solve. Therefore, we can never know how/why we have feelings. Since the hard problem can also be stated as how/why we are not zombies, this can never be answered either. Harnad says that it's impossible for two beings to be completely equal down to the last molecule with one having feelings and the other without feelings. I agree with this because something in between the entire being and the tiny molecules is creating the feeling. Thus, we can never find the “thing” that’s responsible for creating this feeling--or if there even is anything at all. We can answer the Hard Problem of why/how we’re not zombies by saying it’s impossible to be ourselves and be zombies at the same time, but this is a circular argument. Thus, I’m not really sure where to go from here.
ReplyDeleteA passage that I found interesting was the distinction between 1st person and 3rd person. In Dennett’s paper, he is arguing that the best way to study feeling is by studying the 3rd person correlates of feelings, such as verbal reports, hormonal reactions and behavioural reactions. In this paper, Harnad is arguing that ‘1st person’ and ‘3rd person’ are really just weasel words for feeling. This is because even if you are studying the ‘3rd person’ correlates of feelings, you are still not feeling what that person is. As Harnad says, “the only one who can actually feel the feelings themselves is the party of the first part, the feeler (the “1st person”). This reiterates the other minds problem yet again. If we are just studying the observable correlates of feeling, there is no way to tell if that person is truly feeling what they are saying. The only way to know is to be the feeler themselves, which evokes the other minds problem. To me, this argument against Dennett was very compelling as it pokes holes in what he was arguing.
ReplyDeleteEvelyn, I also found this distinction to be quite interesting. I can understand how Dennett may desperately want to make this distinction, separating this science of heterophenomenology from others through the objectivity of “3rd person” data. It is through objectivity that he hopes to achieve a purely scientific status in this study of consciousness. However, as Harnad points out, neutralizing the words in verbal accounts and using “objective” biological data like heartrate, hormone changes, etc. is just thinly occluded 1st person data, as these are necessarily tied to feeling, which itself is subjective. Hormonal changes, heart rate, and behavior are simply integrated into our conscious feeling and cannot be separated from it. Harnad broke down many more weasel words that Dennett uses, illuminating the similarities that lie beneath the distinctions Dennett tries to make between thinking, thoughts, consciousness, experience and feeling. If we follow Harnad’s line of thinking, we can still never know the internal state of the feeler through these “objective” data as we will never be able to feel their feelings as they feel them. There are definitively tied to them and with them.
DeleteAs discussed in class, heterophenomenology doesn’t improve our ability to solve the hard problem. Here Dennett somewhat dismisses feelings and experiences as too complex to scientifically explain, thus failing to understand that this exactly the hard problem; explaining the cause and causal role of feelings.
ReplyDeleteUnlike what he offers, we don’t need ambiguous cases where ‘feelings’ are ambiguous in descriptive terms. Feelings are not ruled out by complexity. We can understand feelings in our own experiences, we know what it feels like to feel something or not, just as Searle felt something to not understand Chinese.
By acknowledging the other minds' problem as an impasse to study feelings, demonstrates that Dennett is not asking the right questions. We know feelings are present as we feel them. A detailed report of feelings as produced in htereophenomenological terms do not amount to experiencing them. It is not about proving their existence, nor to capture their fine qualities, which would result in simulation. The question remains to explain feelings, why do they exist? What gives rise to them? Heterophenomenology, which is essentially T3, can only help us get towards a solution to the easy problem at best.
Hi Lola, I appreciate your insight of how using the other minds' problem as an impasse to feelings shows that Dennett isn't asking the right questions because he still fails to explain the causal role of feelings as this is not something that can ever be tested with heterephenomenological methods since feelings are completely subjective states. As I address in my skywritings, it seems that the hard problem is not something that can ever be solved. We are so concerned with why feelings exist, yet we know that feelings are completely subjective experiences that can only be accurately described by the experient, wouldn't it be possible that the answer to the hard problem can also only be described/answered by the experient?
DeleteI appreciated Professor Harnad’s assertion that “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”, that “somehow” being how a bunch of cells working together produce a feeling creature such as a human. What I appreciate is the intellectual honesty to admit that you don’t know. While reading Dennett’s article, what struck me (and also rubbed me the wrong way) is the way in which he spoke of something like feeling with certainty, as if he had the answer to the hard problem, even though he didn’t. I see this sort of certainty in a lot of scientific and academic writing. It seems that such people have forgotten that the starting point for science is not knowing and are too prideful to admit that they were wrong. However, to admit that you don’t know actually gets you closer to the truth.
ReplyDeleteI have, in a past skywriting, used the term “emergence” to describe a possible mechanism by which feeling may occur. While I do still think that feeling may be emergent, I now understand that emergence is not a complete explanation. In order for emergence to be part of the explanation for how and why we feel, an actual causal mechanism by which this emergence occurs must be suggested.
I enjoyed this article, and the fact that it broke down the flaws and missing pieces in Dennett’s article (which, I have to admit, I had trouble reading — his points felt wrong to me, but I had difficulty articulating why). Now, it’s perfectly clear just how far Dennett was from addressing or responding to the hard problem (and the other minds problem) in his article.
ReplyDeleteOn a technical note, in the article, Harnad says that “David [Chalmers] might know it's consciousness, but I know it's feelings” — that the A team, consisting of Dennett, leaves out. He then says that “Consciousness, being half-epistemic, like thought, is equivocal. This is just about feelings”. However, isn’t consciousness just a weasel-word for feeling? Doesn’t leaving out consciousness equate to leaving out feeling?
Firstly, I really enjoyed the format and casual tone Professor Harnad wrote with in this reading, as addressing Dennett point by point simultaneously allowed me to not only better understand the points Dennett was making, but I felt like Harnard's writing style also presented his own points very clearly and logically, which is something I liked, as many of our readings in this class have been harder for me to follow without having to go back and reread a few times!
ReplyDeleteSecondly, skywriting about this reading in hindsight, I can see this reading as a turning point in this class, where the idea of 'feeling' truly came to the forefront and because something that could be considered among the other debates and discussions that have occurred in the course material. Many other readings have used bigger words and complicated explanations, but the boiling down to the idea of feeling is a great other way to think about consciousness, T-3/T-4/T-5 robots, and can be linked back to many topics we've discussed to this point.
In this paper Prof.Harnad mentioned "On the face of it, an emotion is just a synonym for a certain kind of feeling." So if we can't explain how and why we feel, we can't explain how and why we have emotions. There have been 4 main theories of emotions, one of which Prof.Harnad mentioned in the paper is "W. James's (Error)" which I think refers to the James-Lange Theory of Emotion that functional states caused emotions. J-L Theory apparently faced many objections, but I don't think Dennett's heterophenomenology could face the same objections because he is not attributing feelings to functional states, but more similar to the Cannon-Bard theory, which of course doesn't explain how and why we have emotions either but still purely correlating the functional states with emotions.
ReplyDeleteWhen first considering felt and unfelt experiences, I was wondering what we would make of experiences we might feel but not have explicit memory of. For example, traumatic events that we live through but maybe don’t remember might be considered an unconscious experience. However, once I saw that consciousness could be considered the same as feeling, I understood that an unconscious experience would not relate to the memory of the event, but instead the feelings of that experience. Thus, it is clear to me now that a traumatic event as described would require us to feel something but not necessarily have memory of it to experience it.
ReplyDeleteTo continue from my skywriting in 10A, Prof Harnad on Team B has a different stance on the hard problem. Alternatively, Prof Harnad states that mental states are felt states where it feels like to be in a particular state. Prof Harnad differentiates mental states and felt states because while they correspond, it feels like something to think. He states that interpreting the functional states of others does not explain how and why there are felt states (feelings).
ReplyDeleteAs for how to approach the hard problem, Prof Harnad believes that once we have the causal explanations for the easy problem, there is no degrees of freedom left to explain feeling. Hence, the hard problem is “hard” because cognitive science needs to provide a testable casual explanation for our feeling capacity which is (practically) impossible because there is no casual room left; we cannot reverse-engineer our feeling capacity. Once we have T5, identical to a human being, “identical in all the low-level properties postulated by a completed physics”, there is no fundamental force of nature that can account for how and why feelings are produced.
As for the Zombie Hunch, Prof Harnad believes that there is no such thing as a T5 zombie, an insentience but identical me. First off, we can never know for sure whether the Zombie is really insentient because of the other minds problem which concerns whether the other mind feels. Additionally, unlike Chalmers (who is also on Team B) who believes that a T5 will “have no phenomenal feel”, Prof Harnad believes a T5 cannot be insentient. He equates the explanation to the Zombie Hunch, how and why there canNOT be a T5 zombie, to the hard problem, as such explanation would also explain how and why feeling is necessary to be (pass?) T5.
I really enjoyed this reading as professor Harnad basically directly addresses the issues I raised with some of Dennett’s statements in the previous reading. As he mentions in the reading, the capacity to feel and the capacity to function/do are completely different. This means that even with the heterophenomenological method, at best we can only simply observe one’s functioning capacity, which still would never till us exactly how they got to that state or why they’re in that state. This enforces my idea that Dennett was wrong in the previous reading, we aren’t able to use a third-person perspective to answer all the questions without philosophically significant residue. After this reading, I’m not sure that the hard problem can ever be answered, and I’m even beginning to question if it’s worth answering.
ReplyDeleteI also think that Dennett talks a lot about feeling without directly addressing it. He manages to talk his way around feelings. In the premise of 10a, he talks about the two falsities of believing that something that was there was not, and that something was there when in fact, it wasn’t. For someone to base their argument and tool “heterophenomenology” on these premises without addressing how these feelings come about seems rather shortsighted. It fails in creating a neutral 3rd party as it relies on introspection to do the heavy lifting as to how one forms a stream of consciousness without addressing the real questions surrounding consciousness.
ReplyDelete“To suppose that something that is molecule for molecule identical to me could fail to have feelings sounds as sensible as to suppose that something that was molecule for molecular identical to the moon could fail to have gravity”. Is this essentially not arguing about T4, with internal indistinguishability? I was not entirely sure about this inclusion as it is still contentious as to whether the internal biological structure is what creates feeling. I would however like to continue this discussion.
“How / why do we feel? How / why are we not Zombies” is a rather red herring. How / why are we not zombies is answered by the fact that we know we are not zombies (this includes the feeling that we are not zombies), which leads one back to the unanswerable problem, how do we feel? Specifically, how do we feel we are not zombies?
What I enjoyed about this discussion was the importance of feeling in conscious experience. Even though there are obviously flaws in Dennett’s heterophenomenology argument, what it led me to think about was the dearth of feelings and its relaiton to conscious experiences. If event A transpires, with associated feeling of A, but then later on event A is misremembered and therefore elicits a new feeling (call it B), then essentially it creates two conscious “events” from one objective event that occurred at a point in time. This demonstrates the importance of answering the “‘hard” problem as to why / how we feel because it lends support as to how crucial feelings are in human experience and the potential avenues they open.
ReplyDeleteIn this article, Prof.Harnad proposes that the hard problem is simply explaining how and why we are not Zombies.
ReplyDeleteAs being a T3 robot for a whole semester, I always try to think about if I am really a T3 robot that came out of the lab of MIT, how can I send the information to the other people that I am a 'person' who could also feel. In week 2, Alan Turing suggests I imitate human behavior since humans could only observe others' behavior; they could not see my mind. (The other mind problem) But how could I actually have the 'feelings'? Before I know what is feeling, I need to use my sensorimotor system to perceive the world (symbol grounding problem) and learn to do the right thing with the right kind of thing (categorize learning) in case my classmates notice that I am acting weird, which is not doing the right thing. However, just by doing the right thing, imitating how other people are doing is not enough for me when I am asked why I did what I have done. I need to know why other people have certain behavior and use it as my own motivation when I am deciding what the right thing to do (evolutionary psychology) is. Mimic language ability is easy. Language is just manipulating symbols--words, according to certain rules--syntax and grammar, with different interfaces. Before they sent me out of MIT, they built in some assumptions about the units in my algorithm, which gives me the cue to look for the pattern worth paying attention to. (Language acquisition) Since I was being exposed to enormous volumes of language input, I picked up learning OG really quickly; however, I could not understand what they were talking about when they said there exist universal grammar that governs people's thought. Since I don't have such grammar, I decided not to say something that I had never heard from a human. I could make changes to the sentences according to the OG I learned, but I am not creating new phrases and sentences for which I could not find a template in any kind of human language. (UG) It is worth noting that so far, even though I haven't come up with any solution to prove that I am a feeling creature, everyone treats me nicely as an organism with the same feeling capacity as they do. However, I don't quite understand why they assume that I could feel, but the animals can't since they only joked about conducting craniotomy on me while they were slaughtering other sentient animals.
In conclusion, I still have no idea how to prove that I am not a zombie when I am writing my last skywriting of this course. But I believe, as long as we keep seeking a causal explanation of any feeling mind, we will eventually understand what is feeling, what is cognition.
Hi, Peizhao (Eric)! This is really an interesting post that combined big themes that we have learned throughout the semester and I enjoyed it a lot.
DeleteHowever, regarding the problem of proving that you are not a zombie, yes, you cannot prove to others that you are not a zombie, but the fact that you feel that you are feeling is direct evidence, which is visible, or shall I say feelable, only to yourself.
Also, since T3 is already the correct level for reverse-engineering human cognition, you should also have the UG built-in! Besides, instead of using the word "mimic", I think "learn" would be more appropriate.
Hi, Peizhao! I agree with Zhiyuan that there is no "proof" to others that you are not a zombie or a robot. Because of the other-minds problem, whether others feel or not is something of which we can never be certain. So there is no way to prove that to others, if by proof you mean the proof with Cartesian certainty.
DeleteBut I think Professor Harnad is right in saying that it is not very interesting to think of a zombie who is the same to us in every aspect but do not have feelings. Maybe we should not be too worried about how to prove that we are not! Although we do not have a Cartesian proof, I think passing T5 is a good enough "proof" that we have feelings!
In Dennett’s paper in 10a, he distinguishes A Team and B Team. A Team believes that there is no hard problem—solving the easy problem is equivalent to explain the whole problem of cognition. B Team, on the other hand, believes that A Team leaves unexplained the hard problem, which is how and why we feel.
ReplyDeleteDennett is in A Team. He believes that with heterophenomenology, we can properly study and explain consciousness (feelings). Hence, there is no distinction between easy and hard problem—everything can ultimately be explained casually/functionally.
As a response to Dennett’s argument, Professor Harnad’s main point is that heterophenomenology does not really explain “consciousness”—it cannot explain how and why we feel. He points out that heterophenomenology is “an empirical psychophysical data-gathering paradigm for getting all the measurable ("3rd person") correlates of feelings”. But this correlation does not really explain how and why we feel. “For the functional part will continue to hew only to its functional drummer. All those behaviors and neural processes will have their functional and adaptive causal explanations of the usual sorts. But their felt correlates will not. So where does that get us?”
I like this point made by Professor Harnad: “Dear Dan, you keep giving examples of successful prediction of functions from functions, and then an overall causal/functional explanation of the correlation. But when feeling rather than function is what is being predicted, all progress stops with the prediction. There are no further steps to be taken; only regression back to the functional explanation of the functional correlates.”
Yes, correlations are useful. They can yield theories that makes good predictions. But the problem is that these theories cannot touch the hard problem of how and why we feel without reducing it back to the functional explanation of the functional correlates. In doing so, it never really does the explanatory work.