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
But... "Claustrum Nostrum"
Saturday 2 January 2016
10b. Harnad, S. (unpublished) On Dennett on Consciousness: The Mind/Body Problem is the Feeling/Function Problem
Harnad, S. (unpublished) On Dennett on Consciousness: The Mind/Body Problem is the Feeling/Function Problem.
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...
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...
Subscribe to:
Post Comments (Atom)
I vehemently agree with the point that “This is just about feelings.” What else could so-called “first-person science” be concerning itself with? In the words of William James: “OUR judgments concerning the worth of things, big or little, depend on the feelings the things arouse in us. Where we judge a thing to be precious in consequence of the idea we frame of it, this is only because the idea is itself associated already with a feeling. If we were radically feelingless, and if ideas were the only things our mind could entertain, we should lose all our likes and dislikes at a stroke, and be unable to point to any one situation or experience in life more valuable or significant than any other.” (On a Certain Blindness in Human Beings).
ReplyDeleteFurthermore, Dennett seems to lose sight of what is under all our noses: we feel. He states: “You are not authoritative about what is happening in you, but only about what seems to be happening in you”. It’s not ‘only’ what seems to be happening in me, what seems to be, i.e., feeling, is the whole problem here! For Dennett, what matters is “having a phenomenological belief”. Yet, again, as Professor Harnad points out in the paper, it’s not about the belief, it’s about feeling. Dennett’s eschewing of feeling can be characterized as skepticism of feeling’s centrality as the most esoteric problem facing cognitive science. To rebuke this skepticism, a modified version of G.E. Moore’s “Here is one hand” argument might be appropriate: ‘Here is one hand. And here is another. There are at least two feelings of having hands. Therefore, I feel.’ This feeling is precisely what is most obvious yet most troubling.
However, I do take issue with “There is no accounting for feelings functionally. Period.” I believe that we can possibly make a case building from circumstantial evidence, as if we were ‘peaking’ into feeling. The way forward might be:
Delete(1) The fact that we feel is caused by something in the brain.
(2) We can manipulate the brain to find out what it does, using the same logic as in lesion studies.
(3) Putatively, why do we feel? To seek pleasure and avoid pain. [Debatable assertion, but a useful starting point. Refer below for discussion.]
(4) Research Question: Theoretically, does removing the neural reward system (theoretically, since it’s not ‘one thing’) result in the absence of feeling (e.g., no more feeling of hunger, etc.)?
Evidence to suggest this (Note: NOT a ‘V’ study): “Genetically engineered mice that lack tyrosine hydroxylase in all dopaminergic neurons become hypoactive and aphagic and they starve by 4 weeks of age … The results suggest that dopamine facilitates the output from dorsal striatum, which provides a permissive signal allowing feeding and other goal-directed behaviors.” (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2720267/)
This is the type of circumstantial evidence that I referred to; we can infer that in order for the output from the dorsal striatum to allow feeding and other goal-direct behaviors, it has to also allow beforehand or simultaneously the feeling of hunger and the feeling of being driven for other goals.
Importantly, this is all conjecturing for now until someone can design an experiment to do something similar on humans in an ethical manner, which is likely not possible.
Furthermore, assertion (3) is contentious admittedly. This is tantamount to asserting that feeling causes us to seek more feeling of a pleasurable variety and to avoid feeling of a painful variety. It’s a non-starter in this sense. It’s also contentious when considering the ‘zombie’ situation. The question is, in this case, can a philosophical zombie seek pleasure and avoid pain (whatever that means without being circular) without feeling?
I can’t answer the latter question, which is why I surmise that the best starting point might be to simply remove the neural reward system (in essence, remove seeking pleasure and avoiding pain) and see if people would still feel (hypothetically). The problem remains, however, we might just end up creating a zombie who tells us they feel, putting us back at square one.
Despite all of this, feeling is and always will be the most enigmatic problem and the one that makes us most human. William James: “Now the blindness in human beings … is the blindness with which we all are afflicted in regard to the feelings of creatures and people different from ourselves. … We are practical beings, each of us with limited functions and duties to perform. Each is bound to feel intensely the importance of his own duties and the significance of the situations that call these forth. But this feeling is in each of us a vital secret, for sympathy with which we vainly look to others. The others are too much absorbed in their own vital secrets to take an interest in ours. Hence the stupidity and injustice of our opinions, so far as they deal with the significance of alien lives. Hence the falsity of our judgments, so far as they presume to decide in an absolute way on the value of other persons' conditions or ideals.”
DeleteThe above was written before reading 10c. In my comment for 10c, I try to grapple with the causal role of feeling.
DeleteI'm just shooting off a quick reply to your note about a Research Question. I think even the lack of a base sense feels like something. Losing a sense (sight or smell) still has an experiential quality to it. Losing your motivation or your sex drive again still feels like something. I think the choice of subject - non-verbal mice - might be making the results seem more applicable than they are. And I think you on onto something when you say you'd just end up with the Hard Problem - someone who is experiencing something that can't be reached.
DeleteHarnad’s response addressed the main issue I had with Dennett (which was basically that heterophenomenology can’t provide a causal mechanism for, only correlates of, cognitive states like feelings) and exposed this issue on a deeper level for me. I found myself agreeing with the line of argument that even if you created a perfectly functioning causal mechanism for every other cognitive state (i.e. you solved the easy problem), feelings would still be inexplicable functionally. But then this very direct concluding line kind of made me balk: “…I just call a methodological spade the spade it is: There is no accounting for feelings functionally. Period.” I didn’t balk because I disagree, but I think I finally started thinking about the implications involved in calling the how/why of feelings a “methodological spade”. If feelings are functionally inexplicable (i.e. if the hard problem is really unsolvable), no scientific methodology can hope to causally explain how/why feelings are felt. I know that seems like a pretty elementary rephrasing or mental jump or whatever you want to call it, but I’ve pretty much never decided I thought that something was 100% inaccessible to science in my whole adult life, so this connection was a big deal for me.
ReplyDelete12-word summary of above post: I agree with Harnad but had a minor mental crisis about it
Hey Olivia,
DeleteFirst of all, I can totally relate to that feeling of crisis you had and while I am aware of the difficulties of the hard problem, I don't necessarily think it's unsolvable. Although, I don't think Chalmers or Dennett have even come close to creating a methodology for solving it.
One source of hope (and one that I believe Dr. Harnad has faith in) is Turing Testing and AI. Dr. Harnad claims that an AI cannot pass T2 until it can pass T3. If we could create an AI that has all of the capacities of a human and can convince us that it is conscious and describe the feelings that it experiences (although it might only be weakly equivalent to ours), that seems like the first step to functionally understanding feeling. This is Karl says, but I imagine that the sensory input that a T3 AI would experience wouldn’t be that different from what you are experiencing right now. Since somebody made that AI and functionally understands the mechanisms that are causing the AI to feel that way, there might be some hope to crack the hard problem. Not to mention, a T4 AI would give us a hypothetically identical perspective into what human consciousness is like (in theory) and if you believe that a T4 AI could be created, then all hope is not lost.
I agree with you Karl. Though the other mind's problem doesn't seem like it will ever be soluble, we may well be able to gain many insights that we currently do not have before hitting the other minds threshold. For example, in attempting to make a T3 robot, we may discover that there is something unique and quite peculiar about the dynamics of brains (or living things in general, if you subscribe to the 'life-mind continuity thesis') which we find to be a fundamental piece of the puzzle in successfully making a T3. Obviously this doesn't give us certainty that the robot really does feel, but we may have a fair amount of certainty.
DeletePersonally, I agree with Olivia and Harnad that the 'hard problem' poses a question beyond the scope of causal explanation. Simply, functionalistic causal explanations can never hope to explain feeling because feeling itself does not presuppose some function. As such, since science hold causal explanations to be the gold standard, science itself can never hope to account for feeling. This is exactly the point phenomenology as a philosophical method tries to drive home.
DeleteI agree with Yi.
DeleteI told this little story in 10a but I'll tell it again. A drunk man (M) is clambering about under a streetlight when a scientist (S) approaches him and asks what he's doing.
M: I'm looking for my keys.
S: Well, is this where you lost them?
M: No, but this is where the light is.
To speak metaphorically, science can only look at things "where the light is." Don't get me wrong: it's an extremely powerful method. But it can only tell us about things that are testable.
Science can't tell us, for example, exactly how many birds are flying in the sky right now. There's an objectively right and wrong answer, but we just can't test it. (Yet?)
It can't tell us anything about the existence of God. If God was real, and existed outside of space and time (as God is often proposed to be), then what can we possibly do to test it? Presumably, nothing.
The same goes for feeling. Because of the other minds problem, we can't even identify feeling except in ourselves. So, we can't test for it either. And if we can't test for it, then how can we possibly explain it?
I completely agree with Michael. Without solving the other minds problem there is no objective way of qualifying and quantifying feeling. This is also where Dennett's argument collapses. Even though the method of heterophenomenology tries to objectively capture 3rd person and 1st person accounts to answer the hard problem, there is no method of intrinsic cross validation. It will get no where. I also completely agree with professor Harnad regarding the the hard problem being beyond the scope of a causal framework. Even if we build a T3, that does everything we do and presumably feels, there will be no way of explaining how that feeling emerged. However, the very fact that we are aware of our feelings serves as a starting point for an account of for the hard problem. Its a self generated account. An investigation into the mind through a sustained effort on how a feeling emerges conducted by multiple agents longitudinally may serve as starting point. If the accounts are similar across multiple agents regarding how phenomena elicit feeling we will have some kind of a story.
Delete"Even if we build a T3, that does everything we do and presumably feels, there will be no way of explaining how that feeling emerged." (from Soham)
DeleteIf we were able to build a T3, which is out of our current capabilities, could that possibly be due to the fact that we might have also somehow learned more about "how that feeling emerged"? I'm not necessarily saying that I think we will ever reach that understanding, but when we are speculating about problems that are currently out of our understanding, we should consider the possibility that if they were ever solved, it could be due to the fact that we might have solved other "unsolvable" problems to do so.
To address Karl's response. From my understanding T3/T4 AI can only ever give us information and a causal explanation for how we do the things that we do. For example, how we see the colour red or how we pick an apple from the tree. It does not answer how we feel, Searle's room has shown that something can pass the TT without understanding so it probably can pass without feeling too. It would not give a causal explanation for our feeling of red nor why we feel in the first place. T3/T4 seem like they would only give a functional explanation of functional correlates (such as the actions required to ground words, communicate, and pass the TT) but not a functional explanation of feeling those things.
DeleteI have to rebut against Michael's argument.
DeleteThe example you are thinking of is from Abraham Kaplans 'The Principles of the Drunkard's Search'. It is later used by Freedman as the 'streetlight effect' in which people are more inclined to look for things where it is easiest. Under the streetlight effect argument it is in fact the scientist who is searching under the streetlight.
To extend the parable to the hard problem/easy problem, many cognitive theorists are trying to use behaviour (the area under the streetlight) to look for the origin of feeling (the keys). Whether they look under the streetlight, where the keys are not, or in the dark they will not be successful in finding their answer.
I think it is safe to say that Dennett's heterophenomenology is actually just an observation of behaviour (verbal behaviour) and thus runs into the problem of strong/weak equivalence. All the information we have through the method of heterophenomenolgy is input and output, we cannot say or mimic what the process between the two is. In this way Dennett is searching under the streetlight in behaviourism while the hard problem is located elsewhere.
Even if the Great Pumpkin descended from the Orange Heights and assured us all that, "Yes, Dominique' (whether T3 or T4) "does indeed feel!" -- and even though Dominique's mechanism for doing all she does has been completely reverse-engineered, so we can explain, causally, exactly how and why she can do everything she can do -- that still would not explain how and why she feels. Not even the slightest hint. She does feel; we know that now, from the GP. And it's definitely her internal mechanism (whether T3 or T4) that causes that feeling -- somehow.
DeleteWe just don't know how -- or why...
It's alright to say "I believe the hard problem can be solved."
But it's a good idea to keep clearly in mind what you are up against!
To echo Karl & Olivia, this reading in particular was vexing in that it further illuminated just how circular reasoning with the hard problem is. The following quote from Harnad (paraphrasing Turing) is the most apt articulation of grappling with the concept of feelings in a methodological manner.
ReplyDelete" *TURING: "I don't think we can make any functional inroads on feelings, so let's forget about them and focus on performance capacity, trusting that, if they do have any functional role, feelings will kick in at some point in the performance capacity hierarchy, and if they don't they won't, but we can't hope to be any the wiser either way" "
Reasoning with the hard question ends up being circular in that we firstly address whether the hard question is one that is even important (Dennett's camp says no), and if it is important, any possible explanations that arise from the subdisciplines of cognitive science end up only pinpointing neural correlates of feeling (as Dennett's heterophenomenology does). I agree with Harnard that the hard problem is one that is relevant and not one to be flippantly dismissed. The above quote suggests that once an AI at say T4 or T5 level is reverse-engineered that can do almost everything if not everything that we can do, then the feeling will be a by-product of its capacity. Another way of interpreting the quote is that only an agent or system that can indeed feel will be able to behaviorally do everything that we have the capacity to do. But a counterpoint to the latter possibility is that feelings may be wholly unnecessary for the function - I'd like to operate on the belief that they are in fact necessary.
It seems like they should be necessary, but this is in fact pure speculation. Harnad isn't saying that a successful T3/4/5 definitely will feel, but that he thinks it probably will. If feeling is necessary, then it will be there. If it's not, it won't. But this isn't decisive in either direction.
DeleteI think that the directional relationship of whether behaviour is necessary and sufficient to give rise to feeling, or whether feeling is necessary for behaviour cannot be resolved by creating a T3. Again, the other-minds problem prevents us from knowing whether in fact T3 is actually able to feel, so we may not ever have an answer whether performance capacity is enough for feeling or not.
DeleteTo be honest, one of the points within Dennett's article that I found convincing was that 1st person data is subjective. And, because of this, beliefs may be wrong.
ReplyDeleteHowever, after reading Harnad's response, I now see why this is completely irrelevant. Harnad writes, "But who cares about 'beliefs'...? What we wanted to know about was feelings. Not beliefs about feelings, but feelings." I'm definitely glad I read this paper after Dennett's because it helps me contextualize the topic. As Harnad mentions, it doesn't matter if someone's subjective experience is right or wrong, it just matters that they are feeling that experience at all.
Hello Laura,
DeleteI agree with your assessment that 1st person data is subjective. However, I do fear that we may be throwing the baby out with the bath water by saying "who cares about beliefs?". The distinction you highlight is very important, namely feelings do not equal belief. However, belief is important for the other-minds-problem, in that it is what allows us to believe that other entities are feeling like we do. Furthermore, I do think beliefs are very important in passing T2 and even T1. When communicating, humans tend to rely heavily on pragmatics. Pragmatics includes all the non-literal parts of meaning. Therefore, if I say, “ Mary ate soup or salad”, it gives rise to the implicature that she did not eat both. This implicature is calculated by trying to infer what the speaker’s beliefs are and why they didn’t say a different sentence that might have been more informative. Therefore, beliefs about the world are critical in the way we communicate with one another and I am not sure if they can be ignored.
I also did have a question pertaining to: "On the face of it, an emotion is just a synonym for a certain kind of feeling”. If emotion is a different kind of feeling from sensorimotor feeling, what are the differences? I was thinking about this and came up with a possible theory, in the same way that there is a symbol grounding problem in language maybe the same exists for feelings. There may be a foundational set of feelings from which all other feelings are created by combining them in different ways.
Hi Laura and Valentina,
DeleteWhen Harnad dismisses beliefs, I interpreted him as saying that, for our purposes, they should be treated just like any other feeling. It feels like something to believe something in the moment we are believing it. When we are not believing it, it is just stored data. It seems like in our stored data, certain ‘constellations’ have a particular qualitative tag, brought to our awareness when they are primed. I agree that a successful T3 requires the ability to attain, store, and express beliefs.
Regarding Harnad’s quotation:
"On the face of it, an emotion is just a synonym for a certain kind of feeling”. If emotion is a different kind of feeling from sensorimotor feeling, what are the differences?”
I think Harnad was asking this question rhetorically because he believes there are no meaningful differences in sensorimotor feelings if the goal is explaining why/how it is that we feel at all. In any case it gave me food for thought as well.
I think your point about emotions being a sort of symbol system is interesting. We’re born with the mirror neurons to distinguish the 5 (I think…) main facial expressions/ correlated emotions, and mind read in some capacity. Analogous to the linguistic symbol grounding problem, we ground the faces we see, in meaning, SOMEHOW, via sensorimotor grounding.
This is being pedantic, but sensorimotor experiences differ from emotions because emotions are components of sensorimotor feeling. One emotion cannot complete a mental state. Emotions also differ in that they have more of a prominent role in our functioning than other feelings. They seem to be more of a behaviour generator than other feelings (ex: what looking at an apple feels like, or what touching a doorknob feels like.) They also have a prominent role in mnemonic encoding and retrieval.
I did not interpret Harnad’s quote as meaning ‘feeling’ is separated into different categories. Throughout the course I have simply interpreted feeling, as feeling, a cohesive unit of unexplained phenomena. Scientists like Dennett believed that the hard problem may be solved by building T4. To build T4 would mean to build a robot that was indistinguishable from the human brain. Therefore, if the T4 robot was taken apart, piece by piece, it would have the same neurotransmitters and firing patterns as the human brain. Yet reverse engineering the brain in this way would still not allow us to see its hidden function of feeling because as Harnad says in the video he posted, ‘feeling’ is not worn on the brain’s sleeve the way the heart wears ‘pumping blood’ on its sleeve.
DeleteAlthough is feels different to understand a language, to see a color or to feel an emotion for example, I don’t think the hard problem of feeling warrants further distinguishing beyond how and why the phenomenon of feeling emerges from a brain.
Laura, I think you hit the nail on the head with this. I think that I’ve struggled a lot with wondering why anyone even cares if what somebody is feeling is “right or wrong” – what is a right subjective experience, and what is a wrong one? Is a right subjective experience one that aligns with what is shown by an objective test of that experience? Is a right subjective experience one that aligns with someone else’s subjective experience? The more of these questions we begin to ask, the more besides-the-point they start to become. Even if we found a single explanation for how and why we feel the way we feel, it would be virtually impossible to know if it’s the “right” explanation. A relatively cheesy example I encountered lately was that of an AI in a video game who is hacked, and the player has the option to (somewhat naively, I feel) ask the AI, “Did that hurt?” The AI’s response was, “I cannot feel pain. But the purpose of pain in organics is to avoid repeating experiences, and I would very much like to avoid repeating that experience.” This attribution of the purpose of feeling pain to an evolutionary advantage like avoiding harmful stimuli is just one example of how we have come to feel pain, but we could never know if this is the right one.
DeleteHi Valentina, to me it seems that the crucial difference between emotions and sensorimotor feelings is that the former is top down, while the latter is bottom up. We feel heat and pain if we touch a hot stove by accident, but we don’t think about making ourselves feel pain. If your palms start to sweat before a big speech, and your throat feels really dry, then those reactions came about probably because you’re really nervous about giving your speech. By internalizing and thinking about our experiences, however briefly, we cause our body to react accordingly. We don’t feel embarrassed if there’s no reason to be, nor surprised if there’s nothing surprising going on in our environment or our heads.
DeleteSOS. The moment I feel like I have resolved my stance on this Dan Dennett fiasco, I add a level of abstraction to my thinking that sets me backwards. Full disclosure, there are certainly aspects of Professor Harnad’s rebuttal to Dennett that I am having trouble grasping (e.g. what is “lie-detector science/forensic science). Ultimately, what I gather is that basically Dennett “tackles” the hard problem (and Chalmers’ concerns) by “reframing” what is understood by feelings (excuse my passive aggressive quotations). His methodology of heterophenomenology focuses on people’s perception of their consciousness/feelings, and as such tries to argue that the concern over feeling (essence of the hard problem) is irrelevant. It’s just the perception that matters (and thanks to these accounts being “bracketed by neutrality” we can zero in on it all). This doesn’t make sense to me though because this is all only valid insofar as heterophenomenology is the be all and end all. But it obviously isn’t because at the end of the day it feels like something to feel something. Your perception of that feeling is separate. (In fact, it feels like something to perceive what you are feeling). Better explained by Professor Harnad:
ReplyDeleteNever mind "qualia." Just call them feelings. I can misremember, I can misdescribe, but whatever I felt, I felt. Whatever that feeling felt-like (not how I remember or describe it, but how it felt at the time) is what we are talking about here, and not even how it felt, but that it felt like anything at all. That is the warp and the woof of all this. Explain the how/why of that that and you have won me over to your team. Till then, it all just sounds like beating about the bush...
Beating about the bush indeed!!!
"I also don't believe the weaker version of this: that there could be a functionally equivalent but molecularly different pair of entities, and one of them feels and the other doesn't. (In other words, 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.)"
ReplyDeleteFrom what I understand, Harnad is saying is that any T3-passing system should also feel. Anything that can do everything we can do should also feel as we do.
While I'll grant that a molecule-for-molecule copy of me must also feel, we have to at least be open to the possibility that a T3 will not. A hunch that this is impossible presupposes that feeling is necessary to function. It might be, but it might not. In other words, we could build a T3 that was a zombie and, because of other-minds, we wouldn't know.
The Hard Problem is fascinating, but the more I read into it and the more I think about it, the more I see how it's really a dead end. There's definitely an answer to the question but I don't think it's reachable via the scientific method, so we should just focus on other things. To quote Neil deGrasse Tyson: "The Universe is under no obligation to make sense to you."
Hey Michael! I agree that if there is an answer to the Hard Problem, it will definitely not be reached through scientific methodologies. Certainly, Dennett’s HP supports this assertion. For one, science looks for answers by investigating physical correlates, be it areas of cortex, synapses, action potentials, etc., and we have concluded that we cannot learn how/why from studying where/when in the brain as functional localization would still not proffer an explanation for how a mental process functions causally, nor why it has evolved in its way. Moreover, the existence of feelings cannot be empirically justified, since the only evidence we have of “feelings” is our self-assertion that we feel. Our knowledge of “feeling” is entirely subjective. So if science is not the avenue on which to resolve the Hard problem, what is? Metaphysics?
DeleteHey Michael and Jaime,
DeleteTaking a step back from the confusion that was Dennett's article and the humorously repetitive, focused paper Harnad put forth, I wanted to focus a little on that quote by Neil deGrasse Tyson as a thought of mine has echoed it throughout the semester. We discuss the hard problem so much, and how solving it would be pretty incredible, and just what it is we need to look at in order to solve it. It seems we have everything we need except the actual ability to go ahead and figure out the "how/why". Before jumping into metaphysics, as Jaime mentions, perhaps we just have to come to terms with the fact that at present, we can't solve this problem, but perhaps we will be able to in the future (be it serendipity or otherwise). In the meantime, perhaps the present path should be what was suggested in a previous reading – let’s work on scaling up to the robotic TT, and then see what happens.
What I get from this argument is that the hard problem is that it is hard because there is no consensus on how to formulate it, let alone answer it. The cleavage starts when determining what comes first: one team says it is beliefs (team A), one says it is feelings(team C, since team B is quite irrelevant in my opinion). On the one hand, team A's argument is very attractive because it is much easier to talk about and define beliefs, and therefore gives an (illusory?) impression that progress can be made. On the other hand, it is clear that it feels like something to have a belief, but that a feeling cannot be a belief, so feeling must come first. One could answer that we can believe we feel something, but that actually is feeling.
ReplyDeleteFrom this it seems that the formulation is now cleared: we have found what it is about, so we need to figure out how and why it is about that. And since feelings can't be apprehended properly by any empirical method or any known method, we can conclude that the hard problem can not be solved. However I find this too strong to be supported. We are not able to solve it yet, that is true. It can mean two things: either we don't have the proper tools or methods to answer it, or there is still an issue in the formulation of the problem.
I definitely agree with all your points, I had the same ideas as I was reading the articles. A’s argument is attractive because it’s easier to talk about as you say, because it focuses on the easy problem, hence why it can approach it from an empirical point of view with the knowledge and technology we are at. It talks about more concrete phenomenon, whereas it does not mention the hard problem. However, it seems like there is nothing to say about the hard problem at the moment, except that it is hard and unsolvable at the moment. From the last two points you make, I believe that we don’t have the proper tools to answer it yet. It is beyond our knowledge and technology. It is like asking someone how planets are made when they have just figured out that the earth revolves around the sun.
DeleteHarnad's response to the Dennet paper helped clear up some confusion that I had with the Dennet paper itself.
ReplyDeleteFirst of all, what heterophenomenology even is (and why it is hetero). This was not made clear in the Dennet paper, but from Harnad's paper, it seems to me that it's named hetero due to its ability to deal with many different "data" (verbal, neural, thoughts, behaviors, feelings, etc.). I'm wondering how this is any different from neurophenomenology and other types of phenomenology. I'm not sure I understand Dennet's need for a new term here.
I'm also now understanding that knowing how we have certain feelings, does not tell us how we have feelings in general or more importantly, why we have feelings (in any sense of the word). This reminds me of our previous discussions of evolutionary psychology. I wonder if that can help get the evolutionary sense of why we have feelings. What good do they do us to have feelings? Is this a question we will ever be able to answer?
Also, it helped me understand why "beliefs" aren't the best way to phrase consciousness. At one point Harnad says: "The issue with zombies is not 'beliefs' (whatever those are: for Cummins a thermometer has beliefs, for some maybe a book-page or even a wall does); the issue with zombies is feelings, not beliefs (nor beliefs about feelings)." Beliefs are not explanatory and are not giving anyone the ability to mind-read. There are many instances of pseudo-beliefs, where something can be expressing a belief (perhaps, computationally) but not actually have that belief.
This paper also brought up some questions for me.
I was wondering what Harnad's support for this claim was (whether I missed it, or it was implied in some way) - "I think only further correlations will 'emerge.' I don't know what kind of 'connection' Dave has in mind, but if it's causal (and not telekinetic), I'd forget about it! No other candidate functional role seems viable either." - Why are causal or functional connections being discounted here? Is this saying that feelings are just epiphenomenon of some sort of functional process, and this epiphenomenon cannot be explained and so there cannot be a direct functional explanation from brain processes to feelings?
Hi Rebecca,
DeleteIn response to your “telekinetic and discounting causal connections question”:
In 10c, the authors point to 2 reasons why these attempting to draw causal connections about feeling is problematic (1) we have no evidence to believe that feeling is because of some fifth, invisible causal force (like telekinesis) (2) a study by Libet (1985) showed that the feeling of intending to perform a voluntary action actually may follow the initiation of the action – so feeling may not actually even be correlated with the origins of voluntary behaviour, and thus incorporating it into our causal mechanism (for the easy problem) may not actually be helpful.
But at the same time, I totally agree with you. We can’t discount that there is some causal mechanism for feeling…I think it’s just that at the moment we don’t have any way to investigate it except correlate it with our ‘doing’ capacities…which tells us nothing about the feeling except that it was felt…and we already know that.
"Now faced with these failures of overlap people who believe they are conscious of more than is in fact going on in them, and people who do not believe they are conscious of things that are in fact going on in them heterophenomenology maintains a nice neutrality: it characterizes their beliefs, their heterophenomenological world, without passing judgment, and then investigates to see what could explain the existence of those beliefs."
ReplyDeleteI partially understand the role of heterophenomenology in the later part of this paragraph, however I am really having trouble understanding first part. What are examples of when people thought they were more (or less) conscious than what was true, and how is a heterophenomenology supposed to balance these out? Is it the step-by-step interpretation of the situation that the process of heterophenomenology engages in?
I’m having some trouble understanding Dr. Harnad’s position on the role of feelings in human functioning. In the article, you stated that you did not believe in the possibility of a pair of functionally equivalent entities in which one feels and the other does not (what Dennett calls the Zombie Hunch). However, I was under the impression that you also held the belief that if we tried to explain every aspect of how and why we do what we do, we could do so without encountering the need to incorporate feelings into that explanation (which makes the hard problem unsolvable). But what conclusion does this leave other than to dismiss feelings as a “spandrel” of human functioning? How is it impossible for us to perform human functions without feeling if all functions can be explained without them?
ReplyDeleteI have also been thinking about ways in which feelings might be crucial for us to do what we do. One thought that occurred to me was related to the passage of time— is it possible that our ability to function depends on the feeling that we are doing something in the present moment, therefore enabling us to differentiate that from the past and future, and allowing for memories and planning? Without feelings, might we lose the ability to keep track of time, the way we do during slow-wave sleep?
Another example I came up with was understanding, specifically the “aha” moment that arises when we have a good grasp on a concept. Is there some other mechanism that allows us to, instantly, calculate how certain we are about something, so that we could still use this information in the absence of feelings?
Maybe in assuming that the hard problem is insolvable, Harnad naturally infers that cognitive science can't incorporate it into building the T levels. So maybe rather than him meaning we *could* explain how and why we do what we do without encountering the incorporation of feelings, he means we *would*.
DeleteI like the examples you give of feelings being necessary for doing what what we do. To elaborate on your first example, maybe our notion of time is state-dependent. This is a stretch into answering the hard problem, but maybe it feels like something to be awake because it feels like something to sleep. So, we can't explain why one feeling exists, but we could explain one feeling existing because it necessitates for another to be possible.
I’ve pretty sure i’ve seen youtube videos of Dennett where he states that consciousness is an illusion, and that it is also an epiphenomenon (not implicated in doing what we do, just a sort of inner movie that gives ‘us’ the illusion of playing right behind our eyes, for no functional reason whatsoever.) This implies that we are pre-determined zombies, each of us in effect hallucinating that we’re the agents of our being.
DeleteThis thought definitely makes me feel like my pocket has been picked, or feel a nihilistic twinge, so I too would love if feeling were necessary to our functioning. I believe that it is in particular ways, but in much smaller, less significant ways than we tend to believe.
I heard a good analogy for this once that compared the ‘movie’ we see in our heads to our computer desktop. We have no idea how the computer works, how our files are retrieved and brought to the fore, or what’s actually happening in the software of the computer when we create artwork on photoshop. We have the feeling that we have the mouse, but do we actually? How do we know? This is the zombie hunch.
I think the question you raise is really interesting and important. What are the ways in which feelings might be crucial for us to do what we do? I think the passing of time example is a good one. Poets like T.S. Elliot have considered our consciousness like the surf, the buffer between the past (the sand) and the water (the future.)
I think feelings may play a role when it comes to drives. I think computationally, the feeling circuits have the capacity to travel faster and stronger and have the ability to override the attentional demands made by the less emotional prefrontal cortex. When we see a snake, our startle response takes over our mental state in a quick and uninhibited manner that our less emotional circuits could not muster. Maybe by designing a being feel something, is faster and more efficient than making it compute all that is happening in a dangerous situation, and have to plan to protect itself. This is kind of like the Funes argument. Feeling is more efficient than not feeling, because not feeling makes us pay attention to everything. This still doesn’t get at why we have to FEEL it though. Maybe it’s just a T4 physical anomaly. When you rub two sticks together, poof, feeling.
To sum up the argument made in this paper, Dennet’s refutation of first-person experience does not hold water because even a perfect set of data surrounding the physical processes does not provide a causal explanation. This is what we have seen in class, the ‘how and why’ are missing. I believe that this argument demonstrates why the disagreement cannot be resolved, at leats at our current understanding of the universe. Even with a machine which can “do 100% mind-reading via functional correlates of feeling”, Harnad does not believe that this is a causal explanation, and Dennet does. Harnad rejects Dennet’s justification, and Dennet has no reason to justify it farther: if we are nothing but functional correlates, then we can just point to a tricorder scan and say ‘that’s feeling right there, that blip’. However, Harnad’s point is well taken – this does not explain why we feel things, apart from saying that it is an illusion. If it is an illusion, we still beg the question: why is it an illusion, and why do we feel that illusion?
ReplyDeleteHowever, I think that the conclusion that he seems to arrive at – we cannot arrive at a causal explanation – is problematic. If it happens, and it is causal, then there must be a reason it happens. With sufficient time, technology, and brilliance, we can arrive at an understanding of it – otherwise it is not causally related. I agree with him that heterophenomenology does not innately provide an explanation, but there must be one. We may never stumble upon it – after all, in the words of Neil DeGrasse Tyson, “the Universe is under no obligation to make sense to you”. However, unless we are willing to reject the notion of causality because the problem is hard (and thereby throw away science, or at least it’s authority in explaining the world) I do not think we can throw away a search for answers.
“Beliefs schmeliefs! The "Zombie", ex hypothesi, does not feel. It is true of him that he does not feel. We do feel. It is true of us that we feel. That's the difference, regardless of whether or not there can actually be such Zombies...”
ReplyDeleteI agree with professor Harnad’s blatant disregard for the discussion surrounding whether zombie’s exist and their inability to access conscious states because at the heart of the matter, regardless of if the zombie feels in the same way that we do, the problem persists – how/why do we feel?
Once it is established that this is the only question worth harping on, what is the correct direction to go in? Harnad believes Turing was correct “Do the do-able, and explain the explicable; and foreswear the undo-able and the inexplicable...”
Does this give validity to Fodor’s notion of our inability to learn about how/why from studying where/when as the ‘functional correlates of science cannot explain the how/why we feel’?
Team Bennett leaves out the how/ why of feeling. His methodology of heterophenomenology does not answer nor begin to answer the causal mechanism of feeling but rather deals with function and correlation. Team Chalmers believes that consciousness is important and that the hard problem (the how/why of consciousness) is hard but solvable. Harnad believes that the hard problem is feeling and that it is insoluble, outputs are functions while feelings are not. Functions include brain functions (such as seen on neuroimaging data), behaviour and structural properties of the brain. What they feel like is correlated to those functions but HP does not answer how or why said feeling is correlated with said function. The whole debate about zombies etc. is irrelevant and so is the argument of HP versus first person data because neither give a causal explanation of how/ why we feel.
ReplyDeletewe are robots made of robots; we're each composed of some few trillion robotic cells, each one as mindless as the molecules they're composed of, but working together in a gigantic team that creates all the action that occurs in a conscious agent.
ReplyDeleteThis definitely sends us back to Searle with his CRA. One single unit cannot claim to understand, nor FEEL anything by itself. One neuron cannot be claimed to be responsible for consciousness and qualia. Then what does? Could it be the system as a whole again? The comeback of the System Reply? This sounds very epiphenomalist to me. That the from the whole system, feeling would simply “emerge” out of the system. It seems to only permit Body→mind interaction, leaving the mind→body interaction out of the equation.
I had the same question as you. I don't really understand what parts need to feel or understand for the whole system to be considered able to understand or feel. Similarly, there is the argument about experience, and I wonder if it is it relevant to question whether or not it matters if these single units are able to "experience," or is it just the system as a whole that matters?
DeleteRE: “We are not interested in whether your toothache was real or psychosomatic, or even if your tooth was hallucinated, nor in the conditions under which these various things may or may not happen or be predicted. We are interested in how/why they feel like anything at all."
ReplyDeleteI agree that experience, consciousness, thought, qualia are all synonyms for feeling, acting to circumvent the hard problem. Prof Harnad manages to reduce and redirect Dennett’s question to its simplest form: How and why are there any feelings at all?
From my understanding, figuring out how we feel would allow us to leap over the zombie hunch, the same way our knowledge that the earth is round has allowed us to overcome our hunch that it is flat. This would be the answer the elusive “what is it that humans have that robots do not?” question and clearly tell us what/how/why it is to feel human.
Prof Harnad, you clearly don’t think that heterophenomology or computationalism will answer the hard problem. Do you know what will? Or do you think we should abandon the hard problem as it is simply impossible to account for feelings functionally?
I think the hard problem is insoluble.
DeleteBTW, the hard problem (of explaining causally how and why organisms feel) is exactly the same as the problem of explaining how and why there can't be zombies (if there can't be zombies).
If there can be zombies (i.e., "zorganisms" T3- or T4-indisinguashable from feeling organisms, but they don't feel) then the hard problem becomes, if anything, even harder! Because then we have to explain how and why we are not zorganisms.
Every time you look harder at the hard problem, it just keeps looking harder and harder.
In response to the claustrum article posted above, I agree with Prof. Harnad in saying that consciousness is not just awakeness, it is feeling.
ReplyDeleteA being that is awake but not feeling would be unconscious (like a zombie?) just as a being that is feeling but not awake would be conscious (like a sleeping human). In other words, a robot or a zombie can have the correlates of feeling, without feeling.
RE: “That they correlate (feeling and function) is an interesting fact. Explaining how and why is another matter”
ReplyDeleteI had a lot of hope in heterophenomology and its ability to tackle the hard problem. It's appeal lies in its supposed ability to transcend the other-minds barrier with its fancy techniques for translating feeling from first-person (subjective/invisible feeling) to third-person (objective/observable feeling).
The problem is, however, that even if this"heterophenomenological world" was an accurate portrayal of another cognizer's feelings, it would not help to explain "why" the cognizer feels at all. The "why" of feeling cannot be explained by merely unveiling the feelings of a cognizer, nor can it be modeled. This leaves the hard problem untouched and still insoluble.
Hello Manda, yes, I do agree with you that ultimately it seems that we all concede to the fact that the hard problem is untouched and insoluble. Despite Dannett's efforts in explaining heterophenomology and those first and third-person feelings, it seems that the method would explains the causality of doing to answer merely the easy question and leaves the hard problem even more complex and confounding with his attempt to explain his argument by discussing about beliefs and consciousness.
DeleteI, perhaps, was also optimistic towards that there would be some clues as to how to approach about the hard question. Yet, it seems that we are really capped at the solution of the easy questions, as that once we answered the capacity of causal function, there will be no more left for feelings...
This article definitely addresses my contentions in relation to Dennett, as well as to Chalmers, and explicitly frames the Hard Problem – how and why does the brain generate our ability to feel? Answering this would provide a causal mechanism that goes beyond mere correlation between brain state and “feeling”, and would explain how neural states actually conjure up feelings and more so, the adaptive value of consciousness. I know Harnad maintains that there is no answer to the Hard Problem, but what exactly do those who do believe in the solubility of the Hard Problem accept as a causal mechanism? This article mentions Tom Nagel who says that unlike a chemist’s explanation for the liquidity of water from its molecular constituents, third-person sciences do NOT provide explanations for the correlations between subjective experience and objective conditions (i.e. neural correlates, circumstance, etc.). In solving the Hard Problem, are we pushing for an explanation akin to the chemist's (i.e. an explanation for correlation)? If so, I don’t see how predictive power for “feeling” from neural states, experience or anything we do that’s accompanied by a feeling, goes unwarranted. Would it not eventually provide a causal mechanism (at the very least, an answer to "how") for feeling? In the same way that building a T3 would provide an empirical explanation for everything an organism can do, why, if we were to be able to provide accurate predictions under what conditions “feelings” would occur, would that not also provide an explanation for how we feel? I know Harnad says this would only pertain to the “Easy Problem” since I suppose it treats the predicted feeling as an output to some condition as input (feeling as a performance capacity), in the same way that some input would provide a behavioral output (T3). I agree this does not answer “why”, but why is predictability not enough to explain the “how” of the Hard Problem?
ReplyDeleteRE: Science can read my mind better than I can, can predict better than I can, can reinterpret my feelings for me better than I can” is irrelevant… What is this lie-detector functional-correlate science cannot do is explain how/why I feel at all.
ReplyDeleteThis reply unflinchingly pointed out Dennett’s dismissal of the hard problem, and, like we discussed in class, Stevan says that the problem of feelings is insoluble. Discussing why the hard problem is so, is because once we have explained how and why organisms do everything they do, it seems that we have completed the degrees of freedom in casual explanations – just like the T4/heterophenomenology, we would have explained all the observable aspects, including the action correlation of feelings. But we are left with the problem of explaining feelings. In a functional driven way, why aren’t we just zombies that just do things unfelt? It would be to explain why we don’t just see red through merely visual processing, but also feeling something when seeing red.
Seeing as there is no fifth force in the universe (psychokinetic force), the attempt of explaining feeling and the causality seems to be unmanageable. So, what would come next after solving the easy problem? Do we just concede that we will not be able to explain the causal/functional of feelings? It seems that ultimately here, we are faced with the problem of overdetermination.
(Here onwards, I am treading on unfamiliar grounds. So please correct me if I make any mistakes.)
Despite Stevan saying that the hard problem is insoluble, I wanted to venture this more before I yield to the same conclusion. From what I understood, there is some possible cases for solution to an overdetermined system within a mathematical context. But in attempt to extrapolate this and apply as an approach the hard problem seems somewhat futile. I read a philosophical article that talked about overdetermination outside of mathematical context and found it interesting – but ultimately doesn’t seem to get us any closer to solving the insoluble.
Article link for those interested: http://tedsider.org/papers/overdetermination.pdf
I would love to hear what Stevan and the rest of the class think about this.
Hi Grace: "Over-determination," like "epiphenomenalism" or "spandrels" is no explanation at at all. It's just a posit, like "It's there because it's there."
DeleteWhen we're looking for causal how/why explanations (and not finding them, nor even finding how they could be possible), I can't fathom why we would find "It's overdetermination" soothing.
1st-person/3rd-person voodoo is equally vacuous, to my mind.
("It's an illusion," in contrast, is not just vacuous, but downright self-contradictory.)
“None of these qualitative variants matters a whit. It is that anything is being felt at all that is at issue here. Exotic data on priming and implicit processing don't have any bearing on this at all!”
ReplyDeleteI agree with Professor Harnad on that we have to redirect the question in conjunction with “feeling.” Dennett’s article from 10a is a true example of confirmation bias. As Harnad described it, “why not pick an open-and-shut case like feeling or not feeling warm-now?” If we can explain the doing through scientific psychophysiology, could we not gain insight on the “feeling” through experiments that are designed to infer the “feeling”? As Harnad described in his video, he wills his finger to move. The physiological changes and the neural circuitry has been explained on how and why he can move his finger, but not the how and why of his will to move his finger. When he gave that example, it reminded me of the Benjamin Libet experiment (Libet et al., 1983).
In Haggard and Libet’s article (Haggard & Libet, 2001), they ask the following question: “how can a mental state (my conscious intention) initiate the neural events in the motor areas of the brain that lead to my body movement?” I would change the question to: “how can feelings initiate the neural events in the motor areas of the brain that lead to my body movement?” This would be the direction in which Cognitive Science and the disciplines which constitute it should aim toward.
On a side note, the problem with animal use rests in the “feeling” of those that feel that animals do not feel (people who are still partaking in animal use are not feeling it, so am I). But over the course of this course, I am slowly feeling that animals are no different than humans and therefore, they feel. If animals don’t feel, it doesn’t matter (i.e., if they do not have a nervous system). Plants don’t have nervous system. Human beings have nervous system, so it matters. So do animals.
Hello,
DeleteI'm a bit confused by the last bit here, but I think you're trying to say that it feels like something to have a belief that animals also have feelings, and vice versa. Sort of the same way that it feels like something to believe that another human also feels things, and vice versa. For example, robots that look extremely similar to humans give a more eerie and creepy reported feeling to a human interacting with them, as if there is something empty and missing from them. If you've changed your opinion over this course, that you once thought that animals do not feel and now you do, then I think this is an example of the complication of the other minds barrier. It's possible to invoke the feeling of feeling like something feels (please @Harnad can I use another word then feeling now?), for whatever reason, we have the ability to do this which I think is an interesting discussion. Your subjective experience of this class (and I'm sure the other countless PETA facebook posts), have triggered the feeling of other animals having subjective experiences themseleves. This is why I think studying consciousness is still a necessary pursuit, because strange grey area effects can occur on the boundary of what is considered conscious and what is not.
RE:
ReplyDelete*"Is this truly neutral, or does it bias our investigation of consciousness by stopping one step short? Shouldn't our data include not just subjects subjective beliefs about their experiences, but the experiences themselves?"*
"Well, that doesn't sound like our data (we being the E's, the 3rd person's) but S's data (the 1st person). But who cares? Include them if you like! Call S's toothache, the feeling, a part of your data-set too, if you like. We know that sometimes at least, toothache is indeed correlated with S's behavior. So what do we lose if we just suppose that it's always true. The correlation is perfect, 100%, and you, Dan, the hetero E, "own" both, the behavior and the feelings. They're both your data, in the hetero batch.
Now comes the hard part: How/why does S feel? (Please make sure your reply is not to some other question, which could be perfectly well answered without his feeling anything at all. Don't say "The function of the feeling is to draw his attention to..." -- because the reply will always be: Why/how does any of that "drawing" 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" etc. This is where the ineluctable difficulty lies, not in a hybrid data-base or in authenticating correlations/predictions between functional data and feelings.)"
This seems to summarize the crux of the incompatibility between Dennett and Harnad's respective views on consciousness/feeling. Despite my personal interest in the correlation of brain state and feeling, I cannot claim to project a final and certain solution to the hard problem by those methodological means. Nor can I offer Dennett an out in his dismissal of the hard problem entirely. Despite the solipsist impossibility of subjective investigation, no hypothetical "toothache data" should purport to give the "why".
The debate is almost tiresome, and as Harnad laments, too many "talk past" each other because of the semantic and terminological pitfalls of attempted differentiation, interchanging, or equivocating "consciousness", "feeling", "thought", "experience" without circling back to the same non-explanations, or worse yet the irrelevant distractor i.e. Zombic brain.
The humour of this piece is necessary for the amount of times that professor Harnad has to repeat himself. I think the repetition is necessary and efficacious for both understanding and remembering what is wrong with Dennett’s article. If it weren’t for the humour, I think it would be unbearable to read. It is made clear that predicting feelings is not enough to explain how/why it feels like something to think. But if Harnad doesn’t think an answer to the hard problem is achievable, which he doesn’t think is, then what to make of all this literature which fails to address the very problem it poses? The article is funny and effective but leaves me feeling both hopeless and uncreative, is it suggested that the hard problem should be dropped?
ReplyDeleteHi Krista! I think that while professor Harnad doesn't believe that the hard problem will ever be solved, he still promotes research in this area because it can possibly tell us a bit about the easy problem. I don’t think he is encouraging people to stray away from solving the hard problem, but instead emphasizes that he wants people who are making claims in the field to have a proper grasp of what the hard problem entails in the first place. Professor Harnad really breaks down the main points of Dennett’s paper in a digestible way and provides very valid arguments against Dennett’s ‘heterophenomenology’. I think the main issue here is that people in the field are quick to start explaining the hard problem in terms of functional correlates like “bodily reactions” (i.e. hormones, molecular processes, emotions) but this still doesn’t answer the overarching questions of how and why. There is also a lot of debate and need for clarification for the language used to describe these issues (as Harnad pointed out in lecture). For example, Dennett loosely uses the word ‘mental activity’ which professor Harnad feels is really just a way of bending the language. In all, it seems that before answering the hard problem, one must truly have a grasp of simple terminology used to describe feeling and consciousness to really understand what the question are even about. Ultimately, once one fully understands this problem, one can accept that it may never be solved.
DeleteTo expand on this point a little further, I think that we can go back to our discussion of Fodor’s article for a moment. Fodor’s main argument was that there is no point in studying the where/when because it gets us no closer to knowing the how/why. I would agree with you, Annabel, that I don’t think prof Harnad believes in abandoning the hard problem even if he believes it will never be solved. I think that much of the frustration lies in the roundabout ways that people have used language to try to find solutions to what they think is the hard problem, but are really just flawed projections of the easy problem to sound like the hard problem.
DeleteThis critique by Harnad was eye opening for me. While reading Dennett’s paper, I felt a bit that he did have some sort of a point that using third person data could provide us with useful information. However, I see now that while third person data may contribute to our knowledge, it is not explaining how or why one actually feels. If someone is providing their subjective experience while you are measuring physiological data from them, it doesn’t prove any causal mechanisms. Dennett oversimplifies the issue a lot, and I think I actually did, too. My gut reaction was that if we can find functional correlates or localize certain functions then that would help explain how we experience the phenomenon of feeling. However, all of these findings only explain our capacity to do something, they don’t do any good to explain how those processes lead us to be able to feel.
ReplyDeleteRe: “Frankly, this pseudo-puzzle looks like it's just a consequence of the highly counterfactual premise: 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.
ReplyDeleteThe trouble is, that in the moon's case it is easy to explain causally/functionally, exactly how and why this was so unlikely, whereas in the case of my functional clone, my Turing doppelganger (or myself, for that matter), it is (I think intractably) hard... “
I really liked this quote because it clearly demonstrates the complexity of what is at the center of this issue. Creating robots and machines to replicate human capabilities has been so problematic because we do not know how or why we are able to feel; all we know is how/why we’re able to do certain things that we can do (i.e. our performance capacities). Therefore, since we don’t have this casual/functional explanation of our capacity to feel, we don’t understand how unlikely it would be for a robot that is identical to us to not have this ability. However, because we don’t have this explanation we need to accept that we can only build Turing machines that can do things we can do, as this is something we have more of a functional and casual explanation for. As a further extension of this, and as bizarre as this idea might be, if we were able to understand the causal mechanism behind our ability to feel, would this then be enough to replicate this in non-human/animal objects such as machines and robots? Thus, is the lack of a causal mechanism the only roadblock in this problem?
I found this piece by Dr. Harnad entertaining, but also very focused – it’s easy to see that he made this a reading to really nail down his main point of how and why we feel, without skirting around it. Not only does he go through Dennett’s confusing article on heterophenomenology and take it apart, he does so in an amusing (what was that poem on the 3rd page?) but also repetitive way (to make his point, I assume). As one of the latter articles in our class, it goes back to the ideas we've been visiting since January and brings into focus the one that perhaps ties up all of what we’ve learnt.
ReplyDeleteThe 10b reading here brings us from Dennett's heterophenomenology back to the hard problem: How and why are there any feelings at all? After reading, I am even surer that heterophenomenology (or scientific methods like imaging) won't help to solve the hard problem. And I agree with the idea that, maybe, when we try getting to the answer of the hard problem, we won’t have a *causal* explanation of it.
ReplyDeleteI like the claustrum article clarifying that consciousness is not just awakeness, but I still believe awakeness is still a quite important part of consciousness, of our feelings. Awakeness may not be the focus of the discussion here, but don't we need to be awake to feel (most of the time)? This is not helpful in solving the hard problem, but I don't want to say that the discovery from the article “Scientists discover the on-off switch for human consciousness deep within the brain” is totally useless in knowing consciousness, I think they just used the wrong word “consciousness” to talk about “awakeness”. If awakeness is a (little) part of consciousness, still it might give some insights to us in knowing consciousness; just like what we tried to know about cognition from computation. Even if we do not get the whole picture, a little bit of information might still be better than none.
If JNDs distinguish what we feel and what we don’t feel, I’m just wondering how does it explain for feeling of the difference between two things. For instance, according to Weber’s law, we need a larger difference in order to feel a difference between two things (it’s easier to feel the difference between a book in left hand and 20 books in right hand, but it’s harder to feel the difference between 150 books in left hand and 169 books in right hand because they are both heavy even if the difference between 2 hands is both 19 books).
ReplyDeleteI assume the “zombies” used in this context are people who do not have feelings for something right now. However, I guess we all have this experience of being very sick in bed but still somewhat conscious of what’s happening out there but you just can’t have a verbal reaction. And I also doubt that vegetative patients are not feeling anything at all. But in a later class it was mentioned that “if you’re not feeling you’re not feeling”. But does feeling have to be effable?
ReplyDeleteFeeling is ineffable (what does it feel like to see blue?) except to someone who has had the feeling, or something like it.
DeleteWhere nothing is being felt, nothing is being felt. Not "something, but something that is ineffable."
Remember LAYLEK? This is an example of an uncomplemented category: What does it feel like to not feel? -- Not to not feel X but tofeel Y instead: To not feel anything at all (i.e., what does it feel like to be a Zombie: Answer: nothing).
Uncomplemented categories always create puzzles. Consider the self-denial paradoxes: "This sentence is not true." (Is it true or is it false?)
I guess I’m struggling with figuring out why cognitive science bothers with feeling in the first place. If it wants a universal causal explanation for the hard problem, what methodology can penetrate the other minds problem? A T3, a T4, a T5.. the problem will still be just as impermeable as it is right now. What science looks at is behavior, and so far behavior (including language) has only helped us become closer to solving the easy problem by reverse engineering a Turing Machine.
ReplyDeleteScience can capture, test, theorize about a spectrum of phenomena related to behavior, but all the proposed explanations of finding an explanation for feeling (Dennets, Chalmers, creating a T3 etc.) are either irrelevant or circular. All fields have boundaries to what they can explore. For anthropology, this may be generalizability, for cog sci, it may be feelings. Perhaps cognitive science needs to accept that the easy problem is the limit to what it can explain, and not keep throwing out “explanations” such as the “Claustrum Nostrum” or “heterophenomenology,” which end up not explaining anything. Other fields that accept as a premise that feeling can only be experienced, and that experience can be subjectively shared, should be given more authority when it comes to feeling. It’s just frustrating to see so much energy, time and resources devoted to scientific studies that say they are closer to finding an “explanation for consciousness” when there isn’t any that can be formed within the scientific method.
Nimra, not wanting to defend scientists, I can only say that they are a bit like mountain-climbers, who want to climb it "because it's there." On the face of it, feeling (sentience) is clearly a biological trait. Some (not all) organisms have the trait (animals with nervous systems do, plants and microbes probably don't). It's perfectly natural to ask, as with all other biological traits: How does it work? How does the genome (and the brain) cause that trait? And what is it for (evolutionarily speaking)?
DeleteI think that's at least a partial answer to your question of why cognitive science "bothers" with trying to explain feeling.
Now Turing's answer is that there's no point trying because the only thing science (including biology and physics) can ever explain causally is structure and function (let's call those "doings"). Doings are observable, directly or indirectly, measurable, explainable. Feelings are not observable or measurable. And as for being explainable: It looks as if once all doings have been explained (the easy problem), there are no causal degrees of freedom left to explain feeling; feeling seems causally superfluous (since there is no "5th force").
So that's where all the non-starters keep coming from (“Claustrum Nostrum"... “heterophenomenology”, "Global Workspace," etc.), along with the nonexplanatory metaphysical labels ("Dualism, Epiphenomenalism, Interactionism, Identity Theory, Functionalism" etc.). Lots of words, no causal explanation.
I don't know whether (as you say) "all fields have boundaries." I actually think they are all interconnected. Rather, I think the "hard problem" is unique -- to feeling.
As to giving up on trying to explain it (Turing's advice): I agree. But there's nothing to be complacent about there, because, as I tried to show in Week 10, it is only feeling, and hence feeling organisms (not science or engineering or art or metaphysics or occultism) that matter in the universe. Hard to stop puzzling about how and way...
(Not sure what it means to "give more authority" to other "fields": the ones that just say feeling can only be felt and shared. It's a rather delicate area on which to cede to authority. But the other-minds problem is certainly no justification to hurt... and in fact, when you come right down to it, the only thing that really matters about feelings is not causing hurt. A universe with only pleasure would matter as little as a zombie universe.)
Although I recognize the frivolousness of trying to solve the Hard Problem, I think it's almost in humane nature to want to find an answer. As Prof. Harnad says above, scientists are trying to climb a mountain because it is there. I agree that the scientific method will most likely prove insufficient in solving the hard problem and that perhaps these efforts should be thrown out the window all together, but I think one of the things that is so remarkable about the Hard Problem is that all humans have considered it to some point in their lives in that everyone has questioned why they feel and whether or why someone else feels differently. So perhaps the question is silly to look at from a cognitive science perspective but I think its a high hopes to ask all of humanity to ignore it all together.
DeleteWhile I agree that we will probably never be able to solve the hard problem, I think that the efforts of the scientific community to find an explanation for feeling have been beneficial in other areas! I definitely think that this whole debate about why humans feel and how we feel has encouraged us to consider how/if other species might feel. While we may never be able to develop a causal mechanism of feeling, I would hope that the parallels that we have been able to draw between humans and other species have directed more focus and attention towards animal rights issues, and generated more support for animal rights movements.
DeleteWith a rather different approach to heterophenomenology, Harnad responds to Dennett boldly and with much more clarity in his explanations. The bottom line is that the hard problem cannot be solved no matter how completely we end up solving causal mechanisms and their properties. Once you solve the easy problem and explain everything an organism can do, including a T3 or T4 being, there are no degrees of freedom left to explain feeling. There is no causal mechanism left to explain feeling because such mechanisms are limited to explaining “doing”. Whatever explanation would theoretically exist could only be based on finding a causal mechanism that explains all the data and not feeling. Thoughts and feelings are two different types of experiences, and although thoughts can be observable through data and science, feelings are generated within oneself and one could only find correlations between the two. These T3 and T4 forms of AI only help us solve the easy problem, giving causal explanations for how we do what we do. The phenomenal experience of feeling remains subjective and brings us back to the other minds problem. Thus, heterophenomenology seems like a far-fetched idea since it is very unlikely that there can be a bridging between first person experiences and third person observations. Even with language, hearing, speaking, and understanding all ‘feel’ different and this is what makes ‘feelings’ so special. Feeling will always remain a private experience unobservable by beings other than the one who possesses that certain feeling at that very moment.
ReplyDeleteRE: Consciousness, being half-epistemic, like thought, is equivocal. This is just about feelings. Aboutness has nothing to do with it. It's just the how/why of feeling that the A Team (and everyone else who has a go) invariably leaves out. To not leave it out would be to answer the simple question: "How and why do T3 robots like ourselves feel?" Why don't they just go about their Turing business (including the emailing you and I are doing right now) zombily? Why feelingly?
ReplyDeleteThere can't be zombies, you reply? But I agree! All I am asking is how and why not!
Madison says--but by feeling you mean consciousness. You insist that it is feeling but you mean more than just c fibers firing. There is something it is like to feel pain. That "what it is like" is the feeling you are referring to, not pain (which is a feeling). You do not mean the feelings of hot or cold or sad or happy. You are referring to what it is like to experience these feelings. It is aboutness. I get confused when you say “ It's just the how/why of feeling”. To this I feel compelled to respond with, we feel a certain way because of the hormones and axonal systems of our body (We feel pain before we feel temperature and this is BECAUSE more myelinated fiber results in faster conduction resulting in the feeling of pain vs. the less myelinated fiber resulting in in slower conduction therefore the feeling of temperature. When you take ecstasy you increase hormonal levels of dopamine and serotonin which provides a sense of feeling happy, when you take melatonin you have a sense of feeling tired. So, we can say how we feel x,y,z, is caused by hormonal levels of a,b,c.
RE: (That is does hurt, and that that hurting correlates perfectly with some functional story, is not the how/why explanation we were seeking...)
My response is that you aren’t seeking the causal mechanism to feeling, but rather causal mechanism to how we are aware of feeling.
RE: But my "Zombic hunch" is not that there could be a Turing-equivalent frog that did not feel! My Zombic "hunch" is that I know a how/why explanation when I see/hear one, and there's none in sight for how and why the frog either isn't or cannot be a Zombie. It may very well be the case that it cannot be. But I want to know how/why (and not just "that," or "just-so")!
Say we got rid of the frogs c-fibers, and then maybe her visual pathways, and maybe then her auditory pathways, then somatosensation. How many “feelings” do we have to get rid of before the frog turns into a zombie frog? If we had an emotionless frog it would still be aware its internal state? We are not looking for feeling, we are looking for the awareness
AWARENESS of feeling.
you can't feel without being aware that you are feeling.. but what makes you aware? it is not that you feel
DeleteMaddy, 2nd-orderism doesn't help! "Conscious" is a weasel word. It's just a synonym for feeling. To be "conscious that you are feeling" is just to "feel that you are feeling." Well of course: If you weren't feeling it you wouldn't be feeling that you were feeling it.
Delete2nd-orderism is sometimes useful in social contexts: "I feel that you feel that I feel that..." There the object of the feeling is slightly different in each iteration, but it's still all just feeling.
And we're not really asking here about how and why we are feeling this or that (or even feeling that we are feeling this or feeling that we are feeling that). We're wondering how and why we are ever feeling anything at all!
Is the answer not as simple as, if we didn't feel there wouldn't be anything at all? Without feeling, everything is a laylek which means there would be nothing. If we didn't have feeling (and we were just moving through the universe as zombies) there would be no interaction between two forces, because the zombies wouldn't be a force. But then again a stream doesn't feel and it dynamically interacts/ influences a mountain side. So never mind !
DeleteIt appears Dennett is missing the point of the argument. He seems to provide many examples in which problems were solved from cause and effect scientific observations yet never proposes a solution to how and why we feel things (at the obvious frustration of Dr. Harnad). It is clear that science has established many associations between neurotransmitters, hormones and behavior yet this does not bring us any closer to answering the hard problem. All we have here are associations, and associations are not good enough explanations. In addition, in response to the Zombie hunch how, if we directly replicate the molecular structure down to the physical properties of the protons and neutrons of the atoms of oneself, how could this not generate a conscious being? You cannot just replicate the exact copy of a human and just say this being lacks consciousness definitively.
ReplyDeleteZombie speculations don't have to go to the metaphysical extreme of agonizing over whether there could be two identical T5s, such that one of them could feel and the other could be a Zombie. The full question is already there with two (different) T3s. If there cannot be T3 Zombies, why not? And if there can be, what's the difference (and why aren't we T3 Zombies)?
DeleteI completely understand and agree that Dennett has missed the hard problem. He does not address it, except to reply to Chalmers at one point. However, isn’t solving the easy problem the way to go about solving the hard problem? Is that not all we can do right now, and what Dennett and Chalmers are trying to do with their methodologies, however unsuccessful they may be? Although I agree that Dennett’s paper should have given more weight to his theories about the hard problem, I still appreciate that he is trying to figure out the mechanisms. That might be why I also feel like the implicit and unconscious processes are interesting to study and might help us understand human beings more. It might be just because of their elusiveness and mystery that I do not want to toss them away, but would not figuring out why so many unconscious processes affect our feelings, help us understand the hard problem? It might be another facet of it, but it still contributes to our feelings. How and why do unconscious processes lead to feelings? Why does it feel like something long after I see it? Why are some hidden?
ReplyDeleteFrom 10a Dennett: “I figure that Turing’s genius permitted him to see that we can leap over the Zombic Hunch. We can come to see it, in the end, as a misleader, a roadblock to understanding. We’ve learned to dismiss other such intuitions in the past–the obstacles that so long prevented us from seeing the Earth as revolving around the sun, or seeing that living things were composed of nonliving matter. It still seems that the sun goes round the earth, and it still seems that a living thing has some extra spark, some extra ingredient that sets it apart from all nonliving stuff, but we’ve learned not to credit those intuitions.”
ReplyDeleteThe problem that I’m having here with Dennett is that the earth does actually revolve around the sun even though it doesn’t seem like it, but that analogy doesn’t work with feelings and zombies: it’s doesn’t seem like human beings have feelings, they do have feelings that cannot be ignored! But this is different from saying they cannot be measured. I really liked how the Harnad in 10b points out that the Turing is not leaping over something that it isn’t there, but sidestepping the hard problem; even if we were pretty sure we could generate feeling, we still couldn’t be certain because of the Other Minds problem. Thus, its more relevant to focus on modelling the doing capacity (which reverse engineering a TT-passing bot is trying to do).
I'm surprised that Prof. Harnad didn't refute Dennett's explanation of Turing's question: "Turing: How could we make a robot that had thoughts, that learned from experience (interacting with the world) and used what it learned the way we can do?"
ReplyDeleteI would have thought that the professor disagreed in a couple ways.
Firstly, it doesn't seem like Turing cared about making robots that had thoughts. He just wanted them to be able to do what we could do. It seems liek Turing probably would have known that it was likely impossible to determine whether or not another mind had thoughts/feelings
I do think he got it right with the "learned from experience part," as this seems to suggest a T3
Finally, (granted this might be a bit nit-picky on language), Dennett seems to suggest that Turing wanted robots to learn in the same way that we do. This sounds to me like strong equivalence, and as we have seen Turing didn't care whether or not what passed TT did it through weak or strong equivalence, just that it was able to pass.
While I completely agree and appreciate the humour of repetition that emphasizes how thoroughly Dennett failed to grasp the hard problem, I still struggle with an aspect of how we’ve repeatedly approached the hard problem. Dennett doesn’t see it as anything of importance, Harnad believes it is insoluble and others believe it can be explained away by T4 – and none of these seem like a satisfactory answer to me. Perhaps it’s just stubbornness, but I struggle to resolve to the idea the hard problem is unimportant/insoluble in time or space. I’m comfortable with the notion that currently we are incapable of explaining the hard problem or even just why a T4 – identical to use molecule for molecule could potentially still fail to grasp the why of feeling. Perhaps a blind faith in science, which I ultimately believe to be a mechanism by which we explain things that works (albeit slowly), I fail to see how it would ultimately fail to penetrate the Other Minds barrier when we’ve previously examined the idea that becoming the system (ie, Searle’s periscope) penetrates that barrier. I don’t have a good answer, or even am sure if what I think is remotely close to accurate but it seems that to dismiss something that is currently unanswerable as always unanswerable is far from a productive stance on something worthy of scholarly debate. The fact we currently still cannot adequately explain the physiology of the brain itself, let alone the billions of interactions and molecules etc that work together is encouraging to say that we could still learn about the brain that could contribute a deeper understanding as to the how and why of a mechanism more complicated than the current simple reflex arcs we can explain through and through, but equally as intimately part of the brain.
ReplyDeleteI'm wondering if biocentrism can help account for the how and why of feeling. While farfetched, biocentrism is the belief that everything in the universe is just what we personally perceive it as. While I don't think it's necessarily true, I think it would be interesting to evaluate how and why we feel from this perspective to maybe shed some light. In the same way we think the sun looks like it orbits around the earth, maybe this could be an analogy for feeling. For us, it seems one way and that's all that matters to us. But viewing it from this POV could help undercover the actual scientific reason (like the earth orbiting the sun).
ReplyDeleteIn my last skywriting, I was having trouble putting into words a lot of thoughts I was having on heterophenomenology as a solution to the hard problem. I feebly attempted to put these thoughts on paper, but Harnad addressed them much more eloquently in this paper. First of all, he repeatedly pointed out (as was necessary) that Dennett’s heterophenomenology does not even address the hard problem AT ALL. I kept thinkng, “how does Dennett think that he can take ‘raw data’ from people’s words and expressions and bodily reactions, and turn that into an explanation of what they are feeling?’” And the point is, that he can’t! There is still a gap between what other people say and do and what it is that they feel. As I mentioned before, the whole point of subjective experience is that it is subjective - no one else can try know what it is another person is feeling. Harnad also points out that what Dennett is attempting to do is ‘predict,’ while the hard problem is more than that, it is knowing, understanding, the how and why of feeling. So prediction of peoples feelings is not explanation of how and why people feel. I also felt an overhwelming sense of relief when Harnad stated that the Hard problem simply cannot be solved, because that’s what I had been thinking all semester, but was under the impression that despite being beyond the scope of this course, the hard problem did have a solution. Even if we come up with a robot that has perfect output of human feelings and emotions, feeling is not doing so we still cannot know how or why one feels.
ReplyDeleteWhile I still feel like the heterophenominal method is useful for allowing us to better understand what somebody else’s thoughts and beliefs are and how they may be fallible, I now better understand that it’s not at all addressing the hard problem of how and why there is any feeling involved with the things we do.
ReplyDeleteWhile Dennett’s writing is a joy to read because of his sense of humor, he often complicates ideas in a way that makes it a lot more difficult to understand what he’s trying to communicate. His use of language, such as weasly type words like “qualia” (which Prof Harnad calls him out on) help to blur the matters at hand and make it seem like feeling can be understood. His method of observing consciousness in others is purely behavioral, which is as best as a scientific approach can do in exploring matters of the mind, but ultimately comes nowhere near explaining feeling.
I believe that Harnad is right to criticize Dennett’s proposal of using heterophenomonology, which is essentially psychophysiological data, to study feelings (which is the best term to describe consciousness and qualia and all similar synonyms). This (third person) data is entirely useless by itself. Let’s say we see a part of the brain light up in response to a picture of a snake. Dennett might conclude from this that we have learned something about how fear works in the brain (presumably this part of the brain lights up in response to fear-related stimuli). However his conclusion completely contradicts his premise that feelings can be described solely in terms of third-person data. Setting aside the problems with correlative neuroimaging data for a moment, how does Dennett deduce that snakes (and therefore an image of a snake) elicits a fear response in most humans? Because people say “I’m afraid of snakes”, a form of first person data. How do any of us know what fear is? We know fear from our own experience (first person data) and the information we get about the experiences of others (second-hand first person data). How can you know what physiological data means if you are not able to link it to first-person data. For example, let’s say you were measuring increases in sweating - a physiological response to multiple emotional states (fear, sexual arousal, stress). How do you know what emotion is causing this response without making use of first person data?
ReplyDeleteAfter reading Dennett’s article this much was already very clear to me. I thought that heterophenomonology was actually an interesting approach, but would only be useful if we were to link the objectively measured, third person data to first person data (subjective experiences of feeling). This would allow us to form correlations between psychophysiological data and “feeling”. However after reading Harnad’s article I began to question whether this approach would actually help us solve the “hard problem” at all. After further consideration, I do not believe this approach would help us to solve the hard problem, at least not in the way Harnad conceptualizes the hard problem. The questions of “how and why we feel” requires that the causal mechanisms for feeling be elucidated. I’m not sure if correlations between psychophysiological data and feelings would allow for such causal mechanisms to be found. This data would do nothing to explain why we are not all feelingless Zombies?
I am also glad that Harnad explained that “consciousness”, “qualia” and other words used to refer to an individual’s own internal subjective experience can simply be encompassed under the word “feeling”. It makes it much easier to understand the hard problem as well as make sense of the other readings using this lens.
I agreed with this commentary on Dennett's paper and felt it addressed many of the objections I had when reading Dennett's paper. Third person information may be able to inform us in many ways, but it does not explain the how and why of feeling- the hard problem. Third person information will always be "third person" - It will always pertain to DOING, the easy problem, and not feeling, the hard problem. Solving the hard problem requires reverse engineering a causal mechanism of feeling. Only in this way would we be able to understand how and why we feel. But because we cannot know if something else is feeling we have no reference point from which to reference engineer. This seems like a major and maybe insolvable issue.
ReplyDeleteThis pinching idea was interesting. Pinching is nothing life threatening. I guess you could say it’s a warning? But realistically, it doesn’t tell us much, but we still feel it. We don’t necessarily find it super painful or even painful at all. It’s understandable to feel pain when it is life threatening, but when it’s just a prod, what is it’s purpose? Also, another thought, but would the case of “emotional numbness” count as a lack of emotion? Or does that in itself count as a feeling or felt state? And when we are not thinking (imagine a moment of meditation or just a moment where you really aren’t thinking or aren’t aware that you’re thinking), does this also count as a felt state?
ReplyDeleteRE: “If Turing had instead said:
ReplyDelete*TURING: "I don't think we can make any functional inroads on feelings, so let's forget about them and focus on performance capacity, trusting that, if they do have any functional role, feelings will kick in at some point in the performance capacity hierarchy, and if they don't they won't, but we can't hope to be any the wiser either way"”
I understand that the problem is not about whether there could be any zombies (robots passing TT without feelings) at all or not but I am wondering if knowing for sure that there can't be one would be helpful in asking why and how there can't be one. We will run into the other mind's problem if we talk about Turing's t3 robot and whether it can feel or not but if we can say that there is no way for a t2 machine to do everything that we cognizers with feelings can do without having sensorimotor capacities, what is our obstacle in expanding this argument to say that t2 cannot be passed unless the robot can feel things it is talking about. We said sensed states feel like something, so if the robot/machine/engineered model in t2 can talk about something based on having sensorimotor grounded experiences, I don't see how the sensory part of the grounding can be completed without generating any feelings.