Extra optional readings:
Harnad, S. (2011) Minds, Brains and Turing. Consciousness Online 3.
Harnad, S. (2014) Animal pain and human pleasure: ethical dilemmas outside the classroom. LSE Impact Blog 6/13 June 13 2014
Dennett, D. (unpublished) The fantasy of first-person science.
"I find it ironic that while Chalmers has made something of a mission of trying to convince scientists that they must abandon 3rd-person science for 1st-person science, when asked to recommend some avenues to explore, he falls back on the very work that I showcased in my account of how to study human consciousness empirically from the 3rd-person point of view. Moreover, it is telling that none of the work on consciousness that he has mentioned favorably addresses his so-called Hard Problem in any fashion; it is all concerned, quite appropriately, with what he insists on calling the easy problems. First-person science of consciousness is a discipline with no methods, no data, no results, no future, no promise. It will remain a fantasy."
see also How/Why The Hard Problem is Hard: http://andara.uqam.ca/Panopto/Content/Sessions/e77674a1-a902-40e0-a9ac-7a96185c4399/78025b8e-68cf-4256-b663-3ac51c8ed100-f13e20f4-93f1-4af9-9ba0-8d2c221be233.mp4
and also this (from week 10 of the very first year this course was given, 2011):
“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 non-living matter.”
ReplyDeleteThese “intuitions” really don’t seem equivalent to me. He’s comparing the “intuition” of the Zombie Hunch to the intuition of the middle ages; of a geocentric universe and of a life force. The latter were certainly intuitions but were based on primitive anthropocentric notions that were running rampant before being overthrown by more scientific reasoning and the requirement of empirical evidence.
It seems to me that, if anything, the zombie hunch isn’t a hunch but a reasonable conclusion to come to in the absence of evidence and is, in fact, the polar opposite of the intuitions of the middle ages that accept beliefs as fact based solely on subjective intuition without any objective evidence.
I think that to refute the Zombie Hunch would in fact be an intuition and scientifically regressive. To accept that there would be feeling in the absence of direct evidence, of a Turing machine for instance, is what would actually be similar to the intuitions of the middles ages that Dennett talked about.
Explaining how and why there cannot be zombies is just one (of many) variants of the hard problem of explaining how and why organisms feel.
DeleteI have noticed that I may have misinterpreted the Zombie Hunch to only imply the belief that a Turing machine does not necessarily feel. But, it also encompasses the belief that a complete replica of myself, down to the last molecule, but one that cannot feel, could theoretically exist and that it would behave exactly as I do, give the same output in response to the same input, and would therefore claim to feel, just as I do. But, as I've said in the next skywriting, that doesn't seem possible (just as Dennett and you have said). In fact, I don't even think it should even make sense to someone who believes in the zombie hunch if they believe that a zombie would give the same output in response to the same input. Because introspecting and noticing that I feel, feeling that I feel, is an input which would produce the output of me claiming to feel. A zombie would not have the same input, and by their definition, would not give the same output, so it should not claim to feel.
DeleteIt's a puzzle what zombies (if they had been possible) or even an insentient T3 or T4 would be referring to when they talked about feeling. But it could be their detected but unfelt internal states. Think about it. No use asking Gabe, because rely the same thing a feeler would...
Delete“On the other hand, unless you claim not just reliability but outright infallibility, you should admit that some--just some--of your beliefs (or verbal judgments) about your conscious experiences might be wrong. In all such cases, however rare they are, what has to be explained by theory is not the conscious experience, but your belief in it (or your sincere verbal judgment, etc).”
ReplyDeleteI disagree with him. He is trying to discredit the need to explain experience by equating it to beliefs about one’s own experience from what I understand. It seems to me that our beliefs about our conscious experiences can’t actually be “wrong”. One of the examples he uses is the instinctual “belief” that the detail of our visual field is uniform all the way across to the peripheries. Clearly that belief is false, we know that our focal point in our visual field is the most detailed and the visual field gets gradually less detailed as we move further away from the point. However, even though the belief is false, that does not mean that our conscious experience of it is false. We may not see in detail in our peripheral vision, but we are not consciously aware of that, therefore we do not consciously experience it and that is factually true. In fact, the focal point in our visual field is disproportionately represented in our visual cortex (it’s larger than it should be relative to the whole visual field) compared to the more peripheral areas, which suggests that our brain is processing more of that visual center and, in a sense, ignores the less detailed periphery. So, our belief intuitive belief that the visual field is equally detailed in all dimensions is false because its based on our attention being focused on the focal point of our visual field. So, we consciously perceive more of the center and pay less attention to the periphery, which is a factual conscious experience. I’m not really certain that I understand why Dennett seems to be saying that that would mean that the conscious experience is false?
If a person had an auditory hallucination, he does have a conscious experience of hearing a noise and believes it stems from an external source. What Dennett is saying would be similar to saying that since that person’s belief that his perception of sound originates from an external source is false, then his conscious experience of that auditory perception is false. But whether his belief is true or not, his conscious experience is a categorical truth.
I'm not sure if I have misunderstood Dennett's argument?
You've understood: I can be wrong that there is something wrong with my tooth (or even that I have a tooth), but I can't be wrong that I'm feeling a toothache (when I'm feeling it). That's just what is (or ought to be) Descartes.
DeleteAnd reflect a little on what it means to "have" a "belief": it's like having a toothache: My belief may be false, but the fact that it "feels" true cannot be.
It feels like something to believe something.
Welcome to the hard problem (and its kid-sibling, the "other minds" problem: don't mix them up).
While I want to agree with Prof. Harnad, consciousness isn't just computation, I find it hard to disagree with Dennett's Heterophenomenology. My one point of contention was how he said heterophenomenology takes what people say, their perceived experience, into account as well as their physical data, like hormones, heart rate, and other physiological signs. At this point I was left with, "he seems to be jumping over consciousness and ignoring it." And the very next sentence answered it. If we ourselves can't know all that goes on psychologically, how is someone studying another person supposed to know that? So they just study how these unknown things–such as priming, let's say–affect one's behavior because that's how it manifests.
ReplyDeleteI feel somewhat torn because Dennett is a computationalist, and I just do not agree with that, but heterophenomenology seems to be a systematic way to study a person's experience. It side steps the other minds problem by taking what people say at face value. My concern, though, is how on EARTH does one "try to systematize" someone's experience? I feel like the explanation of heterophenomenology and how it works (using the overlap of false positives and false negatives) makes sense in an abstract way, but I don't see how the implementation could lead to gaining any ground in understanding consciousness.
I agree with your point of contention against heterophenomenology but I don't agree that Dennett answers this point. Particularly, I find that just because I don't know what goes on in my brain that it shouldn't be attempted to be understood. I think Dennett really does ignore consciousness rather than leap over it.
Deletesee the reply below
Delete“First-person science of consciousness is a discipline with no methods, no data, no results, no future, no promise.”
ReplyDeleteThis conclusion by Dennett coincides with the idea that introspection cannot offer any information about the “why” and “how” of the brain. Although I do agree that there needs to be some method of study or of experiment more rigorous than introspection, the idea of the third-person study of the first person does not seem like the correct method. While Dennett would say that this is the Zombic Hunch that I need to leap over and that I have no suggestions about how to scientifically research answers to the Hard Problem, I am not entirely convinced by Dennett’s research. Particularly, I find it difficult to actually see how it answers the hard problem at all if you are just taking first-person statements as data. I find that this does not solve the other-minds problem and that there would be no way to distinguish whether something is actually feeling or not. I do not believe Chalmers has any manner in which to do that either, but I don’t think there is a scientific way of answering the other-minds problem and the hard problem. This is because even if the entirety of a brain was reverse-engineered and made into a robot, we would also have no idea of whether or not the robot is actually feeling. So I think my conclusion from this article is that I think that both Dennett and Chalmers do not have the ability to answer the hard problem.
I agree with you. I don’t think heterophenomenology adds anything to solving the hard problem (how and why we feel). Reading Dennett’s paper reminded me of the mirror neuron and brain imaging discussion of week four, where we concluded that correlations between brain areas and behavior don’t tell us much about how and why the brain can do what it can do. While heterophenomenology, is a step in the right direction by moving away from pure introspection, I’m not convinced that it does anything more than to offer a detailed description of a person’s conscious experience (i.e. feeling), rather than explaining the how and why of this experience.
DeleteGabe, "heterophenomenology" is just T4, and that solves neither the other-minds problem (whether, what) nor the hard problem (how, why).
DeleteMarine & Erica: spot-on.
I completely agree that heterophenomenology doesn’t even attempt to solve the hard problem, although it seems like Dennett believes it does! The problem is that he conflates the third-person interpretation of a first-person description of an event, such as blushing, as the experience of that event. I can describe the emotion I had when blushing, say the thoughts I was thinking when blushing, and Dennett can record my behavior when I was blushing, but none of these would capture what it feels like to be Taylor blushing. The heterophenomenological world doesn’t record that subjective experience.
DeleteAnother point that I want to highlight is the bias of heterophenomenology. In the article Dennett states, “We move, that is, from raw data to interpreted data: a catalogue of the subjects’ convictions, beliefs, attitudes, emotional reactions, . . .” It seems to me that the so-called heterophenomenological world is not a neutral portray of what it is like to be a specific person, but an interpreted picture of what it is like to be that person. Science is not immune from bias or human interpretation of an event. I don’t see how interpreted, third-person data is useful in describing a supremely subjective event such as feeling.
While I agree that heterophenomenology doesn't add anything to solving the hard problem, I think that Dennett may already know that. In this paper, he says "Heterophenomenology allows us to proceed with our catalogue of a subject’s beliefs leaving it open whether any or all of them are Chalmers-style phenomenological beliefs or mere zombie-beliefs. In fact, heterophenomenology permits science to get on with the business of accounting for the patterns in all these subjective beliefs without stopping to settle this imponderable issue. And surely Chalmers must admit that the patterns in these beliefs are among the phenomena that any theory of consciousness must explain." From this what I understand is that those in support of heterophenomenology understand that it isn't helping answer the hard problem directly, but it's a route that will help at some point. This approach and the mention of it above reminds me mirror neurons and how those research findings also don't necessarily help us with the hard problem, but there's a belief that they will help us in some way later.
Delete“From the recorded verbal utterances, we get transcripts [...] from which in turn we devise interpretations of the subjects’ speech acts, which we thus get to treat as (apparent) expressions of their beliefs [...] but then we adopt a special move, which distinguishes heterophenomenology from the normal interpersonal stance: the subjects’ beliefs (etc.) are all bracketed for neutrality.”
ReplyDeleteThis makes the mistake of overlooking the bias of interpretation – behavioural (or verbal) external and supposedly subjective observations are still being made by a person. It is overly optimistic and erroneous to consider these interpretations as neutral.
“heterophenomenology gives you much more data than just a subject’s verbal judgments; every blush, hesitation, and frown, as well as all the covert, internal reactions and activities that can be detected”
It seems to me that he is vastly over-estimating a person’s ability to interpret micro-facial expressions or tones of voice and everything else that is supposedly conveyed by a person’s speech - and not just detect them but infer the appropriate correlated sentiment, context, and all of the first-person experiences. This leads to the Other Minds Problem! People can’t know for sure what is going on inside someone else’s head – their thoughts, feelings and so forth - and can’t even say for certain if they have anything going on in their head at all.
“So heterophenomenology's list of primary data doesn't leave out any conscious experiences you know of, or even have any first-person inklings about.”
The question isn’t the first-person’s impression of their thoughts, and how we could create a machine who says that they perceive the same internal states – the goal is to create a robot that can do everything we can do (easy problem) and that can feel what it’s like to do everything we can do (hard problem) – and everything we can do isn’t necessarily limited by what we KNOW we can do.
“what has to be explained by theory is not the conscious experience, but your belief in it”
Why would we not aim to explain conscious experience? Why would belief be a sufficient end-goal of cognitive science?
“Heterophenomenology is nothing but good old 3rd-person scientific method applied to the particular phenomena of human (and animal) consciousness”
This seems to rely on behaviorist principles of observation (as Chalmers later says as well)... but even if we could make interpretations in an objective and neutral manner (which we can’t), this would still only be describing WHAT we do/believe, not HOW we do/believe.
“the adoption of agnosticism is so firmly built into practice these days that it goes without saying”
I don’t understand the us of agnosticism in this context?
(response continued in comments)
(continued:)
Delete“we are giving you total, dictatorial authority over the account of how it seems to you, about what it is like to be you [...] Of course you might be lying, but we’ll give you the benefit of the doubt.”
What reason is there for giving subjects the benefit of the doubt? He doesn’t appear to offer any reason for fully relying on people’s accounts other than suggesting that there is no alternative. But why should we believe people?
“And if some subjects in our apparatus tell us that their qualia do shift, while our brainscanner data shows clearly that they don’t, we’ll treat these subjects as simply wrong about their own qualia”
Is his ultimate goal to locate events in the brain? If so, we’ve already discussed in class that locating brain activity that correlates with capacities does not itself EXPLAIN CAUSALLY HOW these brain activities and related behaviors came to be. Reducing consciousness to brain activity also has yet to account for the hard problem of what it is like to feel like something that can do everything we can do.
“You can protect qualia from heterophenomenological appropriation, but only at the cost of declaring them outside science altogether.”
Just because his theory is rejected does not mean another scientifically-based theory can’t come along. That’s just hyperbolic.
“Chalmers allows that zombies have internal states with contents, which the zombie can report [...] these internal states have contents, but not conscious contents”
I think one of the essential differences that Chalmers and Dennett quibble over is that the internal states being referred to by Chalmers are simply biological properties, like those required by T4. However, for Chalmers, biological events, just like brain activity, does not necessarily encapsulate conscious experience/feeling states. On the other hand, Dennett seems to directly equate those internal events and the ability to express them verbally as sufficient for consciousness.
“Chalmers believes he gets his justification from his “direct evidence” of his consciousness. So does the zombie, of course."
This describes the Other Minds Problem.
“[...] it is telling that none of the work on consciousness that he has mentioned favorably addresses his so-called Hard Problem in any fashion; it is all concerned, quite appropriately, with what he insists on calling the easy problems. First-person science of consciousness is a discipline with no methods, no data, no results, no future, no promise.”
Not yet having a theory or method of evaluation does not eliminate the possibility of developing one and does not mean that the pursuit is not valuable. Being in the process of trying to develop appropriate methodology does not devalue the significance of what is trying to be explained. Put another way, the notion that feeling states/conscious experience are important for cognitive science (which seems self-evident to me) is not dismissible because it is not observable by traditional behaviorist approaches to science.
Lastly, Dr. Harnad has said in class that he believes that the hard problem is unsolvable; yet he doesn’t disqualify its legitimacy or importance. I suppose it depends on whether or not someone is willing to consider that there may be phenomena outside of our understanding?
I agree with a lot of your observations.
DeleteHere are my takes about 3 aspects of “heterophenomenology” which are variant of yours:
1) Problem of interpretation: Interpretation of verbal reports seems to be one of the big pillars of heterophenomenology, whereas Dennett explains it as “[devising] interpretations of the subjects’ speech acts. […] we move from raw to interpreted data”. While I agree that you cannot escape the need of interpretation when analysing data, I am just wondering if Dennett can be that confident about the interpretation of the verbal reports. The way he seems to put it, in verbal reports there are either true beliefs about feelings, false positive beliefs about feelings, or false negative about feelings. But indeed, language (and categorization made with it) is approximate. So, I am wondering if this method of interpretation can be anything more than “The examinator has a belief X’ that subject A believes that X.”. I am not saying that it is completely impossible to have a strong theory of interpretation, but I feel that the difficulty of this task is a bit banalized by the Dennett article.
2) What and when (and where), but with feelings: The way heterophenomenology is described, it seems more to me like a attempt to describe feeling rather and explain it. In other words, even the method of verbal report will just tell you that “Subject A believed X at moment Y”. Dennett want to combine this method with the 3rd person of analysing brain physiological activity (among other things). But again, this just seems like a “when and where” description, but with verbal reports in addition.
3) Beliefs vs feelings: In the Appendix part of the article, Goldman answered this to Dennett: “a "doxological" (or representational) reductionist like yourself will want to reduce feeling states to dispositions-to-believe “. This makes me wonder about why one would consider beliefs as something separated from feelings. Indeed, the way I understand how Dennett put it, beliefs are something that you hold for true. But can you hold something for true without feeling that this is true for you? (i.e. you believe it).
Dan Dennett (a student of Gilbert Ryle) is, at heart, still a behaviorist. Whether or not an organism feels is not just a matter of "beliefs" -- neither the organism's beliefs nor my beliefs about the organism. And the clincher: it feels like something to believe! So an organism that does not feel does not believe. It just behaves.
Delete"Qualia" is just another weasel-word for feeling. Yes, we can believe a T3 or T4 feels (and we're probably right). But we cannot know for sure whether we're right (except about our own feelings -- that's Descartes' Cogito again), nor, if we are right, how or why it feels.
The hard problem is explaining, causally, how and why organisms feel. What it feels-like is part of the other-minds problem, not the hard problem.
We cannot "dismiss" either feeling or the hard problem, whether or not the hard problem is solvable. Each one of us knows that there's at least one sentient (feeling) organism in the universe.
Whether others feel is a matter of interpretation. Whether I feel is not. Whether there is something wrong with my tooth (or whether I even have a tooth) is a matter of interpretation. Whether I'm having what feels like a toothache is not.
"Heterophenomenology" is just T4. (Dennett, by the way, was the first to describe cogsci as reverse engineering.)
See above replies about what a Zombie would say about believing and feeling. Whether it really believes, hence feels, is another matter. (Ask Gabe.)
The hard problem is how and why we feel. The other-minds problem is whether others feel and how do I know it. Though they are not the same, and I'm not sure whether I'm feeling it right, I have a hunch that they are somewhat closely correlated as to that solving either one of them would lead to the solution to the other:
DeleteIf we know the reason and mechanism behind our feelings, then by observing the same behavior that mechanism displays in others, we can infer whether they are feeling something and how it feels like.
On the other hand, I cannot think of any other way to figure out whether others feel and what it feels like other than to solve the hard problem first. Therefore, if the other-minds barrier can be penetrated, it seems like the hard problem won't be that hard anymore.
However, neither of them can be solved by introspection, as Dennett pointed out: "first-person science of consciousness is a discipline with no methods, no data, no results, no future, no promise. It will remain a fantasy". But I doubt any of them can be answered by heterophenomenology either, since except for a bunch of subjects' data collected, I cannot see it leads anywhere closer to the reverse engineering of our feelings.
The OMP is about how to find out whether and what others feel.
DeleteOMP and HP are related, but, without a solution we have no idea how solving one would solve the other.
For normal adult human beings there is no OMP: it's virtually certain they all feel. But that does not help solve HP.
(The only ones for which there is real uncertainty about whether they feel are people in a coma, very early stage embryos, species that lack a nervous system, and single cells.)
Psychokinetic dualism (a 5th "mental" force in the universe) would have been a solution to HP, but it's false: all evidence is against it. If it had been true, and the 5th force had been measurable, then it would have solved the OMP too. But it's just a fantasy.
Try to imagine a viable solution to either HP or to OMP and then describe how it would solve the other. (Most candidates are just variants of psychokinetic dualism -- and that's part of why the HP is hard.)
Are there any readings you would recommend on Psychokinetic dualism? I might've misunderstood in class that the 5th force argument might explain feeling. Would like to know the evidence against it.
Delete“What a good idea: we can let subjects speak for themselves, in the first-person, and then we can take what they say seriously and try to systematize it, to capture the structure of their experience! And we could call it heterophenomenology.”
ReplyDeleteHeterophenomenology, as Dennett describes it above, relies heavily on a subject’s verbal capacity to explain their subjective experience. I believe this methodology is quite flawed, since it seems to leave out the experiences of potential subjects who lack or have impaired verbal abilities (e.g. babies, aphasics, etc.), who are undeniably conscious too. How can heterophenomenology capture their subjective experiences, if verbal report is out of the question? I wonder if (and how) Dennett addresses this issue. My guess is that we could try to imagine the phenomenological experiences of a baby or aphasic person, but wouldn’t the other-minds problem prevent us from achieving this?
The other-minds problem (OMP) is not just a problem about whether babies, aphasics or nonhuman species feel; it's just as much of a problem about talking adults.
DeleteFinding out what any other body feels is part of the OMP too. And with talking people we can get a good idea, because they can tell us (and if the words are grounded for us, we can understand what they are saying). But the challenging part of OMP is not what but whether.
Dennett’s perspective directly contradicts the idea of Chalmers as he so directly put in the introduction of his essay. He believes that Chalmer’s perspective is based on what he calls a “Zombic Hunch” and that Chalmer’s belief is merely based on intuition. However, as Dennett attacks Chalmer’s belief, I believe that it is essential to know just how hard the hard problem is. The hard problem goes further than just having a robot do everything that we can do, but it is also to have the robot feel everything we can feel. We know that we feel, however, it is not something tangible or even describable. When we think about our 3rd grade teacher, we know we can feel something, but how would you differentiate in describing how you feel when you think about your 3rd grade teacher or about your 5th grade teacher. If we are unable to describe it, how are we able to mimic it in a robot or even know that we have achieved feeling in the robot. Unlike the easy problem, there is no physical behaviour we can see. On the other hand, it is hard to imagine that if a Zombie has all the same parts that we do, and all the same inputs that we do, that a zombie would not be able to feel. If the zombie were not able to feel, I doubt it could give us the same output.
ReplyDeleteDennett’s point about heterophenomenology is an interesting one. I find it intriguing how he used a systematic way to try to break down feeling and human experience. With that said, it almost seems that he ironically also missed the idea of consciousness, as we consciously feel. Dennett uses an example that we don’t experience the different resolutions that our eye can see at different points in space. Although I understand he point in that we are not able notice that difference when we go about our day seeing, I don’t see how that point would invalidate our feeling of seeing in one resolution. What I feel consciously can’t be negated by what is true subconsciously. Just because my reality does not match with what is actually happening doesn’t mean that my reality, or what I feel, is not actually true.
Not knowing, by introspection, how and why we can do all the things we can do, is not the hard problem. We have to wait for cogsci to solve the easy problem and explain it to us. But the hard problem of explaining how and why we feel is cogsci's problem too, whether or not it is ever solved.
DeleteThe hard problem can't be missed if you phase out the weasel-words: You said Dennett "missed the idea of consciousness, as we consciously feel." He could not have missed the redundancy if put this (absurd) way: "missed the idea of feeling, as we feelingly feel."
Try tightening it up more by substituting for the other weasel words, like experiencing, noticing, seeing, subconsciously...
Amy, when you said "What I feel consciously can’t be negated by what is true subconsciously. Just because my reality does not match with what is actually happening doesn’t mean that my reality, or what I feel, is not actually true." It brought to mind some instances in which my brain fooled me into believing something that wasn't true.
DeleteFor example, I have distinct memories of trying to move my arm, and moving the wrong one. What I felt consciously was me moving my left arm, but in reality I moved my right arm. Does this not make my reality untrue?
"My challenge to you is somewhat different: to show that I misdescribe the standard methodology of cognitive science when I say it adopts the neutrality of heterophenomenology."
ReplyDeleteIn the same vain as some of the other commentaries above, it seems that heterophenomenology is an attempt at being the "happy medium" between introspection and behaviourism. With this being said it sounds a lot like people in arm chairs finding correlations between other people thinking in arms chairs. It doesn't actually address how latter arm chair people are thinking (and more importantly, how the "feeling" of thinking occurs) because there isn't any reverse engineering involved.
I am also a little confused as to how heterophenomenology is a new technique, because is this not the same methodology of any psychologist or psychiatrist? You tell them about your thoughts (first-person analysis) about how you feel in relation and then the doctor gives you their (third-person) perspective on what's going on and what your "symptoms" mean?
In conclusion to my thoughts above, it seems that yes, the neutrality and more rigorous scientific method of heterophenomenology are valid techniques to implement, however it doesn’t address the actual problem of cognitive science, the hard problem.
On another note, I find it interesting that Andy Clark and David Chalmers are in opposing camps as they cowrote the "The Extended Mind". The Extended Mind is the idea that a physical object, a notebook for example, could be included as part of one's mind as long as it is reliable, available and trustworthy to the user. Although, I don't believe in the extended mind paradigm myself, I wonder how Chalmers could as the leader of the "B Team"? The B team is concerned with the hard problem of cognitive science (how we feel), so if the argument is that the mind can be extended to inanimate objects (non-feeling things), how would Chalmers justify this? Would he propose that the entirety of the mind does not need to be feeling? Even if he did it seems contradictory to me.
Heterophenomenology is just T4. Talking about what is going on (T2) is part of T4.
DeleteClark & Chalmers on the "extended mind" used to be part of this course. Your question is the right one to ask: "mind" and "mental" just means "feeler" and "felt."
You can "extend" your cognition with remote tools and remote data and google. But a headache -- or any other felt process -- cannot extend beyond the width of your head.
If you want to see what the skywritings on that looked like, they are here and here.
“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. “
ReplyDeleteHeterophenomenology is advertised as a third-person approach to consciousness, using subjective reports of conscious experiences in conjunction with physiological data such as heart rate, hormone levels, and facial expressions. It doesn’t actually offer any insight into solving the hard problem of cognitive science, as it can only tell you what someone is feeling at a certain time. Knowing that a person *has* a conscious experience does not explain *how and why* they are having said experience, and knowing that there are physiological changes associated with these self-reported experiences once again does not answer any relevant questions.
Another gripe I have with heterophenomenology is that we often have no idea why it is that we think the way we do, and we harbor many unconscious beliefs we that cannot express (implicit biases, buried trauma, etc.). Heterophenomenology relies on verbal fluency and trust that the person/entity describing the experience is *actually* having the experience. For the latter, we will never know for sure (just because it is being said does not necessarily mean it is being felt); therefore, heterophenomenology also does not solve the other minds problem.
Obviously, introspection is a flawed methodology for the study of felt states but heterophenomenology is definitely not the magic solution. It does not do the things that Dennett claims it does (false advertising!), and is completely useless in solving anything that is worth solving.
"First-personal" is just another weasel word for felt.
DeleteAnd heterophenomenology is just weather-forecasting -- figuring out what you are feeling from what you are saying and doing (including what your brain is doing -- T2, T3, T4) but not actually reverse-engineering it!
I’m hitting some roadblocks in understanding Dennets argument, starting pretty early on in the article. Dennet emphases ‘experience’ as being central to the ‘B team’ stance on this issue, but to me it’s unclear what he means by experience or that its an accurate assessment of where folks on the (b) side stand (is he talking about the same thing that others talk about when they say ‘feeling’?).
ReplyDeleteIm feeling a bit dense reading this, but in service of what concept is he advocating for heterophenomenology? Is it just to push towards saying that the solution to the hard problem is more in reach (via a turing ish framework) than B team folks think?
Re: Kevin LaFleur
DeleteHi Kevin! I think you raise some interesting questions that I also thought about. Here is what I understood from the article:
"Dennet emphases ‘experience’ as being central to the ‘B team’ stance on this issue, but to me it’s unclear what he means by experience or that its an accurate assessment of where folks on the (b) side stand (is he talking about the same thing that others talk about when they
say ‘feeling’?)."
Yes, I believe that Dennet is referring to what we know as "feeling", though he uses weasel words like "consciousness" and "qualia" instead. For instance, Dennet clarifies that "[Chalmers] insists that…A team leaves out consciousness." The B team also argues that Turing leaves out what Dennet calls the "Zombie Hunch" which essentially describes the fact that we know what feelings feel like because we directly experience it ourselves, even though these feelings are inexplicable. Chalmers believes that Turing simply does not address cognitive science's "hard problem" (how and why we feel).
"Im feeling a bit dense reading this, but in service of what concept is he advocating for heterophenomenology? Is it just to push towards saying that the solution to the hard problem is more in reach (via a turing ish framework) than B team folks think?"
Heterophenomenology is Dennet's solution to the hard problem -- an objective, scientific method to explain and know ineffable subjective experiences or "consciousness" (feeling). He argues that by studying one's verbal reports, behaviour, and internal conditions, we can understand their feelings. Therefore, Dennet would indeed say that "the solution to the hard problem is more in reach" through some observational paradigm which I think is similar to a computationalist's perspective -- we can study cognition by examining symbols and the rules by which they are manipulated (although Turing was not a computationalist).
I hope that helps and please do add on or correct me if I misunderstood anything!
Hi Vivian and Kevin,
DeleteI agree with you, Vivian, it seems that he’s using “experience” to refer to “felt states”.
I just wanted to add that in his “answer” to the hard problem (re: heterphenomenology) – a scientific method in which he observes verbal reports, behaviour and internal conditions in order to determine what an agent is feeling – or what is feels like to be that agent – he is completely ignoring the question “why”, and also the question of how this was developed /evolved for in the first place. It seems to me that he’s dodging the hard problem and diverting the question to something more or less irrelevant.
Yes, "experience," "consciousness," "qualia" etc. are just weasel-words for feeling. Knowing what people are feeling does not explain how or why they feel. HeteroPh is just part of T4, which would solve the easy problem but not the hard problem. What Dan Dennett is actually trying to show is that there is not hard problem because "feeling" is just "believing." Doesn't work, because believing is felt too! So it just begs the question.
Delete“We start with recorded raw data. Among these are the vocal sounds people make (what they say, in other words), but to these verbal reports must be added all the other manifestations of belief, conviction, expectation, fear, loathing, disgust, etc., including any and all internal conditions (e.g. brain activities, hormonal diffusion, heart rate changes, etc.) detectable by objective means”
ReplyDeleteIt seems like heterophenomenology only offers a methodology to test a T4 robot. It includes a bunch of physical parameters that are irrelevant to cognitive scientists beside the fact that they keep the cognitive beings (us) alive. However, heterophenomenology is not cognitive science. It does not offer any explanation about how that T4 being is able to do what it does. I don’t see how we could study cognition by only observing. We need to built something in order to understand how it works. Cognitive science’s job is not to play God and argues about what is conscious or not. Cognitive science’s job is to provide explanations that will allow us to built something that will be able to do all of what we can do-- whether or not we are zombies that are fooled into thinking that we feel. Behavior is all of what we can observe, and for that Dennett is right. But he misses the point. We want to understand how and why humans do what they can do, and to that heterophenomenology cannot answer.
Hi Chloe!
DeleteI agree that heterophenomenology doesn’t offer any explanation about how that T4 robot is able to what it does. In a way, heterophenomenology aims to capture whatever it can with regards to human behaviours (including self report) and hope to make an understanding of how people have the type of experiences that they have. First, it looks like a messy way to study the brain, because of all the issues with self-report differences. Second, and more importantly, heterophenomenology doesn’t make our job any easier for dealing with the hard problem. Heterophenomenology’s method is similar to introspection, with the twist that it is from a 3rd person view. However, we know that this is not a productive way of studying the brain since you cannot get a meta-level explanation of how your brain works, just sitting and feeling and reporting about how it works. It is like “let’s explain the homunculus in your brain”, however it doesn’t explain how that homunculus works and we get infinite regression…
On the other hand, you said “ We need to built something in order to understand how it works.” I do not agree with that part. Yes, that would be a way of explaining how it works. However, the study of brain (without the production) might come up with some explanation with regards to how the brain works.
Hi Chloe and Onur,
DeleteRegarding the comments about reverse engineering – I think I agree with both of you, in that in order to explain the brain we need to (theoretically) reverse engineer it. So yes, I think that “production” isn’t necessary so long as we unambiguously do understand the causal mechanisms for how the brain does all that it does – but if we understand the causal mechanisms well enough to explain how we do everything we can do, we have reached the level where reverse engineering (building a machine that can do everything that the brain can do) is necessarily possible.
Also I just want to add that I don’t think this answers the -why- problem, and this debate doesn’t actually seem to touch on the “why” problem in a meaningful way. That is, if we take away an agent’s ‘feeling’ capacity – does that actually interfere with their ability to do the things that they can do? If the T3 robot can theoretically do everything we can do – is it necessarily conscious, or if it isn’t does that prove that feeling is completely separate from doing?
We can be fooled into thinking others feel but can we be "fooled" into "thinking" that we ourselves (or rather I myself) feel? Descartes, with whom this course started, suggests the enswer is no (with his Cogito [Sentio]).
Delete"Heterophenomenology" adds to cogsci's usual tools (observing and causally modelling behavior and brain function) the "nuclear weapon" of language, which allows mind-reading and "brain-writing" (as Dennett called them): producing and understanding language, with which we can, among other things, describe to others what we are feeling (introspective report). But language is already part of T2, T3 and T4...
If Gabe feels, the HP is to explain how and why; if he does not feel, the HP is to explain why not.
In the course, we have talked about how introspection does not offer a causal explanation for both the easy problem and the hard problem of cognitive science. Introspection can’t tell us how and why the brain does what it does. So, at first, I felt inclined to agree that “first-person science of consciousness is a discipline with no methods, no data, no results, no future, no promise.” Although at first glance, it doesn’t seem like a bad technique to ask people how they feel; the problem is finding a systematic way to analyze these responses, which I can only imagine how variable and inconsistent they could be. Moreover, the other minds barrier cannot be broken… how would I know that what someone says they feel is what they actually feel? I could never know if the next person I talk to is actually a zombie or robot, nor could they know if I am.
ReplyDeleteMoreover, it feels like the hard problem is very underdetermined—much more so than the easy problem could ever be. We can argue that weak equivalence is “enough” for the easy problem. But I don’t think weak equivalence will ever be enough for the hard problem. The hard problem is hard for a reason: it’s unsolvable. But it doesn’t mean that people do not feel and it doesn’t mean that it isn’t valuable to know what and how people feel. Maybe it can’t be solved by cognitive scientists, but it is valuable to know how people feel in daily interactions—that is, if we assume they do feel.
Hi Erica!
Delete“The hard problem is hard for a reason: it’s unsolvable. But it doesn’t mean that people do not feel and it doesn’t mean that it isn’t valuable to know what and how people feel. Maybe it can’t be solved by cognitive scientists, but it is valuable to know how people feel in daily interactions—that is, if we assume they do feel.”
I agree with what you said about the hard problem being unsolvable and the fact that this does not mean that people do not feel. There is no doubt that despite our difficulties in tackling the hard problem, we know that other people around us feel things because we know that we ourselves feel things. You are right in saying that the Other Minds Problem prevents us from knowing with 100% certainty what and if other people are in fact feeling, but I think it is safe for us to assume that all other humans around us feel things and feel things similarly to the way we do. I think the Other Minds Problem would be more relevant when wondering if a robot feels or if someone in a coma feels (as mentioned by professor Harnad in a previous Skywriting post).
I think the point you bring up about the value of knowing what and how people feel is an interesting one. At first, I was inclined to believe that knowing what and how people feel could in some way help us solve the hard problem. This might seem pessimistic of me, but the more I thought about this, the more I began to think that this was not the case. I don’t believe that gathering more information about what and how people feel could help us solve the hard problem. This is similar to the way that I don’t believe studying Mirror Neurons could help us explain the easy problem. I’m not saying that studying what and how we feel isn’t interesting! In fact, I think this is extremely interesting, but I don’t think it provides any valuable information for cognitive scientists. I believe gaining knowledge about this would simply satisfy our curiosity to learn about the brain and human functioning. What and how people feel, just like Mirror Neurons, do not address the question why. Why is the most difficult, but the most important question to answer. Reverse engineering definitely has perks when trying to answer the how questions that we have, but it still has difficulty helping us answer the why questions that we have.
“heterophenomenological world”
ReplyDeleteFirst, a part of the raw data for the first-person experience is verbal reports, language, yet we know such description is always an approximation while the internal conditions can be reported with exact values. It seems that some correlation can be constructed between the two, but since a felt state may need an infinitely long description to be described *exactly*, which is practically unrealistic, it is as if we never know what we are missing. Besides, even if we know exactly how the feelings match with internal conditions, it seems that it still does not solve the problem of how and why we feel.
Also, the meaning of the same words may be different for different people -- yes, a square is always going to a square (mathematically), but what about those less tangible or less well-defined concepts, for example, love? How a person gets the meaning of love is through experience, and since different people may have different experiences, the definition, or the feeling of “love” would be different, and thus how a person would use this word would be different. And, it seems that the definition for such word is dynamic, and what is more important, the process of “updating”/”acquiring” a meaning of a word is not always “narrated” with words –how is this going to be taken into account in the heterophenomenological world that he proposes? According to Dennett, “if some of your conscious experiences occur unbeknownst to you (if they are experiences about which you have no beliefs, and hence can make no "verbal judgments"), then they are just as inaccessible to your first-person point of view as they are to heterophenomenology.”
Imagine an infant that yet knows any words, what she is feeling will not be able to be examined by heterophenomenology because there is no “text” and only internal physiological condition (is being an infant only physiological? “So heterophenomenology's list of primary data doesn't leave out any conscious experiences you know of, or even have any first-person inklings about.” Does not an infant “know” anything just because she does not have “verbal judgements”?), yet arguably it is the cornerstone; or a less extreme case, someone who is experiencing something that she does not understand (yet) (let’s name it X) – she feels, and she can use words to describe what is happening, but she does not know what it means or what other people call it (i.e. the word for it, which is X), and later on, she learns what it meant and what it is called – Was she experiencing X when she experienced it? With all being said, it occurs to me that the construction of a heterophenomenological world is quite chaotic and arguably unrealistic, even theoretically speaking, and most importantly, it does not help us solve the how and why probelm.
I agree with your notion that the use of verbal reports to judge another person's "beliefs", Dennett's weasel word for feeling, can only be approximate. We can't definitively explain what we are feeling, we just know that we feel. Taking someone else's verbal reports about what they feel at face value does not solve the other minds problem - there is no way for us to validate whether they actually feel what they are claiming (& feeling at all). Evidently, just explaining our feelings while simultaneously measuring other physiological responses during these feelings also does not inch us closer to solving the hard problem either.
DeleteWhile I do feel your example of the abstract and subjective notion of love is a bit confusing, I overall agree that Dennett is deeply flawed when he claims an inability to make verbal judgements on your feelings mean that you do not feel anything about it at all. Similarly, just because you are wrong about a judgement doesn't mean you are not actually feeling that something is wrong (e.g., something may not be wrong with my tooth but I can't be wrong that I do feel like my tooth is hurting). You bring up a good point with your examples of the baby and person who doesn't understand something - knowing, understanding, experiencing - are all weasel words for feeling (it all feels like something to know, understand, experience). Evidently, Dennett misses this point. Due to this, it is clear to me that heterophenomenology completely ignores the core of what makes experiences 'conscious' - how and why we feel.
(Thanks for your reply! I just want to clarify my point of bringing up the instance of "love", which is a cliché.) The point that I was trying to make was *not* that "different definitions (I am not quite sure about "wrong" judgement in this case; I would say they are just different) of love from different people makes people feel no 'love'" (if I understood your interpretation correctly), what I was trying to say is that even when people are using the same word, love in this case, it may still have meant different things. And how did that happen? People felt different things when they were "grounding" the word "love". This is why, in order to have "personalized" definitions in “heterophenomenological world”, we would effectively need to write every "grounding" moment down. And then we came across the problem that when we were grounding the word, we may not even have the words to describe (approximate) what we were feeling yet (or were we even always conscious when we ground words?), thus it cannot be written down.
DeleteI think there is widespread agreement as to the fact that heterophenomenology lacks explanatory power when it comes to hard problem or the other minds problem. But, as Prof. Harnad suggested in an earlier post -- "try to imagine a viable solution to either HP or to OMP and then describe how it would solve the other".
ReplyDeleteThis is virtually impossible without venturing into the world of a fifth "mental" force or ideas that graze psychokenesis. So will reverse-engineering consciousness ever be possible? And if not, what does this mean for advancement in cognitive science? All proposed solutions have indirectly emphasized a need to bridge the Cartesian gap between mind and body. I did a bit of research in any approaches that were going on today to try to address the hard problem, and one study by Stephen Grossberg, a cognitive scientist, was the first search response.
A study that Grossberg conducted in 2017 attempted to propose a research mechanism called Adaptive Resonance Theory (ART). ART computationally examines cortical processes with "processes of consciousness" to examine links rather than separate what we would call the "hard problem" and the "easy problem". While his research techniques do seem to assume this sort of other "mental force" the general gist intrigues me, mainly the fact that we should not examine "why we do what we do" from "why we feel what we feel" (even though we know they are not separate, we have discretely separated the two for the purpose of cognitive science research).
I have mentioned this before, but can't the easy problem offer us insight into the hard problem? Perhaps we should start to emphasize that reverse-engineering "why we do what we do" *may* (I'm really not sure, but I want to think that this is true) offer us insight into "why we feel what we feel".
It is ironic that in the efforts to philosophically combat Cartesian dualism, a new dualistic theory of cognition has emerged, separating functions as "hard" and "easy" despite the inevitable reliance of consciousness on *doing*.
Hi Gili!
DeleteThis is very interesting!
I think it is interesting that Grossberg chooses not to separate the question "why we do what we do" from "why we feel what we feel". I have always thought that this separation was necessary because (I personally am inclined to believe) that cognitive scientists are unable to answer the hard problem. Without separating the easy problem from the hard problem, cognitive scientists could not make much progress and could not provide many plausible solutions to the easy problem.
As I mentioned in a previous skywriting, I am uncertain that reverse engineering the brain could help us answer the hard problem. This might seem pessimistic of me, but I do not see a way in which reverse engineering the brain could help causally explain why we feel what we feel. It might provide us with some information as to what mechanisms in the brain are occurring when we feel things, but I don’t think it could ever answer why. I feel as though reverse engineering is a much more practical approach to solving the easy problem. Explaining how and why we do things is simpler (although, not simple!) than trying to explain how and why we feel things because it is much easier to systematically and objectively observe actions than feelings.
Gili, I find it probing that you compare our distinction between the easy and hard problem to a new dualistic theory (I have never thought of it really like that until now). Like Amanda, I think the separation is necessary given that to do and to feel are not the same. While I do not doubt that consciousness and doing rely and affect the other, I fear that conflating the two ignores what is inherently different between them and the true complexities the mechanisms underlying how and why we can feel. As Fodor suggests, if we were able to reverse engineer a T4-passing robot (like what heterophenomenology strives for and what I assume Glossberg is attempting to do with ART), we would know what / where a psychological mechanism occurs in the brain but we would still not know how or why we feel something about anything. While dualists are criticized for explanatory impotence, I must agree with Amanda that separating doing and feeling probably has allowed us to gain more traction with the former (because it is easier to tackle). While there is much for cogsci to answer (with the hard possibly impossible to solve), I'm not sure if field would have be any closer to explaining the mind without this separation.
DeleteAs we have been looking at consciousness from the terms of the "Team A" (emphasizing what we do) and "Team B" (emphasizing feeling), as put by Dennett, it is interesting to look at some other perspectives such as ART you brought forward Gili.
DeleteIn another class, I remember doing a reading by Vallacher and Michaels which proposes a dynamical system of consciousness that emerges from the integration of lower level, often non-conscious, processing. They put forward this whole system, using theories of fixed point attraction and action identification theory and so on.
While explanations of consciousness like these can be interesting and entertaining, I can't help but feel as if they are just people using fancy terminology and shooting in the dark. It has been pretty validating (after years of wondering what these guys in my assigned readings know about consciousness that I don't know as a conscious being) for Prof Harnad to tell us to cross out the weasel words - its just feeling!
I could tell I was going to despise this article from the minute I read Dennett's "interpretation" of Turing, in which he claims Turing attempted to answer the following question:
ReplyDelete"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?”
The first part of this question is by far the worst. First, Turing never attempted to answer this question. I’m guessing here that Dennett is really talking about “feeling” or consciousness (although I know that word is slightly frowned upon in this course). Turing was simply hypothesizing a machine that could do everything computation can do; that is computative by its very nature (and it could do what mathematicians do). It doesn’t seem like he ever hypothesized that these things could “think”, and even in that case, we’re talking about computationalism (which, at this point in the course, I strongly believe is false). They’re manipulating symbols not by their own volition (which would require thought and intention), they’ve following a script or a set of rules.
The second part of this question “learning from experience”, is something Turing failed to do. We all know that Turing wanted to describe T3 capacities but did so using a T2 model. We also know why that doesn’t work (i.e. the symbol grounding problem betrays the computer).
Dennett tries to make Turing’s work into something it’s not: he’s answering an entirely different question. We now know what the answer to Turing’s question is, whereas we still don’t have any satisfying answer to the hard problem.
I agree with you that this "interpretation" of Turing, made by Dennett, is way off base. Turing was in fact not trying to make a robot that has "thoughts" and learned from "experience" - he was simply trying to make a machine that operated computationally and at a T2 level. This does not include "experience" as Dennett would have it. The T2 machine that Turing was proposing would not be able to interact with the world - in fact this is what Turing left out! So Dennett's claim about what Turing was trying to do is completely wrong and misses Turing's point. Dennett is essentially explaining a T3 model (which has sensorimotor experience) however, this is not what Turing proposed.
DeleteI have some additional issues with Dennett's piece. The first being his statement "We 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." This statement is very confusing to me because it seems as thought Dennett is claiming that there is something *over and above* these cells that compose us. It seems here that he is advocating for some kind of *extra* thing like consciousness(feeling) or a 'soul'(feeling). However, we have also seen that Dennett thinks 'consciousness' is some sort of illusion. So, what does he think makes up this *extra* thing in us that allows us to have this feeling created from molecules which he claims are all mindless? What is the extra ingredient?
Additionally, I have an issue with Dennett's whole heterophenomenology method which he explains "moves from *raw* data to *interpreted* data". But the person interpreting the data is not the one who is feeling it...how would the interpreter be able to correctly interpret if the feeling (or belief/emotional reaction as Dennett calls it) isn't coming from them?
Dennett talks a lot about 'unconscious mental activity' and (feelings) which the person is not aware that they are feeling. I agree that we can not be aware of what's going on in our minds a lot of the time and perhaps there are underlying reasons for our actions or feelings. But, if we are talking about the hard problem I'm not sure if this really applies. If we switch "unconscious mental activity" to felling the issue seemingly goes away. The idea of "unfelt feelings" seems very weird. 'Unconscious mental activity' doesn't seem like the right thing to be talking about if we're trying to address the hard problem. Additionally, I don't know how someone else would ever have "better" access to someone's "unconscious mental activity".
Lastly, I know this has been mentioned before, but Dennett uses a lot of weasel words which could all be reduced to 'feeling'. Dennett talks about "subjects’ convictions, beliefs, attitudes, emotional reactions" which can all be reduced to feeling. I believe this would really clear up some things and untangle a lot of Dennett's 'arguments'.
Hi Flora
DeleteIn regards to your reply to “We 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”, I’m not sure if Dennett is implying some overarching thing that composes us. It seems to me that he’s using this explanation to say that there is no extra element of “consciousness”, as he thinks we are just “robots made of robots” and *our actions* resulting from the coordination of our robotic cells are the only ingredients required for a conscious agent. His later argument stating there is no heterophenomenological difference between Chalmers and “Zombie” Chalmers, supports the notion that there is no added element of consciousness (as Chalmers and Zombie Chalmers are made of the same things and express the same beliefs). Dennett disregards “feeling” chalking everything up to raw behavioral data expressing “beliefs”.
Heterophenomenology is wrought with mistakes and assumptions in uncovering consciousness. In fact, the validity of his argument is circular. All of the errors lead back to the Other Minds Problem. How are we to determine whether others feel? Dennett believes heterophenomenology is the answer to this unsolvable problem. The data used by Dennett is collected through highly fallible instruments. Even if such measurements were precise, the interpretations are all answering the wrong question: what and where. Heterophenomenology cannot determine how and why consciousness occurs. Dennett uses heterophenomenology as a way to ‘[leap] over the Zombie Hunch’. However, I find this to be the same criticism Dennett condemns Chalmers for: using intuition… the reverse hunch (of Chalmers) that we can just ‘[leap] over’ an unsolvable problem.
ReplyDeleteDennett seems to be addressing the question “what are conditions for feeling something?”. And then he argues that feeling can be measured as the sum total of verbal reports and physiological measures. However, it making these arguments Dennett seems to be making multiple assumptions (some of which have been pointed out in earlier comments).
ReplyDeleteFirst, it seems like Dennett’s arguments rest on the premise that feeling is the sum total of all that is measured using heterophenomenology. But is feeling something just the sum total of it? Feeling is not just the sum total of elements of heterophenomenology. As pointed out by others as well, Dennett is confusing feeling with beliefs, and assuming that verbalisation can completely capture feeling.
And even if we were to assume that feeling is just a sum part of the elements of heterophenomenology (which is not) how will all the data in heterophenomenology, be integrated in order to explain a feeling? How do we know how to weigh each experience or thought or physiological measure? To use a computer analogy, is there something our software is doing (like an additional level of code) by integrating all this data that results in feeling/conscious awareness? Or certain differences in our basic neural anatomy that causes the same algorithm to generate different feeling. Given the differences in our memories, there are differences. And so, the proposal that “if some subjects in our apparatus tell us that their qualia do shift, while our brainscanner data shows clearly that they don’t, we’ll treat these subjects as simply wrong about their own qualia”, is not well-founded, and assumes that neural signals should capture all aspects of feeling.
We do not know what generates a feeling or why. And Dennett does not offer a way in which feeling can be derived from heterophenomenology. As Tom Nagel critiqued heterophenomenology can at best provide us with correlations between the feeling and described experience, but does not explain how or why the feeling arose.
Second, although anything can be described by language, in every day utterances, such as when the subject is describing their experience, given that they have finite time as well, words used are often an approximation of subjective experiences. The depth of the description will also depend on the subject’s precision in describing their experience. Dennett addresses this issue by saying that all physiological measures and the etiology of false belief can together account for this, and this practice can be backed up by scientific data. But this again is another assumption, and additionally, a description is confused as a feeling. On the matter of taking into a subject’s responses to understand feeling I agree with Chalmers as he says “having a phenomenological belief doesn’t involve just a pattern of responses, but often requires having certain experiences” (or feeling something).
------------------------------------------------------------------------------------
I really like the idea Stevan proposed in class of feeling being a sixth force present in the universe. Studying that force is of course challenging as we are studying feeling with feeling, or rather, like we discussed in class, we do not know how it is to be without feeling.
Certain spiritual traditions hold a similar idea in that there is feeling of consciousness (we can call it feeling) that exists in the universe and is present in all sentient beings. And in that sense, all sentient beings are connected by that same force of feeling, and the distinction we perceive between members is formed by learning and misidentification of our physical body as the boundary of the field of feeling. Our ability to empathise, i.e. feel that “another” is feeling, indicates that perhaps the boundaries of feeling are not limited to the physical body.
On re-reading earlier comments I realise that the fifth force argument has no evidence supporting it. I am curious to know more and know some good readings on the topic.
ReplyDelete10a
ReplyDelete“A bounty of excellent heterophenomenological research has been done, is being done, on consciousness.”
Here, I am skeptical about how such a research could yield to valid results to help cognitive science. There is still a problem of interpretation. Surely the physiological response and the verbal output are something observable across researchers( confirmed inter-subjectively), but as soon as the researchers start to investigate into the meanings of such data, it only pushes the question back to the same old problem of the first-person science. The response is objective, but the interpretation of the response is not necessarily. Some freedom of interpretation is left to individual researchers who are influenced by their own characteristics and experiences. An iced-cold hand and shaky voice may be interpreted as feeling nervous, frightened or simply cold. Which one is the true interpretation of the experience? And also, behavioural and physiological responses have their limitation sometimes in how correctly and precisely can they reflect the true state of the consciousness. For example, sometimes the language could have unintended meanings and understandings to the listener as compared to the speaker. One still cannot know for certain that based on these objective evidence, the result they derive at is necessarily valid.
My thoughts exactly. By creating a system to mitigate the 'first-personness' necessitated by introspection, Dennett inevitably just creates another lens of subjectivity. It seems as though heterophenomenology excuses its inability to obtain objectivity, or truth for that matter (Re: your example of problems with interpretation) for 'neutrality'. The assumptions that Dennett holds for his argument to be true are tenuous at best from my understanding, and as you've articulated, the layers of interpretation required in his methodology only complicate things for the worse.
DeleteWhen replying to Chalmer's zombie Hunch argument, Dennett says: "This speech act is curious, and when we set out to interpret it, we have to cast about for a charitable interpretation. How does Chalmers’ justification lie in his “direct evidence”? Although he says the zombie lacks that evidence, nevertheless the zombie believes he has the evidence, just as Chalmers does. Chalmers and his zombie twin are heterophenomenological twins: when we interpret all the data we have, we end up attributing to them exactly the same heterophenomenological worlds. Chalmers fervently believes he himself is not a zombie. The zombie fervently believes he himself is not a zombie. Chalmers believes he gets his justification from his “direct evidence” of his consciousness. So does the zombie, of course."
ReplyDeleteI find Dannett's argument not convincing at all. First, he uses the weisel word "believe" which completely misses the point on the fact that all we are sure of is that we "feel". He replaces the phenomenal feel described by Chalmers with "believing to have a direct experience". Basically he is arguing for the Others Mind Problem here, saying that because I don't have access to Gabe's T3 internal states or even to my Mom's internal states or here the Zombie's one I have no idea whether or not she has an experience of the world. For him, the Zombie believes he has one so as I do so why not taking it for his word?
He then continues,
"Even if his “phenomenological beliefs” suddenly ceased to be phenomenological beliefs, he would be none the wiser. It would not seem to him that his beliefs were no longer phenomenological."
The problem here is that there is something it feels like to believe. So he is begging the question. True, I may not have access to the Zombie's internal states so I may never know if he feels or not but this has nothing to do with myself or himself believing that we have a phenomenological experience since the simple fact of believing is embedded in the phenomenological experience.
I am for team B. While the reformulation proposed by Turing is interesting and offers another perspective to consider when attemting to answer the Kanting and Cartesian question, it is not an absolute replacement and the answer to it will not fully answer the first two. I agree the reason is experience. Both Kants and Descartes questions clearly ask for answers which involve the human experience. While Turings various machines make attempts at involving this, their effectiveness is still debatable, so we cannot be sure that Turings questions fully encompasses Kants and Descartes.
ReplyDeleteThe interpretation of subject’s beliefs involved in heterophenomenology is clearly subject to some experimenter bias. Already without interpretation of verbal responses (analysis of which often must also be subjective and often qualitative (not numerical)) as a way of analysing study results, there is already an incredible amount of room for researcher influence in studies. From formulating the way that questions are asked, forming the hypothesis to deciding how best to analyse and categorize the data, there is an immense amount of room for human error.
ReplyDeleteI was on the fence about which "team" I agreed with, but as soon as I saw this part in Dennett's paper I instantly knew I was not on his "team".
ReplyDelete"Chalmers must maintain that the zombie’s conviction is false. He says that the zombie doesn’t have the same beliefs as us “because of the role that experience plays in constituting the contents of those beliefs,” but I don’t see how this can be so. Experience (in the special sense Chalmers has tried to introduce) plays no role in constituting the contents of those beliefs, since ex hypothesi, if experience (in this sense) were eliminated–if Chalmers were to be suddenly zombified–he would go right on saying what he says, insisting on what he now insists on, and so forth. Even if his “phenomenological beliefs” suddenly ceased to be phenomenological beliefs, he would be none the wiser. It would not seem to him that his beliefs were no longer phenomenological."
To think that experience has no role in creating the beliefs we have stems from having completely different definitions of cognition, cognitive science, etc. Heterophenomenology seems to be an attempt of creating a "scientific" approach (one with physical evidence as in medicine) to answer something that isn't the hard problem or the OMP. From what I've read in the replies before me, heterophenomenology is just T4, and T4 cannot answer the hard problem or the OMP. If this is the case, then what exactly is heterophenomenology trying to answer?
The whole idea of heterophenomenology is based on refuting this "zombic hunch".
"Of course it still seems to many people that heterophenomenology must be leaving something out. That’s the ubiquitous Zombic Hunch...The Zombic Hunch, then, is Chalmers’ conviction that he has just described a real problem. It seems to him that there is a problem of how to explain the difference between him and his zombie twin."
Dennett himself clearly doesn't seem to agree that explaining the difference between a person and their zombie twin is a real problem, however to be able to understand the hard problem we have to understand what's distinctive about us that the zombie twins are missing. By finding the distinctions, we're able to continue on with Turing's approach to the question, but also apply it to heterophenomenology if that's what Dennett and others want to do. However, I don't think heterophenomenology is adding anything of substance to the conversation. Rather, it seems to be a waste of time because it's not giving the field anything more to think about when answering the hard problem, or any problem for that matter. This isn't based on a feeling that something is missing from heterophenomenology, but rather the lack of physical contribution of something meaningful. (a Dennett style argument)
Even from the beginning of this article (as many mentioned above), I'm not convinced that what Dennett has deemed Turing's question, "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?" is the right kind of thing Turing was even hinting at, nor do I think that even replaces Kant's question. Turing wasn't trying to answer this question, he was concerned with the verbal outputs, not the internal states of the system.
ReplyDeleteAlso, heterophenomenology as a method for empirical data collection does not seem like it will produce reliable enough results. It seems like there's just too much room for interpretation and error with this method: "we devise interpretations of the subjects’ speech acts, which we thus get to treat as (apparent) expressions of their beliefs, on all topics. Thus using the intentional stance (Dennett, 1971, 1987), we construct therefrom the subject’s heterophenomenological world. We move, that is, from raw data to interpreted data: a catalogue of the subjects’ convictions, beliefs, attitudes, emotional reactions, . . .". Even beyond that, the OMP still stands in the way of what Dennett is trying to achieve with heterophenomenology.
"Maximally extended, it is a neutral portrayal of exactly what it is like to be that subject–in the subject’s own terms, given the best interpretation we can muster."
ReplyDeleteI think this was one of the largest problems with Dennett’s complete faith in the heterophenomenological approach. Furthermore, I find it ironic that he mentions it (albeit flippantly) from the beginning. I find it difficult to believe that the felt state of a person could reliably be translated to a speech form (and subsequently in written form). Is this not just a reliance on the introspection of the subject ? As we’ve discussed earlier in the course, introspection being a closed door when it comes to figuring out cognition in general? So, don’t the assumptions of Dennett’s heterophenomenology all tread on these problems? The paper itself seems to me to be quite self-serving, and if I am understanding his argument alright, heterophenomenology comes no where close to solving the hard problem (even though Dennett believes that by dropping the first person, we can potentially do so).
Dennett seems to believe we need a causal mechanism. Dennett thinks heterophenomenology (or, T4) explains something: that it gives us the causal explanation for all of human capacity, including feeling, we have been looking for. He denies the potential of “3rd person science”, and opts for a form of introspection-based inquiry into the causal mechanism. But this is flawed, and here’s why:
ReplyDeleteLike the heterophenominological robots Dennett thinks would be so helpful, we possess T4 (and a little more) capacity. We can make claims about our own internal states (feelings). But, we cannot use this introspeection to describe the mechanism. Otherwise, we would not have the problem we’ve been trying to solve: the hard problem would be soluble.
But where does that leave us?
Both 1st and 3rd person accounts seem to leave us without a clear way of discovering the underlying causal mechanism. Certainly we could continue to attempt to create machines that do the things we do. But behavioural (doing) equivalence is certainly no indication that there is feeling going on underneath (per the OMP).
I strongly disagree with Dennett’s argument on heterophenomenology (answering the question to whether others feel or not). I think it just goes back to the Other Minds problem. If feelings can be measured, we would have to sample people from all ages and all cultures, because there is variability in all of these factors. And if it’s measured does it mean that it is simply inputs and outputs? Therefore, going back to computation and we have already stated that computationalism is far from being the answer to the Hard Problem.
ReplyDeleteRe: Veronica Rusu December 20, 2019 at 6:42 PM
DeleteI totally agree with you! I think it all ties back into cognitive impenetrability, and essentially Descartes’ cogito – you can only be sure that what you think. I think there is also the added challenge of language, where the boundaries of what is considered to be a specific emotion change. Take the example of the German word “waldeinsamkeit” for example, which refers to the feeling of being alone in the woods, given that we have no word for it in English, would that fall into loneliness? Or fear? Or serenity? We can’t even measure feelings in different people and different cultures because we don’t have the same language to refer to them, making boundaries blurry, and even when we do, we can’t be sure that we all have the same understanding and experience with these words, tying back, as you said, to the Other Minds Problem.