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Dennett, D. (unpublished) The fantasy of first-person science.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):
While the idea of heterophemenology is attractive, it seems fairly unrealistic. From my understanding of its methodology, every objective aspect of the human brain that can be measured, must be measured. Any left out detail could confound the data by making it incomplete. Even if we suppose that this wealth of data will provide you with everything you need to know about a subject’s mind, there is no way of testing it (there are simply too many objective measures to record at once).
ReplyDeleteFurthermore, it seems that one’s mental state is frequently changing much faster than can be verbally reported, and can include feelings that can be difficult to put into words. Even if one were able to accurately report his/her feelings, they could have quickly changed over the course of a single sentence. This can cause difficulties in the temporal dimension when trying to make correlations.
Heterophenomenology is just T4. And it solves neither the hard problem nor the other-minds problem.
DeleteHeterophenomenology is T4, it cannot solve the hard problem or the other minds problem as professor Harnad has clarified. The part of Dennet’s article that explains this well is his section about zombies. In the article, Dennet includes Chalmer’s definition of zombie. Chalmers describes a zombie as looking and behaving identically to Chalmers himself. The zombie would be able to report its internal states, be able to do the right things with the right kind of things, and the zombie would even be able to claim it had consciousness. What makes the zombie a zombie, according to Chalmers is the fact that the zombie does not actually have consciousness despite it reporting it does.
DeleteThis shows that even with the use of 'conscious' experiences as primary data, there is no way of solving the other mind’s problem or the hard problem. We cannot simply rely on a subject’s introspective beliefs alone. There is no way to confirm the zombie is (or is not) feeling.
Below, professor Harnad mentions that Dennet believes finding a causal explanation of behavior is also a causal explanation of feeling.
Have we not already refuted this behaviorist idea by determining that even if T4 is built we cannot solve the hard problem?
The sum of the zombies parts does not create consciousness the way it does in Chalmers but there is no way of confirming this.
Hi Nadia, you bring up some really good points. I agree that even if we find a causal explanation for behaviour, that doesn’t imply that we have found a causal explanation for feeling. Even though there are correlates that show there is a link between behaviour/actions and feeling, this relationship is not clear. It has been shown that sometimes feeling even predates the neural firing that accompanies an action, perhaps due to anticipation. So it’s extremely unclear how finding the causal explanation for behaviour could feasibly serve as one for consciousness.
DeleteHaving read the reading for 10b, i.e., Professor Harnad’s responses, Dennett’s paper can be characterized as an argument for a certain methodology. But this methodology, i.e., heterophenomenology, is not unique at all. It does nothing to explain how/why we feel. It only addresses how/why we give “verbal reports”. That is essentially giving a functional account of how/why we do what we do.
ReplyDeleteThree additional problems came up as I read Dennett’s piece.
(1) Re: “We started with recorded raw data.” What does raw data mean? There is no such thing as raw data. I’m inclined to invoke Martin Heidegger’s phenomenology whenever people speak of “raw data”. No rawness at all. We approach things already having encountered them as “ready-to-hand”. We encounter a hammer as something ready to hammer with: nothing raw about it. Accordingly, it is wrong to say “we move, that is, from raw data to interpreted data”. There is no moving. We approach things ‘as they are’ but this means having already understood the things’ purpose in relation to us. Along these lines, John Dewey made a relevant point about experience (read experience as: feeling): “What we experience is shaped by our habits of expectation and there is no basis for extracting from this complex process the kind of ‘thin given’ beloved of sense datum theorists. We experience all sorts of objects, events and processes, and we should not follow philosophers who seek to impose a distinction between the thin uninterpreted data of experience and the inferential processes which lead us to interpret what we experience as books, people and so on. The dichotomy between the passive given of experience and the rich results of our active conceptualization is not supported by our experience. It is yet another of the philosophers’ distortions.” Therefore, raw data deserves the type of vitriol which Chef Ramsay expresses when he often yells “It’s raw!”; it’s hogwash.
(2) What is hetero about heterophenomenology at all?! Having rebuked the senseless distinction between raw and interpreted data, there is no meaning in saying “we construct therefrom the subject’s heterophenomenological world.” From what I understand, Dennett’s methodology seems like a weak attempt to do what Searls’ periscope does, which is to actually become something/someone to check if it has mental states. The problem, for Dennett, is that this is fundamentally impossible. He wants to “do justice to the most private and ineffable subjective experiences, while never abandoning the methodological principles of science”. This seems unjustifiable; one does not get first-person from third-person just as one does not get ‘ought’ from ‘is’, referring to Hume’s guillotine. This relates closely to how we can’t get how/why from knowing where/when things happen in the brain. How is it that third-person verbal reports get us to first-person feeling?? Dennett’s answer seems to be that we can use “the intentional stance”. At first glance, this seems to be an attempt to ‘become’ the other person to ‘see’ what they’re feeling, similar to Searle’s periscope in the sense of ‘becoming’ another. But then he pivots to saying heterophenomenology can “explain the existence of those beliefs”. This is about feeling, as the reading for 10b points out, not about beliefs. Thus, Dennett strays even further than what was initially understood. Feeling will always be privative, exactly the type of ‘thing’ that third-person POV cannot access. You feel. I feel. But I cannot feel what you feel. That’s it. No hetero involved here. To do a play on words, it’s homo (vs. hetero) all the way down just as it’s turtles all the way down (referring to the symbol grounding problem).
In response to (2), I believe the issue you take with Dennett's methodology can be explained in terms of heterophenomenology's agnosticism. While no one disputes the objectivity of skin conductances and other physiological measures including brain activity, subjective verbal report seems to straddle the boundary between subjective and objective measures of consciousness. I think this is because of the function of verbal language in its natural setting is exactly to communicate the contents of a phenomenological subjective world in an objectively accessible way to other people. So heterophenomenologists appropriate this feature of language to access the subject's internal phenomenological world (conscious experience).
DeleteBut then they employ a fatal move to defend their scientific rigour, heterophenomenologists assert their agnosticism regarding the validity of people's beliefs. I think this is a crucial misstep for them. As I pointed out, subjective report as language is meant to communicate first-person data in an accessible third person metric. The contents of subjective report exactly is that first-person data, by being agnostic and mistrusting the contents of consciousness as reported, we are essentially denying their subjective experience already. We have already assumed a priority of physiology over phenomenology. Thus, heterophenomenology because of its agnosticism directly reduces the significance of subjective reports. This is besides the point that first-person experience might not comprehensively lend itself to be communicability.
Austin, "interpreted" or "uninterpreted," the problem with experience is explaining how and why it is felt.
DeleteDan Dennett, when all's said and done, thinks that there is nothing to feeling that T4 does not explain. Yet we all know that there's everything about feeling that T4 does not explain. (He differs from the behaviorists only in that he thinks that to chronicle, predict and control behavior is not yet to explain it: you need a causal mechanism too (T4). But he rejoins the behaviorists in imagining that a causal explanation of behavior (doing) is also a causal explanation of feeling.)
Yi Yang, yes, language is our most powerful means of mind-reading, including finding out whether someone feels something and what they feel. (Not with 100% Cartesian certainty, but good enough for life.) But whether and what does not explain how or why (any more than where and when does).
(If I were you, I would lose the 1st-person/3rd-person talk. It's not explanatory either. Just as the only consciousness is feeling ("phenomenological consciousness") so the only experience is "first-person" experience. All this soothing jargon is just ways of papering over the fact that we have made absolutely no progress at all on solving the hard problem.)
Yi Yang, I feel that the real problem with the verbal accounts is that it is as useless as the mirror neurons in telling us how/why feeling happens. We can have all the input and output we like but still not know what is going on to create feeling. For example, if someone gets rejected from a job application, they might say 'I feel disappointing and worried for the future'. Even with knowing what caused them to feel, and what they feel, we still have no access to knowing why they feel.
DeleteProf Harnad, would a more accurate way of describing what Dennett calls 1rst and 3rd person be feeling and output? The first person would be feeling, because that's really what he is trying to say when he talks about personal experience, and the 3rd person would simply be whatever verbal output behavior the individual produces?
(3) There is no neutrality. What neutrality is Dennett expressing? Heterophenomenology already presupposes a ton of things. First, somehow raw data gets interpreted; relating to my first point, this is not warranted. Second, somehow we can get first-person from third-person; relating to my second point, this is impossible. Third, and the current main point is that Dennett’s emphasis on the third-person scientific method is left entirely unquestioned, taken for granted, and implied as a positive feature per se. I admit that this is potentially beyond the scope of this argument, but given the central role it plays in his argument, he has to at least defend what exactly he means by third-person scientific method. If his idea of the third-person scientific method assumes the subject-object dichotomy (which science often assumes whether justifiably or not), i.e., “The one is that many aspects of our world are independent of us; the other is that that the world is somehow constituted by or dependent upon our conceptual scheme or point of view” (Blackburn), then Dennett needs to take on the onus of justifying how heterophenomenology, which assumes this dichotomy, can possibly then bridge the divide between privative feelings and third-person ‘interpreted data’. He cannot just assume a fundamental divide between the subject and his/her world, and then purport to bridge the divide between the subject and observers in his/her world. This requires some explanation.
ReplyDeleteAs an aside, in my comment for 10b, I attempt to defend having a third-person scientific method to explan feeling. But it’s mostly conjecturing.
DeleteHi Austin,
DeleteI couldn't agree more with your refutation of Dennett's heterophenomenology. In your second point you suggest that Dennet's position seems unjustifiable because "one does not get first-person from third-person just as one does not get ‘ought’ from ‘is’, referring to Hume’s guillotine". I think this is a very important point because it questions the philosophical validity of an assumption central to psychology and the cognitive sciences. Not only can we not know the how/why from the where and but we also can't know the why from the how (given that we figure out the how). The "how" as we have learnt in class also has the subject object dichotomy that you mention in your third point. As we have learnt, if cognition itself is made up of objective processes like categorization (based on affordances) in relation to a referent, we have already assumed the existence of a referent within an already given world. Does the referent cognize or does cognition cause the referent to exist? In the context of the hard problem "why I feel something", we haven't even answered where the subjective "I" comes from. Given so many assumptions, Dennett's heterophenomenology, first has to address the subject-object dichotomy within the scientific method before attempting to collect any kind of raw or interpreted data.
Soham, the hard problem is to explain the causal mechanism of feeling. T3 (or T4) just explains doing. And in explaining doing, T3/T4-scale (the easy problem) it shows why the hard problem is so hard: Because once you have explained all of our capacity for doing, feeling is left as causally superfluous. "Why" is the question of what feeling is for: What does it give an organism that the T3/T4 capacity to do does not already give it. And if you can't have T3/T4 capacity without feeling, why?
DeleteIf I understood Dennett correctly, I think he is saying that there is no hard problem if we take a 3rd person point of view. He says we are a mindless conglomeration of mindless molecules and all of our behaviour is cause and effect. Dennett says that if we can objectively distance ourselves from our point of view we can see that there is nothing to feeling other than cause and effect and we are simple robots that can be engineered. Dennett also says that Turing answered Kant because he only thinks there is only an easy problem. In my opinion, I don't think that compiling data on feelings shows that we can solve the hard problem by solving the easy problem. Dennett also ignores the other minds problem when he tries to objectify feeling.
DeleteI think Dennett does a decent job explaining/defending heterophenomenology in this piece, but I’m still stuck feeling like there’s something missing from this approach to cognitive science. Dennett’s heterophenomenology (in my best effort to define it the same way he does) is a research program aimed at explaining consciousness that involves obtaining both verbal and non-verbal data from people about their subjective experiences, using this data to reconstruct their heterophenomenological worlds (i.e. their subjective experiences), and explaining all of the discoverable patterns in those worlds (i.e. explaining the subjective cognitive states that can arise). It’s an interesting idea, but I don’t really get how you could actually explain consciousness this way. Fodor’s “Why, why do we go on so about the brain?” piece comes to mind for me – even if we were able to use the verbal and non-verbal data that Dennett describes to perfectly, 100%-accurately correlate every cognitive state to a pattern of brain activity, so what? How is electrical activity in a specific neural circuit making me feel happy, or making me feel like I understand Chinese? What is the causal mechanism underlying these cognitive states?
ReplyDeleteOn another note, I found myself somewhat agreeing with his conclusion that “first-person science of consciousness is a discipline with no methods, no data, no results, no future, no promise.” However, I don’t think that just because Dennett’s heterphenomenology has a method/data/results that it necessarily has future/promise, though – I think it just means that the hard problem of consciousness is hard partly because we don’t know how to solve it.
Hello Olivia, I do agree with you that Dennett’s heterophenomology and his attempt to sway others onto his argument is not convincing to address either the hard or the other-other minds problems. Just like you pointed out that having the accurate correlation of every cognitive state to a brain activity couldn’t explain to us the casual functions of feelings.
DeleteDennett concludes that the “first-person science of consciousness (or, in this case it was easier for me to substitute it as, feelings) is a discipline with no methods, no data, no results, no future, no promise” – this seems to be the closest that he came to acknowledging the hard problem, and yet he seemed to dismiss it as a problem each time it is brought up... As he focuses on heterophenomology and that having the easy problem solved is all that is needed for feelings. To which, most of our class (perhaps under some of Stevan's influence) conclusively decides that it isn't not enough to make that leap of faith to at all explain feelings.
Indeed, it's true that we can't solve the hard problem, but we know that feelings isn't a byproduct (or a spandrel, in any case.) Are feelings needed to pass T3? Seemingly, yes, from previous discussion from TT. And in thinking about the hard problem in relation to the discussion of zombie states, we would assume that, if feeling was needed to pass T3, there would be no T3/T4 zombies... (but if there indeed were, then we would have made a bigger mistake than if we had the incorrect causal explanation for doing-capacity).
Dennett presents a metaphorical war within cognitive science, with Dennett’s team representing the idea that the brain is nothing more than a series of cognitive capabilities layered over each other that sum together, while Chalmer’s team claims that there is something beyond the summed cognitive capabilities. This new wave of Dualism that Chalmers presents has no evidence of the “special something” that separates consciousness from unconsciousness, but instead relies on what Dennett calls the “Zombic Hunch”. While Dennett groups in Dr. Harnad with Chalmer’s team, I don’t think Dr. Harnad has ever presented any evidence that he supports dualism or any form of it.
ReplyDeleteAfter looking into Chalmer’s ideas regarding consciousness, I don’t understand why he is so sure that consciousness is ontologically distinct from all other things in the physical world. He readily admits that the systems of the brain are responsible for consciousness (and these can be described and studied via neuroscience), but the consciousness that arises from those physical interactions cannot be understood from their physical substrates. This has the flavor of the inability to get how/why from where/when in the brain, of which I am inclined to agree, but the assertion that consciousness is impenetrable to study seems a little absurd to me.
Dennett, on the other hand, also seems to have a flawed argument. As commenters have said above, Heterophenomenology cannot account for all of the variables that are included in consciousness (as much as it tries to) and the speed of changing states in consciousness is way faster than one can verbally express (and one might wonder whether anyone can accurately describe their own conscious state, no matter how much time they have).
So with both of these ideas’ flaws, it seems that the hard problem is still hard and we are no better off to solve it as before both of these ideas were brought to the table.
The whole discussion on Chalmer's Zombic Hunch seems pointless to me. Why try to derive any conclusions on mere speculations? If such twin-zombies ever exist one day maybe there will be discussions to have and questions to ask about them, but why bother now? We don't know what such zombies would be like, and trying to imagine them will neither provide explanations on why they would be like that (on the contrary, it would beg the question) nor answer any questions heterophenomenalism tries to answer. For someone advocating for a method based purely on data, Dennett relies on a lot of "what if".
ReplyDeleteIt seems to me that what Dennett is trying to do is attempt to explain why intuition is a roadblock/archaic way of determining what you believe in (I understood that part of the passage as Dennett poking fun at Team B for still believing in that old ‘vitality force’ and trusting the very ‘gut’ instinct). And also as you put it, for someone advocating for a method based completely in data, Dennett’s heterophenomenalism seems to be not very strongly supported and I still don’t see how a researcher or team would carry it out
DeleteI’m also struggling to understand Chalmers' argument on Zombie Hunch. He defines the zombie:
Delete“Molecule for molecule identical to me, and identical in all the low-level properties postulated by a completed physics, but he lacks conscious experience entirely… he will certainly be identical to me functionally… and with indistinguishable behaviour resulting… It is just that none of this functioning will be accompanied by any real conscious experience.”
Here, Chalmers has described the “zombie” as a reverse-engineered T4. We know that with T4, we’re faced with the inevitable Other-Minds Problem, where one can never say with absolute certainty whether the T4 (or any other human, for that matter) is entirely devoid of “feeling”. However, Chalmers outright denies that the T4 has any feeling.
“The justification for my belief that I am conscious lies […] in my direct evidence; the zombie lacks that evidence […] The zombie has the conviction that he has direct evidence of his own consciousness, and that this direct evidence is his justification for his belief that he is conscious. Chalmers must maintain that the zombie’s conviction is false.”
In regards to this quote, I am left wholly unconvinced by Chalmers. Though I see myself siding with Team B, that there certainly is a hard problem, Chalmers’ argument is totally unfounded, and is entirely based on assertion. If the zombie has sensorimotor function, then it can gain experience as it interacts with the world. If the zombie has fervent conviction that he has consciousness, then by using Descartes’ justification that humans have consciousness (“I think therefore I am”), surely we can ascribe consciousness to the zombie. Chalmers disagrees. Unless I’m misunderstanding Chalmers’ argument, I find myself asking what exactly are proponents of the Hard Problem hoping for? What more can they ask of a causal mechanism for feeling? If the zombie is totally indistinguishable from a human in function, structure and capacity, and most importantly, holds “[fervent beliefs] that he himself is not a zombie”, how can Chalmers simply justify his assertion that the “zombie” is in fact a mindless zombie?
Hi Amar, I completely agree with what you said. The problems of experimenter bias, ecological validity etc. would make it impossible to get a truly independent accurate report. This is supported by Dennett’s talk about false positives and false negatives as well. Any time the participant over reports or under reports. Additionally, there are certain things that they may not want to share because they feel it reflects badly upon them.
DeleteI also have a problem with accepting this part of the text:
“Once we adopt any such concept of qualia, for instance, we will be in a position to answer the question of whether color qualia shift during change blindness. 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, and we’ll explain why and how they come to have this false belief.”
(I will use feelings instead of qualia to argue this point, because I feel that qualia has too many interpretations.) I find this problematic because while I do believe feelings can change, so for example the color red could make a person feel a certain way when they feel it. However, when they go through a traumatic event that is associated with the color red, they could begin to have different feelings when they see the color red. So feelings can change, however, you cannot say that what someone is feeling is different from what they feel. It indicates that there is a problem in the 3rd person-science that is not recording in detail enough or extrapolating the data incorrectly.
From my understanding heterophenomenology is a research method that attempts to very accurately do what cognitive science has been doing for a while: correlating function with felt experience. But as we know, where/when doesn't get you how/why. As others have mentioned, Fodor's piece comes to mind here. Honestly I don't understand what Dennett is talking about half the time in this paper, it just seems to be circling around the easy problem of function and the hard problem of feeling with lots of equivocation. One of the teams thinks that you can explain away the hard problem, the other team thinks you can't. Am I missing something? Is it any more complicated than that?
ReplyDeleteFrom what I understand, heterophenomenology is less about the brain functions and more about descriptions of qualitative experience (although, I may be wrong as this paper was very dense and difficult). It seems to me that Dennet is unconcerned with the "hard problem", insteading claiming it is not relevant to a science of consciousness. He even ends his paper with:
Delete"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.
To me, this is claiming that a "first-person science" (which he basically claims doesn't exist), is the science of the hard problem and that his heterophenomenology deals instead with the 3rd person scientific approach to consciousness. I don't think Dennet is arguing that you can explain away the hard problem, but saying that the hard problem is not actually a concern for cognitive science, and that you can get the whole idea of consciousness without a 1st person/hard-problem perspective on the topic. However, it does not seem clear to me what part of his argument actually proves this point. As others have mentioned, he seems to be using a Searle's periscope approach but assuming that it does not matter whether first-person consciousness is a part of functional cognition.
Dennett's paper was rather confusing for me and definitely need some clarifications on his ideas.
ReplyDeleteTo start off, I wanted to confirm that my interpretation of heterophenomenology (HP) is correct. HP involves a 3rd person method that includes verbal reports, physiological changes, neural correlates etc. to describe phenomenology and to never abandon the methodological principles of science. Similarly to our past readings on localisation (Fodor), it correlates function with felt experience but does not explain WHY or HOW. Where I'm confused about is regarding to Dennett's response on Turing. Dennett tries to relate it to Turing's question of "how to make a robot that has thoughts". I'm confused as to why HP (describing phenomenology) has anything related to Turing's question of 'why' and 'how' we feel. Is it because the idea of HP is similar to Fodors argument of solving the hard problem? So HP is related to T4 in a sense where you can mind-read what someone else is doing and act as if they are feeling but due to the other minds problem, feelings will always become unexplained. So does that mean it is neither a solution nor an explanation to the other minds problem or the hard problem?
Also, to clarify:
Team A = doesn't have a hard problem. It denies feelings and leaves out consciousness because it just predicts what you feel without explaining how and why you feel anything at all.
Team B = has a hard problem. We feel - there is feeling but does not explain why we "have" to feel.
Unless I’m very wrong, Dennett’s approach/method doesn’t really solve anything, nor does it give us a means to advance our understanding. While I think it’s an interesting concept and could yield interesting data (if we could even collect that much without confounding it, and then have the time and resources to sift through an analyze the sheer amount of data – all just for one subject, let alone an entire sample), I don't see how it answers the problem of why we feel. So in my opinion yes, it is neither a solution nor an explanation to the other minds problem or the hard problem.
DeleteI have a hard time seeing how HP turns first- person narrative into third-person neutral data. Dennett describes HP as the methodological approach to cognitive science that involves all other quantifiable scientific data (neuroimaging studies, biochemical signals, physiological responses etc.) and first person verbal reports to create an HP world for that person. I am unsure how this complete understanding of one person's world as it seems to them would explain cognition? HP seems like a weird mixture of extreme introspection (the HP world being "a catalogue of the subjects’ convictions, beliefs, attitudes, emotional reactions") and case study (how many people could you seriously gather this much information about?). What is the benefit of knowing someone's entire life history and if the when/where can't answer the how/ why question then how could feelings or the output of cognition answer it? HP seems like the classic listing what we already know and doesn't create hypotheses about how it all occurs. HP is said to be neutral because "all the subject's beliefs are bracketed for neutrality." Does Dennett explain what this bracketing is and how it obtains neutrality? How do you know what experience only 'seems' that way to subjects compared to something they actually report/ experience correctly? Isn't making a judgment on judgments not neutral by nature? I agree that verbal reports are important data for studying cognition but knowing someone's HP world is like knowing the entire brain structure/ function and still not knowing how cognition arises from its parts.
ReplyDeleteAlso it seems like Chalmer's zombie is akin to a T3/T4 (a robot with verbal capabilities and sensorimotor capacity) that has passed the Turing Test (a life long test for whether it passes as a human). It would answer or provide more information about the easy problem (how we do what we can do) but does not broach feeling. This is because Chalmer's zombie has the same internal configurations, identical resulting behaviour, and is able to report on its internal states, but does not have any conscious experience of doing so. It would seem that such a zombie or T3 has the how without the why or feeling.
DeleteJust a minor comment here, but it seems that Chalmer's zombie goes beyond being a T3/T4 - it has the affordances of a T5 in that it is virtually identical to a human. The zombie has the same neurophysiological makeup and symbol-sensorimotor capacity. The even trickier thing is that even if the zombie did and does feel, we can't confirm that due to the Other Minds problem! I think the zombie analogy is ultimately a futile one to try to rationalize as it doesn't get us any closer to solving the hard problem (if the hard problem is even soluble).
DeleteI agree that we are sometimes just like robots, but there’s could be difference in terms of the equivalency of input and output between us and robots. That is, whatever underlying robots’ behavior might not be the same as our neural networks that give rise to an epiphenomenon that allows us to think and feel.
DeleteKathryn, I completely agree with your arguments against Dennett. In fact, I find his argument sort of circular. It seems like he is saying that all experiences are subjective (like his example with the coloured blobs) but then is trying to make a neutral account for all feeling.
DeleteZhao, I think that to pass TT the whole point is that there must be equivalency of input and output. However, your second point, that the underlying robots' behaviour might not be the same relates to strong or weak equivalence. As we have seen, Turing didn't care if it was strong or weak when he created the TT.
So a scientist is walking home one night and sees a drunk man staring at the ground as we walks in circles in the area illuminated by a street light. "What are you doing?" the scientist asks.
ReplyDeleteMan: "I'm looking for my keys."
Scientist: "Well is this where you lost them?"
Man: "I'm not sure, but this is where the light is."
Science is sort of like the streetlight. It's an extremely powerful tool, but it's only useful for things you can study with the scientific method. Now, that covers a huge amount of ground, including everything that Dennett talks about, but the Hard Problem is outside of that scope. Dennett seems to be extremely confused about that.
"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."
He takes the above passage as pointing out a flaw in Chalmers' zombie argument, when in reality he's pointing out precisely the Hard Problem. I'm agnostic as to whether such a zombie could actually exist, but the fact that Chalmers and his twin will report the exact same things and act in exactly the same way, and yet none of that information can be used to tell us that one is conscious and the other not. This is not a problem with Dennett's method, which seems thorough enough, it's a problem with consciousness: no scientific tool can identify it, let alone explain it.
There's a further problem as to how Dennett can deny the existence of his own "direct evidence" of his first-person experience, but maybe he's the real zombie.
Dennett seems to say that nothing exists outside the streetlight, which is odd considering that's precisely where he came from.
DeleteThe street-light koan unfortunately casts no further light at all on the hard problem or why it'a hard.
DeleteAs to zombies: The problem of explaining how and why organisms feel, rather than do, is exactly the same as the problem of explaining how and why organisms are not (or cannot be) zombies.
Maybe I did not fully understand the concept, but I am having trouble seeing how heterophenomenology is not just a more advanced version of the Turing machine. While the machine is able to interpret more of what is going on (such as with visual cues like blushing), it does not mean that the machine truly understands, as we humans do, what is happening. Just as with the Chinese Room Experiment, even though the machine and human were equally able to pass the test, neither were able to actually understand what was happening. Is the idea of heterophenomenology not just a trickier version of this test? While the machine would be able to do more and play a more realistic and convincing part in society, I cannot imagine that it would understand that same as humans do.
ReplyDeleteAs Prof. Harnad has pointed out elsewhere, heterophenomenology is just T4. So yes, as you said it is just a more advanced version of the Turing machine. However I'm not sure if I agree with your point that one such T4/heterophenomenology machine/zombie would not be able to understand the same as humans do. I think this is the whole point of the hard problem, we have no idea if it will understand the same as we have no way to see place ourselves in its mind.
DeleteThis comment has been removed by the author.
DeleteI have to disagree with you Lucy. Heterophenomenology definitely seems like a version of the Turing test. However, I disagree with the fact that you think the machine will understand as we do. Only T5 could, which would be an exact replicate of us, down to every chemical and physical structure. If we left it at T3 or T4, I doubt the machine would need feeling to do what it does. Certainly not below T4. There is something else with us, connected to something in how we feel, that will lead us to answer why we feel. The point of the hard problem is not whether we will be unsure if that T4 robot will feel or not. It probably won’t. The hard problem is how and why. If we solve that in a reverse engineered machine, it will ‘understand’ (as you say it, but feel would explain better) the same way as we do.
DeleteThis comment has been removed by the author.
DeleteAs I like to put it, 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.
ReplyDeleteI'm curious about Dennett's choice of phrasing here. He calls both us and our components (our cells) "robots". I can understand his point, about the unfeeling, robotic cells inside of us who themselves are not conscious, yet together act to help us do what we do. However, when he says that we are robots as well, he places us in that same description and level as the "mindless molecules". It seems to me that giving human beings and cells equivalent descriptions of roboticism is flawed, as there is clearly a difference in consciousness.
I agree, I think that he is putting the cart before the horse here. We are made from robot parts, which are put together in a way we don’t understand. The underlying process is robotic, as far as we understand. Yet we experience consciousness – and it feels like something to experience it. This is a process that is not exhibited by other robotic things. Therefore, consciousness is a process we haven’t accounted for. The idea of ‘mindlessness’ exists because we experience having a mind, and can see things which do not appear to have one as well. It may be finicky language play, but how can we be mindless robots if we (by definition) have a mind?
DeleteHi,
DeleteI also thought that exert was a bit strange and meant more to play on popular Man = Machine debate then critically think through any statements on consciousness. I suppose he is trying to express that consciousness can arise from robotic parts by virtue that conscious states can be embodied by robots thus allowing us to study it. "We are robots made of robots" is a piffy way of trying to communicate that mechanical things can come together and form a mind.
I think by making this statement that "we are robots made of robots", he is again ignoring the hard problem. I believe he is aware that we are different from our robotic cells as we are conscious and have a mind, but by labeling us as robots he reduces this ability of ours to be something that can be replicated in a robotic machine, like Cassie said. This to me is representative of his viewpoint that there is no hard problem and feeling and consciousness can be solved by looking at causal mechanisms, making it something that is thus able to be generated in robots as well.
DeleteI think there's another underlying issue with this as well. As many neuroscientists have said, we have a fairly good understanding of many details regarding the functioning of the brain, but we don't have an overarching theory of brain functioning as a whole. Similarly, we have a very detailed understanding of how living things operate in their parts, and though many people use this as a reason to say that we know 'understand life', it's not so obvious that we indeed have a coherent theory of what living things are as wholes. In both cases, this is basically the issue of autonomy, and what constitutes an autonomous system. Many people (including prof. Harnad it seems) take the question somewhat for granted. Others argue that it's argue at the heart of the issue of understanding the mind. So yes Dennett is right to say that we are made of robotic parts, but those are not autonomous. The question is then how non-autonomous parts can combine to form an autonomous whole?
DeleteDominique, I think that the point of this particular phrasing of “we are robots made of robots” bears too much resemblance to computationalism to ignore. Each of our cells is a “robot” in the sense that it can be boiled down to computation, just like the heart can be symbolised with computation. But as we’ve repeatedly seen throughout the course, cognition cannot merely be computation and that’s where Dennett’s statement seems to break down. The components we are made of are robots, yes, but there is something else going on here. Dennett’s denial of the existence of the hard problem is precisely the denial of the existence of this “something else going on here”.
DeleteI had trouble following certain parts of this article. For one, i'm a bit confused on what heterophenomenology actually is. From what I can tell, heterophenomenology is just taking a 1st person's subjective account of experiencing and turning it into a 3rd person objective account. At first, this seems like a good idea - objectivity is the basis for science. However, I'm having a hard time understanding how experience wouldn't be necessary for explaining consciousness. It seems the experience is the basis of consciousness. Can someone fill me in on what I'm missing?
ReplyDeleteAlright, Dennett is a great and funny writer and I want to give him where credit is due. That being said I think he gives himself way more credit than warranted. From what I understood, heterophenomenology examines subjective experiences in an anti-armchair fashion; incorporating the different aspects that make a method empirical. I do not deny that his method adheres to scientific principles, but I think it simply examines the mental by sophisticated means. It seems like Dennett feels he has won the battle with Chalmers by virtue of presenting a scientific method simply because “First-person science of consciousness is a discipline with no methods, no data, no results, no future, no promise. It will remain a fantasy." What I mean is, even if first-person science of consciousness has a methodological flaw, heterophenomenology does not automatically elucidate consciousness by addressing this flaw. I would feel differently if I was convinced this heterophenomenology explained consciousness…but I don’t.
ReplyDeleteI agree with you. I think Dennet spent too much time and energy discounting 1st person accounts of science of consciousness and uses the same several examples of how heterophenomenology is applicable to consciousness, without actually explaining what it can get us in the end. I understand it could possibly explain some aspects of consciousness, such as blindsight, but I'm not sure how this can be extrapolated to shine light on anything new about consciousness or the brain that we don't already know.
DeleteDenette posits that verbal interpretation is a fail-safe mechanism of sorts to protect against the possible fallacies of conscious experience taken as raw data. Furthermore, he makes a distinction between what someone reports to be feeling (judgement, verbal), and what (s)he is actually feeling (conscious experience). The example he makes of the chemical compounds reminds me of our debate about whether where/when can explain how/why, and also why it is easier to reverse engineer a heart than a brain. His explanation seems to support that we can get how/why from where/when since Ramachandran and Gregory were able to predict this novel and subjective phenomenon from a model of how vision works. He says that heterophenomenologists would 'grant' that some parts of experience are ineffable. But how does their proving that some aspects of experience are unable to be described, support their theory of the primacy of heterophenomenology as the means to stud consciousness?
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteI'm finding myself struggling with two different aspects of Dennett's article.
ReplyDeleteThe first, is that Dennett seems to find flaw with first-person science but not heterophenomenology. He seems to simultaneously equate first-person science and heterophenomenology while concluding that "the first-person science of consciousness is a discipline with no methods, no data, no results, no future, no promise."
I think a good chunk of the debate between the two disciplines lies in the distinction between Chalmers using "subjective experiences" as his central data while Dennett focuses on "subjects' beliefs of subjective experiences". Perhaps I'm simply struggling with the material but I'm failing to find a substantial difference between the two concepts. At the end, you have one person explaining their subjective experiences. Is not a subjective experience inherently what one feels to have experienced or seen?
Additionally, perhaps this is a sign that I'm spending too much time in science-critical disciplines but is there not a subjectivity that accompanies all forms of 3rd person science. So, Dennett's claims that third person science will be free from the subjectivity of the person experiencing it seems flawed to me. A caveat - this is a problem that all of science deals with and I recognize that it's not necessarily 'fixable' in either instance, it merely exists to varying degrees. I do take issue with Dennett not noting that even the analysis of the heterophenomenological data can be subjectively skewed by other humans actors.
**previous comment deleted due to spelling errors. this is an edited re-post.
I think the beginning of this article allows for some confusion between 2 uses of the word “experience,” the first being to receive input from one’s surroundings, and the second being to have feelings (e.g. being aware of a thought, sensation or memory at a given moment). It’s important that these not be confused, as I don’t think the ability to receive input from the world would entail that a being has “experiences” (feelings).
ReplyDeleteI also think that Kant’s question is somewhat vague— “what are the conditions necessary for the possibility of experience […] at all?” The question does not differentiate between a necessary and sufficient condition. If we are talking about a necessary condition, then I think Turing’s question is an acceptable replacement. Surely no being can have a thought (of theirs) without the ability to have thoughts, and learning from one’s surroundings is something that I think would be required as well— after all, thoughts are not self-creating and need to be based on something. However, there is no reason to believe that the presence of thoughts and experiences is enough to produce feelings (although it might be). If this is what’s being implied by Turing’s suggestion, then I don’t think his question does very much to help in answering Kant’s, since it doesn’t provide any way of ensuring that the robot would have “experience” in the feelings sense. Therefore, although Turing’s question is a good method of answering how people are able to do what they do, it does not bring us any closer to answering why we feel.
As for heterophenomenology, it could maybe be a way to confirm the presence of feelings, if we were ever to produce a robot that had them— we could do this by comparing the correlations of reported feelings and other outputs (like a polygraph) to determine whether they were consistent with the reported feelings of humans. Although we still couldn’t be certain about the presence of feelings, it may be the closest we can get to surpassing the other minds problem for a robot. However, in terms of answering the question of why and how we feel, the correlational information from heterophenomenology would not, on its own, provide an explanation.
This comment has been removed by the author.
ReplyDeleteHeterophenomenology seems to be premised upon the fact that we can scan and observe absolutely everything about the act of consciousness. Through taking this ‘3rd person scientific perspective’ on cognition, this field establishes a “total, dictatorial authority over the account of how it seems to you, about what it is like to be you”. However, this seems a bit overreaching to me – especially for someone claiming to be completely agnostic as to conclusions about conscious experience, and whether (or not) it exists. If I’m not mistaken, the ultimate direction that he takes his argument is that we do not actually have free will, it just seems that we do, and that what seems to be choice is just neurological processes. Heterophenomenology, then, is what we should use to examine this seeming, and thus understand ourselves better.
ReplyDeleteIt strikes me that we should not rule out free will on behalf of a technology that does not exist, and we should not disbelieve people who claim to be conscious because that same technology can predict their behavior. That being said, I think that his argument is sound, that a sufficient heterophenomenological investigation might provide an understanding of the brain which seems to be determinate. If I know everything (and I mean everything) about someone, I could predict their behavior. Does this mean they do not have free will? In the sense that free will is an unconditioned choice amongst unlimited choices, yes – but this requires a power to completely constitute one’s own reality. In the sense that it feels like I make choices based on a situation, and that I have flexibility to choose or not to choose, I don’t think it does. In a given situation I will make a certain choice, but it feels like I make that choice because I decide to be the sort of person who makes that choice in that situation. If I had a reason to be otherwise I could be, at least hypothetically speaking. His prediction for a future of deterministic technology presupposes that we are fundamentally right about everything physical, chemical, and existential. If we understand the world as a truly determinate place then of course we are determinate – but can we truly make that claim with certainty?
Dennet seems to believe we are automatons, and simply experience ‘choice’ and ‘feeling’ as an emergent illusion, for whatever mysterious reason. I cannot disprove him, but it does strike me as a non-agnostic point. I think that a deterministic conclusion in this case overestimates our understanding of the universe, ourselves, and our capacity to understand these things.
“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 am confused as to how the zombie can get justification from its ‘direct evidence’ since it does not have access to conscious states or consciousness in general. As noted earlier, “Chalmers allows that zombies have internal states with contents, which the zombie can report (sincerely, one presumes, believing them to be the truth); these internal states have contents, but not conscious contents, only pseudo-conscious contents.” So if there is only ‘pseudoconsciousness’, how can the zombie use its consciousness as direct evidence in the same way that a human being like Dennet does?
Heterophenomenology as Dennett explains seems an interesting theory, but I don’t understand how it would work in practice. While he proposes a method to describe subjective experience/consciousness in a scientifically rigorous and objective way, how is it any more than an idea that work well only on paper? “All behavioural reactions, visceral reaction, hormonal reactions, and other changes in physically detectable state are included within heterophenomenology” – such a vast amount of data, countless hours of footage to view and then analyze and then interpret, I don't see how this approach could ever make sense. There is the argument of restricting oneself to only a certain subset of the physically state but that just counters the whole point of this method – to take everything in its whole. While it could be very useful, I don’t know how if could be put into practice. Finally, I still don’t see how this explains how we feel things, apart from having more causal things to point at.
ReplyDelete• I agree, I don't see how "heterophenomenology maintains a nice neutrality". How is it possible that there is no passing judgement when investigating beliefs? In a human world, I think it is impossible to separate the notion of judgment from investigation. It seems like the author is trying to argue for the idea of using the speakers' truth of experience and not imposing a moral truth (whether an experience actually happened), i.e. if someone believes they saw something, it is their reality although it is not an absolute reality. This is applicable in a wider context as the reality that two people will experience in the exact same circumstance will differ. I would have to side with Levine in "that conscious experiences themselves, not merely our verbal judgments about them, are the primary data to which a theory must answer." Dennett's argument, although logical, seems to just circumvent the hard problem (i.e. how/why we can do what we do).
DeleteDennett’s description of heterophenomenlogy is that it captures more than explicit verbal and action output, but “every blush, hesitation, and frown, as well as all the covert, internal reactions and activities that can be detected, are included.” Like Stevan says, that heterophenomenlgy is merely a part of T4 and passing this would only answer the easy problem. But, from the article, it seems that is all Dennett believes – thinking that the problem of doing (easy problem) is the only problem.
ReplyDeleteIn Dennett said that, “heterophenomenlogy… doesn’t leave out any conscious experience you know of, or even have any first-persons inkling about.” It really shows that he believes it to be all-inclusive into T4. Almost seemingly that he is down playing or just outright denying the hard problem of how and why do we fee and explaining the causal function of feeling. By focusing on what is observable either internal computations or external dynamic aspects (T4), Dennett maintains quite a behaviouristic approach towards feeling.
I don’t find his arguments quite convincing and I felt that it became quite circular and almost confusing when he started to discuss about beliefs; however, in the end, it just seems to be another weasel word for feelings. Because, ultimately, beliefs are felt as well (the unfelt beliefs merely seems to be just states and conditions).
In Denett’s video, he poses the question of whether we can have access-consciousness (easy problem, the “doing”) without the phenomenal consciousness (hard problem, the “feeling”). I wonder if blindsight would be considered to be the alter ego of such a situation, ie. having the feeling without the awareness. But, coming back to Denett’s question, I wonder what he would think of people with Cotard’s syndrome would be an example of having A-Consciousness without the P-consciousness. Indeed, the people with this disorder do know that they are conscious, but they don’t report feeling anything, and will report having the delusional belief to be dead. The absence of feeling is so strong that they will deny their self-existence. Is this then an instance of having only access to 3rd person information, without the 1st person’s?
ReplyDeleteRE: “I figure that Turing’s genius permitted him to see that we can leap over the Zombic Hunch”
ReplyDeleteAccording to Alvin Goldman, Turing was not only a computationalist, but also believed that feeling (or as Goldman phrases it: “intuition/Zombic Hunch”) is an unimportant by-product of cognition.
The purpose of the TT, however, was to provide a causal mechanism for performance capacity. Since the other-minds problem (hard problem) made parts of cognition invisible, the TT focussed on parts of cognition that are visible (doing). In no way, though, did the TT seem to imply that in explaining “doing” one need not explain “feeling”.
Dennett’s heterophenomenology (HP) is an attempt to establish an objective scientific method for the study of subjective consciousness. While Dennett’s HP may seem like a plausible means to quantify mental processes in theory, I don’t understand how it would play out in practice. For one, the use of verbal accounts seems restricting in and of itself. Though I agree with Dennett that speech is a medium which, by its nature, is multi-dimensional, encompassing all of one’s emotional subtleties, intonations and first-person conscious accounts, speech lacks a communication of subconscious accounts or aspects of experience that may be difficult to express in words. Moreover, Dennett denies the existence of the hard problem since there is no “feeling” (or qualia). My confusion pertains to Dennett’s argument about third-person qualia. Dennett says “the price you have to pay for obtaining the support of 3rd-person science for your conviction […] you have to grant that what you mean by how it is/was with you is something that 3rd-person science could either support or show to be mistaken”. For instance, “if some subjects in our apparatus tell us that their qualia do shift, while our brain scanner data shows clearly that they don’t, we’ll treat these subjects as simply wrong about their own qualia”. How can a subject even be wrong about their own qualia if qualia are inherently subjective? Here, the concept of studying consciousness in a 3rd person objective way seems counter-intuitive. Isn’t the point of studying consciousness to study one’s subjective perception of (or beliefs toward) an experience? If we discount a subject’s perception of his/her experience, are we still even studying consciousness or just another objective experience? I agree that one’s subjective perception of an experience can be objectively wrong, but shouldn’t the study of consciousness include the inherent messiness that consciousness involves?
ReplyDeleteI am left to wonder where Dennett’s confidence in grouping himself and other pygmies as being part of the A team (ironically places himself as the captain), while calls other interpreters as belonging to the B team led by David Chalmers. The arguments presented in class for the hardness of the Hard problem is left unaddressed in Dennett’s attempt as a pygmy to resolve it to mere heterophenomenology. As discussed in a previous reading by Fodor (1999), the raw data with which Dennett swears by is neither raw nor data which supports Dennett’s view. Every aspect of the brain that can be recorded, as presented by Fodor, has no causal explanation, but correlation (i.e., brain regions activated during a certain behaviour). No data is raw data. There is nothing objective about the rawness in which Dennett relies on to formulate his heterophenomenology.
ReplyDeleteI believe that through the long academic life, Dennett has forgotten the most foundational idea of science—science is another representation of the world (not an equal phenomenon; just a simulation in the terminology of our class).
“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.”
On the following statement, I would like to suggest that efficiency does not always allow advancement in our endeavours. One cannot begin to “get on with the business of accounting for these subjective beliefs” if one cannot firmly grasp the question being asked. I would even say that this is childish on the basis of not understanding the orderly fashion of procedures.
“I will not contest the existence of first-person facts that are unstudiable by heterophenomenology and other 3rd-person approaches.”
Also, exactly that unstudiables by heterophenomenology is what we are trying to answer. So what does Dennett mean by the unstudiable?
“First-person science of consciousness is a discipline with no methods, no data, no results, no future, no promise. It will remain a fantasy.”
I can agree with Dennett on the first-person science of consciousness as having “no methods, no data, no results.” However, I would not go so far as saying that it does not have no future and no promise.
Exactly, no data is raw dat a and at the end of the day we're working with simulations. I'm still trying to wrap my head around an article which embedds the basic structure of it's argument in what the field itself fails to establish. Sometimes I really do wonder if the priority is to report scientifically accurate information or to ego boost one's self in the field of Academia by pitting themselves against others. Rather than work toward more explanatory models, of more accurate simulations we see bickering between individuals, and grouping into A and B categories with intellects who may very much disagree.
DeleteDennett’s paper is incredibly difficult to follow and a little bit of a drama to read. His theory of heterophenomenology is still pretty unclear to me, but the take home point (for him) is that first-person report is enough to validate a proper study of consciousness. From what I understand, the reported difference between objective facts in the world, and the ability to alter them through perception is in a way a valid report of perception. This is the method of heterophenomenology and is compared to what anthropologists do, of interviewing informants so to later investigate it against the truth. This is old news for anthropology, post-structural anthropologists today actively work against this tendency to compare report to any objective truth. Anthropologists don’t actively seek to compare against any objective truth because it has done nothing for the field in the past but motivate elitist/colonial projects. This is what makes Dennet’s piece dramatic; he entitles his point as valid where claiming his theory as rightfully doing as others do. But they aren’t, and maybe neither should he. I don’t disagree with objective facts, but using them (when they aren’t even facts) as a tool to validate one’s own theory makes for a sassy self-affirmation.
ReplyDeleteHi Krista,
DeleteI agree with you (and most people int he class it appears) that heterophenomenology was poorly described in this paper. I understand it as an attempt to use third (not first) person reports to study consciousness. The B team argues that first person accounts are the only way to understand consciousness (ie. through feeling), but he believes that heterophenomenology breaks that wall and is sufficient to getting a snap shot of someones conscious state. I think we disagree a bit on the idea of objectivity however. I think that modern anthropologists are realizing that they weren't doing objective human science because they were comparing it to the subjective white male colonialists. I think that any science of the human needs to compare against some level of objectivity, I think specifically anthropology needs to work on what that means within their respective field. For cognitive science however, finding an experimentally verifiable method of study the mind is extremely important. I like his argument in that he wants to make a science of consciousness, which is really the only way to establish causality. But you're right in that we need to think about what self-report, environmental simulation, brain imaging (or whatever heterophenomenology is measuring) are able to provide as objective measures of an inherently subjective thing (consciousness).
ReplyDeleteIn most cases, I see the flaw in verbal self reports, however, this paper didn’t make clear to me how heterophenomenological data can be collected in a neutral, third person manner having been untouched by the subject’s inevitably biased perception of what it is they are feeling.
I may be misunderstanding the reading, but to me, this idea of Dennet’s sounds like ivory tower philosophy. Surely cognitive science would benefit from such data, and prosody, body language, heart-rate, hormones, and facial expressions are credible data that we often leave out in psychological research, but but how exactly would heterophenomonology work? How exactly could we collect data pertaining to the subject’s heterophenomenological world in a neutral, empirical, objective manner?
I also take issue with the complete denial of the value of first person science: “First-person science of consciousness is a discipline with no methods, no data, no results, no future, no promise. It will remain a fantasy.”
I am a big Sam Harris fan, an avid listener of the Waking Up Podcast, and have become convinced that mindfulness/ meditation is a valid, valuable, 1st person science. It has taught us that it is possible to manipulate one’s attention and separate the self from the mind, through following a methodology refined over thousands of years of practice and teaching. While there is still a lot of work to be done to look at the practice more objectively, FMRI studies find striking differences between when the mind is detached from the self, and when it is not. I realize that Dennett's arguments are mostly targeted at Chalmers' claims, but I think that mindfulness/ mediation and the 1st person science that developed it is valuable to cognitive science, and shouldn't be reduced to 'fantasy.'
I'm having some trouble following Dennett's arguments here, but it seems that his standpoint is inconsequential to the course as we have progressed.
ReplyDeleteDennett's initial rejection of a science solely based on introspection follows our class trajectory (true or not), but after that his paper loses me.
The proposed heterophenomenological method of evaluation paradoxically claims to "account" for those subjective experiences untouchable by scientific data or inquiry, and yet offers no explanation of how to bridge the gap between functional correlation and the how/why, much less the subjective (first person) and objective (third person).
Perhaps Dennett truly considers the hard problem to be unimportant, but his jumbled argument is self-contradictory regardless.
Had a hard time understanding the heterophenomenology, but no matter what, I don’t think it is the right method to know the how/why of our feelings, or consciousness.
ReplyDeleteRegarding [Harnad, S. (2014) Animal pain and human pleasure: ethical dilemmas outside the classroom], this blog by Professor Harnad wrote the example that “(they) have no way of knowing that it would not hurt that person, even if they were a robot. I then point out to them that they have no way of knowing with one another either, and that that’s the whole point of Turing indistinguishability.”
This reminds me that feeling and consciousness are not testable by a third person and that is why I doubt that heterophenomenology will be able to bring us to the solution of knowing why/how we feel. And it makes almost no sense by recording vocal sounds of people. The way we say or we produce any vocal output does not necessarily show what we are consciously thinking though. And if measuring brain activity and heart rate work in knowing what the person is thinking, then we should have already solved all the problems in cognition.
The only remaining alternative, C(2a), is unattractive for a different reason. You can protect qualia from heterophenomenological appropriation, but only at the cost of declaring them outside science altogether. If qualia are so shy they are not even accessible from the 1st-person point of view, then no 1st-person science of qualia is possible either.
ReplyDeleteThe hard problem will not be solved from studying qualia alone. This would just be a form of introspection. In addition, his example isn’t really an exclusion of qualia. By definition they are an example of a perceived stimulus, or change in stimulus. The inaccessibility described at the beginning was not a shy qualia, it was simply a lack of one altogether. There was no perception of the white door flashing, therefore there was no qualia.
RE: “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 see this as highly problematic because he moment you have someone speaking – their speech can be put into the same constraints and categories that behavioral data or physiological data can be put into. What may not be observable, knowable or analyzable is what it feels like to the person to say the things they are saying. But how is verbal data any more or less explanatory than other correlated data like physiological changes or brain scans? Dennet’s argument that “heterophenomenology” provides a causal mechanism to solve the hard problem has no grounds. And what does he mean by: “Some beliefs that subjects have about their own conscious states are provably false, and hence what needs explanation in these cases is the etiology of the false belief.” How is “heterophenomenology” meant to explain objectively why some beliefs are false, when it cannot even explain beliefs i.e. feelings?
I see a similar issue of the what/where explaining the how/why. The most that verbal communication can achieve is showing what someone is feeling, and is just the same as saying the verbal description of a feeling may be correlated with the feeling. Dennet’s heterophenomenology does not show why or how they feel.
In an attempt to quantify subjective consciousness into an observable study by way of the scientific method, Dennett discusses his idea of heterophenomenology in this paper. However, there is a recurring flaw in that subjective experiences cannot be studied in an objective manner; put simply, it is impossible. Consciousness is only accessible by the entity that possesses and experiences it. Subjective perception of an experience may be different from what is observed objectively but this does not mean that subjective experience can be easily denied. There is no such thing as neutrality in heterophenomenology. Subjective experiences are impenetrable by scientific reports and data, and so first person and third person experiences will never completely overlap. This topic does confuse me but what I take away from it is that heterophenomenology will not provide the causal mechanism to which the hard problem is solved, not even close. How and why we feel is completely private to us as cognizing beings and is only available through first-person access. Only oneself can have phenomenological beliefs and this kind of experience will never be explained by third-person science. Not even something as powerful as language can bridge the gap between first person experiences and what is observed from a third person.
ReplyDeleteHi Neil!
DeleteI agree with the conclusions you have drawn. Dennet believes that “first-person science of consciousness is a discipline with no methods, no data, no results, no future, no promise” and therefore heterphremonology would provide better data to help us to answer questions about our cognitive states. But this still would not answer the hard problem of why and how. Correlational data is not causation and even if we use verbal and non-verbal data and find a correspondance to different cognitive states, this does not become too much better than what first-person science contributes to solving the hard problem and the other minds problem.
When Dennett says “He insists that he just knows that the A team leaves out consciousness. It doesn’t address what Chalmers calls the Hard Problem. How does he know? He says he just does. He has a gut intuition, something he has sometimes called “direct experience.””
ReplyDeleteI found the term “gut intuition” striking. Dennett makes it out as if “direct experience” is hunch. He undermines one’s feeling of existence to what? Possibly illusory? , yet he goes on to talk about heterophenomenology which, in itself, requires ones feelings of existence. I don’t understand how he can deny the importance of the “gut feeling” yet base this entire practice off the feeling. You need to agree that you are experiencing something to report about it.
RE: it involves extracting and purifying texts from (apparently) speaking subjects, and using those texts to generate a theorist’s fiction, the subject’s heterophenomenological world. This fictional world is populated with all the images, events, sounds, smells, hunches, presentiments, and feelings that the subject (apparently) sincerely believes to exist in his or her (or its) stream of consciousness 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.
To use terminology from class, this method is a form of “Cognitive technology” it is explicit rather than implicit, revealing nothing but the tip of the iceberg. Dennett claims that heterophenomenology relies on more than just the third party data, including scientific measures of internal states. However, this to me is just irrelevant fluff (#Fodor). The third-party reporting is irrelevant to cognition. Once feeling becomes language it is no longer felt (cognitive technology) and therefore no longer related to one’s internal state, conveying no isight to the problem at hand (cognition).
RE: On the one hand, 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.
This statement is kind of what I purpose in class often. It hints at the fact that you need a “feeler to feel”, there needs to be that I experiencing inside. However, I disagree that you need to label or verbalize internal feelings to have a conscious thought . There is more to cognition than internalized language because there is internalized feeling. There is something it is like to be a human without said human having language capacity (That was a big statement, that I’m not sure I 100% agree with).
Part f the problem with cognitive science, as I’ve said a million times, and I will say it a million more, is we do not have a definition for cognition. I say this here because we agree that chimps are cognizing but are they cognizing as humans do, or do humans have a different "cognition" because they have language? Cognition/feeling/consciousness is so elusive. Steven said in class that once an animal has a central nervous system it has feeling, but, so, does a drosophila cognize? Feeling is necessary for cognition but it is not sufficient? Are feeling two totally separate entities cognition separate? We keep singing in circles, with the last note so far from reach...
This is the song that never ends, it just goes on and on my friends, some people started singing it not knowing what it was, and they'll just keep on singing it forever, just because this is the song, that never ends, it just goes on and on my friends, some people started singing it not knowing what it was and they'll just keep on singing it forever just because this is the song that never ends...
Qualia are the felt states and they are subjective. It feels like something to see the color red for me, for you probably too, and it will be different between the two of us. I completely disagree with Dennett’s statement that “if some subjects in our apparatus tell us that their qualia do shift, while our brain scanner data shows clearly that they don’t, we will treat these subjects as simply wrong about their own qualia…”. You can’t falsify someone’s personal experience with a brain scanner. What we feel/think/experience might not be accurate all the time, as in illusions and the color blindness, but nevertheless that’s a part of our qualia. That is what we believe we experience. Thus, sometimes, a 1st person point of view can tell us more than a 3rd person point of view. Saying that the 3rd person point of view does not make us abandon science altogether and declare that qualia is outside science. However, we might want to consider the fact that we do not have the technological knowledge and equipment yet to study the most elusive and possibly the most important state of human beings, which is feeling, or qualia as Dennett calls it. Something so subjective will be impossible to explore completely by 3rd person data. There would always be an element that we are missing. The experience is more than what we can explain verbally. Right now, our best hope is through reverse-engineering. 1st person is of course not enough as introspection is not a good way to study cognition and feelings.
ReplyDeleteThere is one point that Dennett makes, which is hard to refute completely. He says, “… the zombie believes he has the evidence, just as Chalmers does.” We can only be sure of our own mental states. However, one can never be a 100% sure, though it goes into extreme skepticism. I personally do believe in cogito. However, I do not think that we can always use it as an argument as you would not believe a robot saying it feels, just as you would not believe Dennett if he said he feels.
“Kant: How is it possible for something even to be a thought (of mine)? What are the conditions for the possibility of experience (veridical or illusory) at all?
ReplyDelete(…)
Turing’s great contribution was to show us that Kant’s question could be recast as an engineering question. Turing showed us how we could trade in the first-person perspective of Descartes and Kant for the third-person perspective of the natural sciences and answer all the questions ¬– without philosophically significant residue”
I really disagree with Dennett here. Kant’s question (as Dennett formulated it) addresses thoughts and experience – something internal. In other words, Kant’s question refers to feeling – the how and why of feeling. I don’t think that Turing and the Turing test answered this question at all. I don’t think Turing even tried to answer the question – He dismissed attempting to answer “thinking” directly. A machine that passed the Turing test would perform indistinguishably from a human, but Turing did not claim the machine would feel, that it would give us knowledge what it’s like to be a machine, or that it would allow us to understand the how and why of feeling, nor did he claim that feeling didn’t exist. Something that passes the Turing test would serve as a possible mechanism of cognition, but it definitely doesn’t “answer all the questions.” The Turing test can only measure performance capacity, which we can use to measure cognitive ability (and thus see if we have successfully reverse engineered cognition) - But it gives us no information about the how and why of feeling and of experience. That “question” is unanswered by the Turing Test.
However, again, I don’t think Turing intended the Turing test to “answer all the questions”. I think Turing may have realized that trying to answer questions about how and why of feeling was futile, and performance capacity is what we should be concerned with because that’s what we can measure and observe – Not that feeling is nonexistent or answered by the Turing test.
RE: A team vs B team
ReplyDeleteI like how Dennett clearly summarized the two sides in a way that was simple and easy to understand. The analogy of humans as robots that consist of robots made me think back to the Turing statement above. If humans can experience things, and humans are robots, does that explain part (B) of the reaction to Turing’s proposal which argues that Turing is leaving out “experience?” If humans experience things and humans are composed of these molecules we are also calling robots, does this mean that these molecules can experience things as well? On the other hand, the big issue that the B team has is that the A team does not take into account consciousness. I find the idea of the Zombie Hunch very interesting. I don’t think I quite understand it fully but I think that it has something to do with what you experience in a certain situation?
There exists the distinction between the private, entirely subjective ‘first-person’ dimension where consciousness begins and the objective ‘third-person’ methodology as described by Dennett in this article. Essentially, heterophenomenology (similar to the philosophical existential practice) brackets off this first person perspective to allow it to be interpreted empirically. Dennett feels that this practice can define all that can be known about the mind and consciousness. In my understanding it seems that Chalmer’s side that argues the about the zombie etc, seems to claim there is more to cognition than the sum of its parts, which bordered into dualist territory where the nature of consciousness cannot readily be explained by its physical components. While I understand that the hard problem is in fact hard … this does not give grounds to dismiss the brain and its components as the causal mechanism for feeling. Dennett's argument remains flawed, but in my mind is perhaps a more optimistic view than the eternal regress that emerges out the concept of a 'zombie' as explained by Chalmer.
ReplyDeleteI’ve noticed that a lot of students are asking questions about heterophenomenology, so I’m going to take a stab at explaining it. From what I understood, the main issue that Dennett is trying to address is the fact that first person experience are central to the study of human consciousness. Because they are subjective experiences, they become difficult to study using conventional scientific methodologies, which Dennett refers to as third person objective methodologies. In essence, Dennett is saying that we cannot study human consciousness using the same methods that we do for studying gravity, for example, because of the subjective nature of consciousness. In order to address this issue, Dennett comes up with heterophenomenology – phenomenology of another not oneself. From what I have gathered from this reading, as well as others, heterophenomenology simply consists in interpreting the vocal outputs of another person. In doing this, you can create a catalogue outlining the subjects beliefs about their own conscious experience. According to Dennett, this is the best way to obtain first person data, while staying within the realm of the scientific method (because introspection, a subjects own report of their beliefs, would not conform to this method). In essence, heterophenomenology would allow us to build a representation of the persons subjective world, through a third person interpretation of their own account. To be honest, I’m not sure how much of this I agree with, but I think that sums it up! I also found this quote from another one of Dennett’s articles that I think would help to conclude the explanation.
ReplyDelete“The total set of details of heterophenomenology, plus all the data we can gather about concurrent events in the brains of subjects and in the surrounding environment, comprise the total data set for a theory of human consciousness. It leaves out no objective phenomena and no subjective phenomena of consciousness.”
https://ase.tufts.edu/cogstud/dennett/papers/JCSarticle.pdf
While it was stated above that heterophenomenology can't solve the other-minds problem, doesn't the OM problem directly refute studying consciousness from a third person perspective? If I only truly know that I feel, and can't confirm that others do, how can I collect data about them, what would it show? It couldn't show how/why they feel if I don't truly know that they DO. In this sense it would be like collecting data about feeling from a Turing machine.
ReplyDeleteMaybe I am experiencing the zombic hunch, but the idea of attempting to explain first person subjective experience, using the third person scientific method, seems absurd to me. What does Dennett mean when he describes heterophenomenology as “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…” Can’t we just admit that we can’t be ‘neutral’ when describing subjective experience? And then agree that that’s not a bad thing? In my opinion, science is never objective, and including subjective experience into the scientific method, and moving beyond a culture of positivism, could be useful in the study of consciousness. I realize that introspection doesn’t tell us how consciousness works, but the first person narrative shouldn’t be left out or discounted. To me it sounds like Dennet is trying to allow the subject to believe what they believe, but then tell them why their beliefs are wrong. Isn’t this the same thing as mansplaining.
ReplyDeleteOn another note, I finished this article feeling confused about the definition of qualia. I was under the impression that qualia was just your qualitative experience - your own perception of phenomena around you. If you don’t notice that something in your environment changes, how does that make the qualia wrong? Even if you didn’t notice a change, your experience remains the same. I don’t understand how “experience” can be incorrect. I think I am on team B with Chalmers on this one. Maybe the science of consciousness can learn something from the social sciences. People in sociology and anthropology conduct research, where the first person perspective is taken into account, and considered to be a valuable component to the data collected. This doesn’t discredit their research as being unscientific.
ReplyDelete“He will be awake, able to report the contents of his internal states, able to focus attention in various places and so on. It is just that none of this functioning will be accompanied by any real conscious experience. There will be no phenomenal feel. There is nothing it is like to be a Zombie”
This is how Chalmer’s describes the zombified version of himself, which is identical to him at a molecular level and therefore what we have described as t5 in this class. Chalmer’s says that he knows he is real, and has direct evidence of this due to his private feeling of having conciousness, but he denies that the t5 could possibly have this even if it genuinely feels like it does. In doing this, it would seem that he is ignoring the “problem of other minds” and by extension should be denying that any human aside from himself is conscious.
As Dennet says in response, Chalmers and his zombie would be indistinguishable through Dennet’s method of accounting for consiousness. I would agree that Dennet’s proposed method is likely the best way that we have to understand what someone else (human, zombie, robot) is experiencing.
A key part of Dennet’s heterophenomenology is using observations and scientific manipulations to find a discrepencies between what it feels like for a subject to feel this world and what is actually happening in the world itself. The split between someone’s personal reality and “objective reality” is a very interesting problem to consider when reverse engineering a human mind. As many studies in cognitive science have shown, humans are largely susceptible to holding false beliefs in highly predictable ways and are very prone to things like the change blindness scenario presented in this reading. When building a system that cognizes as we do (in the spirit of meeting the Turing Test benchmark), it may be necessary to build in intentional deficits that put constraints on things like the system’s perceptual and memory abilities because these flaws are a part of the personal reality that we cognize through.
ReplyDeleteIn my opinion, first-person data is just a synonym for a method we are already very familiar with in psychology, that is self-report. If we accept that using first person narratives is not a valid and reliable scientific methodology this would also discredit the significant portion of social science research that relies heavily on self-report questionnaires. In so many situations we rely on self-reports because (imperfect as they may be) they are the best way of measuring what we hope to measure. For example, how could we hope to measure pain without using a self-report measure? At present, there is currently no reliable nor objective biomarker of pain. We also know that the experience of pain is subjective and varies from person to person. Does this mean that we should not attempt to measure (and therefore study) pain? I agree there are many problems with self-report/first person data. These limitations should be taken into consideration and corrected for wherever possible (for example through the use of standardized measures, precise language, etc). In any case, the use of first person data really isn’t revolutionary, why not try to apply it to consciousness? We already use it to assess countless abstract concepts such as happiness, relatedness, motivation, etc.
ReplyDeleteDennett proposes of “heterophenomonology” to study conscious/subjective experience using third person data by measuring vocal sounds and “internal conditions (e.g. brain activities, hormonal diffusion, heart rate changes, etc.)” in objective terms. I think this approach is interesting but must be combined with first-person data to be useful and meaningful. If the problem you wish to solve is the “Hard Problem” (ie how and why people feel) then you must link any objective measures to internal subjective experiences. We need to understand what subjective “qualia” is associated with an increase in hormone levels.
Dennett has spent much of his time in this article advocating for heterophenomenology as a methodological solution to the backlash against “armchair introspection” and subjective first-person data. He tries to develop empirical methods to bridge the gap between first-person and third-person data, as the biggest criticism of any kind of first-person science of consciousness is its subjectivity, with little to no way of validating its own claims. While I think this is a valiant effort, I’m honestly still pretty unconvinced of the denial of the hard problem that Dennett puts forth.
ReplyDeleteRE: “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...”
ReplyDeleteThat doesn’t make sense to me. If Chalmers did suddenly become a zombie, would he simply not know that he had lost feeling? I think there’s a large difference between never having had feeling and insisting you do, versus having once had feeling and then losing it (like vision or audition). Wouldn’t he notice if he couldn’t/didn’t feel immediately? Would he not be surprised by his reactions to things which would normally elicit feelings from him?
This article made me think about how a lot of the things that we verbalize or even things that we consciously think might not be what we really believe or feel. For example, the IAT test shows in many cases that you might think you’re not biased towards a minority or something, but then you find out that you are unconsciously biased. How can we really trust what people report if we can’t also be sure that they are sure of what they’re reporting? This also relates to the fact that we can never be sure if someone is truly telling us the truth. You can ask what a person thinks they are thinking or what they think they know, but you always must take what they’re saying with a grain of salt.
ReplyDeletei. RE: “Turing showed us how we could trade in the first-person perspective of Descartes and Kant for the third-person perspective of the natural sciences and answer all the questions–without philosophically significant residue.”
ReplyDeleteI don't think this is truly trading the first person point of view with third person point of view of science. Even if we succeed at reverse-engineering the brain and making a robot with our own abilities we have only solved the easy problem while the hard problem which pertains to the first-person view is still unsolved.
ii. RE: “I realize that a "doxological" (or representational) reductionist like yourself will want to reduce feeling states to dispositions-to-believe. A resistor like myself need not deny, of course, that feeling states do have a tendency to produce beliefs.”
I’m wondering whether we should make any distinctions between beliefs and feelings in this case or are they just the same things that make the hard problem hard? We are using the word “feeling” in a very general sense that includes our sensations and all the things that go on in our heads so I am curious to know what is the difference between feelings and beliefs if there is any. Moreover, if beliefs and feelings are different things would they belong to the same category in relation to the hard problem?