If I am understanding this article correctly, Harnad supports Turing in the quest to simply focus on performance capacity in developing a T3 and to just ignore the problem of feelings. Either feelings will or will not arise with the creation of a T3, but that is not of concern because we won’t be able to tell anyways. In order to drive this point home, Harnad shows how Dennett fails to address the how/why problem or how we could get any closer to being able to get to causal explanations. Although its true that doing and feeling are not the same thing, is it possible to have doing without feeling? Why would we have evolutionarily developed feeling if it wasn’t for some purpose? What if there is some function of feeling that allows us to do what we can do? This may be quite an out there suggestion, but what if we couldn’t reach T3 without understanding how feeling plays a role because what if function doesn’t only give rise to feeling, but to some extent there is a bidirectional relationship?
Harnad's article does an excellent job of addressing the many problems and shortcomings of Dennett's article. His main criticism seems to center on Dennett's failure to address the questions of how and why we feel. Harnad also takes issue with the number of "weasel-words" Dennett uses throughout his article (ex. "belief", "experience"), which all essentially boil down to the state of feeling. Harnad argues that without feeling, there is nothing.
Dennett’s heterophenomenological approach entirely misses the purpose of establishing a science of consciousness.
The essay's constant recurring point was how any of the previous philosophies failed to address how and why we feel. A quote by Turing criticizes the solipsistic view that "the only way by which one could be sure that machine thinks is to be the machine and to feel oneself thinking. One could then describe these feelings to the world, but of course no one would be justified in taking any notice." This is the problem of other minds applied to machines, and now what they are thinking but if they are thinking at all. My question is whether seeking strong equivalence in an AI would ever help us understand why it is that we humans feel? I have difficulty drawing conclusions from these purely hypothetical scenarios, but somehow I still feel that even if the causal mechanisms would be the same, it would not be sufficient to conclude that AIs feel or explain why humans feel.
The Zombie Hunch Harnad discusses in his paper brings to mind an example from another class in which I read a chapter of Ramachandran’s book on patients with blindsight. Ramachandran describes blindsight as an example of our “zombie brain” because patients who were completely blind in half their visual field would perform complex motor movements (such as sliding a letter into a letterbox). They completely lacked any visual conscious awareness of the letterbox and yet a part of their brain (the “zombie brain”) was able to take over and perform this task. They would still claim that they did not see a letterbox nor how they put the letter in correctly. I think this raises all sorts of questions about consciousness. These patients who performed this task did it unconsciously, and yet their behavioural output was identical to someone who performed the task consciously. It also begs the question of how much of human cognitive function could operate below the level of our conscious awareness. How much of what we do could we do using only our “zombie brains”? I think it also proposes a specific definition for consciousness, which is probably more constrained than the broader term of “feeling” that we’ve used in this course. Consciousness (as it is defined in most other courses I’ve encountered) is often closely linked to language. I cannot decide myself if we are prone to linking consciousness to language because we are only able to test aspects of consciousness that can be verbalized or if it is because language is in fact very important for consciousness/feeling. For example, in the case of split brain patients where one hemisphere controls one half of the body and the left side is able to pick out objects in reference to an image only shown to the right hemisphere. The split brain patient will often make up a ridiculous story to explain why they picked that particular object because only the left hemisphere produces language. In this case would we say that the decision to pick out a particular object was unconscious? Is this another example of the zombie brain in action?
This might be completely tangential to this course, however I am wondering how we can reconcile these other definitions of consciousness with what has been discussed in this course. Do these examples provide insight into consciousness/feeling? Why or why not? Am I just conflating two very different ideas of what it means to be a zombie? In my mind these examples suggest that consciousness (and perhaps “feelings” as a whole) are not necessary in order for humans to do all that we do. Which begs the question of why we have them? (ie the hard problem). I also think it suggests that there is something special about language.
After reading this discussion/commentary, i am a bit confused in the sense that I feel a bit played tricked given that Denett played right into my initial defiance towards Chalmer's Zombies, to the point that i neglected his "oubli" of feeling/phenomenological experience in his claim. On the one hand, i think that his insistance on using 3rd person report rather than 1st person experience is good, as long as one remains aware that this remains an imperfect proxy. But in science that is often the best compromise (since by the other minds problem, it is really hard to simply trust the subject to report what we are looking for). on the other hand, not only is it a proxy, but it is a proxy that goes back and forth from the researcher to the subject (the researcher frames the question; the subject interprets and responds; the researcher interprets). However, I agree strongly that this reformulation of the Zombie problem: "The Hard Problem is simply explaining how and why we are not Zombies!" is much better. The way I see it, this question should not be dismissed (like Denett does), but at the same time how it should be studied is not Denett's concern, and is already a hard problem itself.
Oh my, I wish I had read this paper a month ago! The paper really made me realize what terms/ concepts I have been mixing up, and made the reasons why I disagreed with papers in the past, but couldn’t quite put my finger on why I disagreed, (such as Dennit’s) clear to me. So many of the debates on the “Hard Problem” are simply arguing different problems, and the lack of clarity and uniformity of what the crux of the issue really is, and terms used to describe such, really obscure whatever progress people think has been made on the “Hard Problem.” After reading the paper, I thought about how the term ‘consciousness’ really obscures the hard problem. I think the term has been sensationalized in cognitive science, and any specific meaning or application of it has been lost. ‘Consciousness’ has an ambiguous meaning across papers, and be used to describe a variety of phonomena like feeling, belief in feeling something, awareness of self…etc., or even as a catch all term for the batch. So, claims that the hard problem has been solved by explaining why/how of consciousness are misguided by the ambiguity of ‘consciousness.'
Harnad’s comment on heterophenomonology as a “Preferred method for what” really stuck with me. What are we measuring, why a person is motivated to do something? How they will act? No! The paper also made me very discouraged. The connection between doing and feeling is ultimately unknowable and seems silly to pursue. This might be a naive thing to say, but I don’t think the hard problem is unique in the way the question cannot be answered. Although it can’t be equated to questions of function/function like “Why does gravity pull?” I think finding an answer to the hard problem is akin to finding an answer to a question like “why/how does gravity exist at all.” We can explain the function of gravity, and have invented equations to predict how it works, but the question of why gravity should ever be needed in the first place, or why there is not a universe that doesn’t need gravity b/c it has even distributions of mass/ energy, is unanswerable. Although saying ‘feeling’ is just a property or fact of ‘doing' for entities with certain internal make-ups does not answer the questions of why/how that is, the same kind of question, even if its not one specifically of feeling/ doing can be said about many other aspects of our world. Kind of like when 5 year olds can’t stop asking ‘why’ and you get to a point where you just can’t answer anymore.
If I am understanding this article correctly, Harnad supports Turing in the quest to simply focus on performance capacity in developing a T3 and to just ignore the problem of feelings. Either feelings will or will not arise with the creation of a T3, but that is not of concern because we won’t be able to tell anyways. In order to drive this point home, Harnad shows how Dennett fails to address the how/why problem or how we could get any closer to being able to get to causal explanations. Although its true that doing and feeling are not the same thing, is it possible to have doing without feeling? Why would we have evolutionarily developed feeling if it wasn’t for some purpose? What if there is some function of feeling that allows us to do what we can do? This may be quite an out there suggestion, but what if we couldn’t reach T3 without understanding how feeling plays a role because what if function doesn’t only give rise to feeling, but to some extent there is a bidirectional relationship?
ReplyDeleteHarnad's article does an excellent job of addressing the many problems and shortcomings of Dennett's article. His main criticism seems to center on Dennett's failure to address the questions of how and why we feel. Harnad also takes issue with the number of "weasel-words" Dennett uses throughout his article (ex. "belief", "experience"), which all essentially boil down to the state of feeling. Harnad argues that without feeling, there is nothing.
ReplyDeleteDennett’s heterophenomenological approach entirely misses the purpose of establishing a science of consciousness.
The essay's constant recurring point was how any of the previous philosophies failed to address how and why we feel. A quote by Turing criticizes the solipsistic view that "the only way by which one could be sure that machine thinks is to be the machine and to feel oneself thinking. One could then describe these feelings to the world, but of course no one would be justified in taking any notice." This is the problem of other minds applied to machines, and now what they are thinking but if they are thinking at all. My question is whether seeking strong equivalence in an AI would ever help us understand why it is that we humans feel? I have difficulty drawing conclusions from these purely hypothetical scenarios, but somehow I still feel that even if the causal mechanisms would be the same, it would not be sufficient to conclude that AIs feel or explain why humans feel.
ReplyDeleteThe Zombie Hunch Harnad discusses in his paper brings to mind an example from another class in which I read a chapter of Ramachandran’s book on patients with blindsight. Ramachandran describes blindsight as an example of our “zombie brain” because patients who were completely blind in half their visual field would perform complex motor movements (such as sliding a letter into a letterbox). They completely lacked any visual conscious awareness of the letterbox and yet a part of their brain (the “zombie brain”) was able to take over and perform this task. They would still claim that they did not see a letterbox nor how they put the letter in correctly. I think this raises all sorts of questions about consciousness. These patients who performed this task did it unconsciously, and yet their behavioural output was identical to someone who performed the task consciously. It also begs the question of how much of human cognitive function could operate below the level of our conscious awareness. How much of what we do could we do using only our “zombie brains”? I think it also proposes a specific definition for consciousness, which is probably more constrained than the broader term of “feeling” that we’ve used in this course. Consciousness (as it is defined in most other courses I’ve encountered) is often closely linked to language. I cannot decide myself if we are prone to linking consciousness to language because we are only able to test aspects of consciousness that can be verbalized or if it is because language is in fact very important for consciousness/feeling. For example, in the case of split brain patients where one hemisphere controls one half of the body and the left side is able to pick out objects in reference to an image only shown to the right hemisphere. The split brain patient will often make up a ridiculous story to explain why they picked that particular object because only the left hemisphere produces language. In this case would we say that the decision to pick out a particular object was unconscious? Is this another example of the zombie brain in action?
ReplyDeleteThis might be completely tangential to this course, however I am wondering how we can reconcile these other definitions of consciousness with what has been discussed in this course. Do these examples provide insight into consciousness/feeling? Why or why not? Am I just conflating two very different ideas of what it means to be a zombie? In my mind these examples suggest that consciousness (and perhaps “feelings” as a whole) are not necessary in order for humans to do all that we do. Which begs the question of why we have them? (ie the hard problem). I also think it suggests that there is something special about language.
After reading this discussion/commentary, i am a bit confused in the sense that I feel a bit played tricked given that Denett played right into my initial defiance towards Chalmer's Zombies, to the point that i neglected his "oubli" of feeling/phenomenological experience in his claim. On the one hand, i think that his insistance on using 3rd person report rather than 1st person experience is good, as long as one remains aware that this remains an imperfect proxy. But in science that is often the best compromise (since by the other minds problem, it is really hard to simply trust the subject to report what we are looking for). on the other hand, not only is it a proxy, but it is a proxy that goes back and forth from the researcher to the subject (the researcher frames the question; the subject interprets and responds; the researcher interprets). However, I agree strongly that this reformulation of the Zombie problem: "The Hard Problem is simply explaining how and why we are not Zombies!" is much better. The way I see it, this question should not be dismissed (like Denett does), but at the same time how it should be studied is not Denett's concern, and is already a hard problem itself.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteOh my, I wish I had read this paper a month ago! The paper really made me realize what terms/ concepts I have been mixing up, and made the reasons why I disagreed with papers in the past, but couldn’t quite put my finger on why I disagreed, (such as Dennit’s) clear to me. So many of the debates on the “Hard Problem” are simply arguing different problems, and the lack of clarity and uniformity of what the crux of the issue really is, and terms used to describe such, really obscure whatever progress people think has been made on the “Hard Problem.” After reading the paper, I thought about how the term ‘consciousness’ really obscures the hard problem. I think the term has been sensationalized in cognitive science, and any specific meaning or application of it has been lost. ‘Consciousness’ has an ambiguous meaning across papers, and be used to describe a variety of phonomena like feeling, belief in feeling something, awareness of self…etc., or even as a catch all term for the batch. So, claims that the hard problem has been solved by explaining why/how of consciousness are misguided by the ambiguity of ‘consciousness.'
DeleteHarnad’s comment on heterophenomonology as a “Preferred method for what” really stuck with me. What are we measuring, why a person is motivated to do something? How they will act? No! The paper also made me very discouraged. The connection between doing and feeling is ultimately unknowable and seems silly to pursue. This might be a naive thing to say, but I don’t think the hard problem is unique in the way the question cannot be answered. Although it can’t be equated to questions of function/function like “Why does gravity pull?” I think finding an answer to the hard problem is akin to finding an answer to a question like “why/how does gravity exist at all.” We can explain the function of gravity, and have invented equations to predict how it works, but the question of why gravity should ever be needed in the first place, or why there is not a universe that doesn’t need gravity b/c it has even distributions of mass/ energy, is unanswerable. Although saying ‘feeling’ is just a property or fact of ‘doing' for entities with certain internal make-ups does not answer the questions of why/how that is, the same kind of question, even if its not one specifically of feeling/ doing can be said about many other aspects of our world. Kind of like when 5 year olds can’t stop asking ‘why’ and you get to a point where you just can’t answer anymore.