10a. Dennett, D. (unpublished) The fantasy of first-person science
Once we can pass the Turing test -- because we can generate and explain everything that cognizers are able to do -- will we have explained all there is to explain about the mind? Or will something still be left out?
Reading: Dennett, D. (unpublished) The fantasy of first-person science.
Instructions for commenting: Quote the passage on which you are commenting (use italics, indent). Comments can also be on the comments of others. Make sure you first edit your comment in another text processor, because if you do it directly in the blogger window you may lose it and have to write it all over again.
“Once we can pass the Turing test — because we can generate and explain everything that cognizers are able to do — will we have explained all there is to explain about the mind? Or will something still be left out?”
ReplyDeleteThis passage raises a thought-provoking question about the limits of science in understanding the brain. Dennett suggests that even if we could replicate and predict all cognitive behavior perfectly, there might still be aspects of subjective experience (what it feels like to think, perceive, or feel) that doesn't have an explanation. I find this idea compelling because it highlights the difference between understanding the mind in terms of functions and behaviours versus understanding the first-person, qualitative experience. It also challenges a common assumption in psychology and neuroscience that observing and modelling behaviour is enough to fully explain 'mental life'. Dennett pushes us to consider whether some elements of consciousness are inherently private, or if they can eventually be captured through objective scientific methods. This passage makes me reflect on how far science can go in explaining our inner experiences and truly how “hard” the hard problem of consciousness is. It’s a reminder that understanding the brain might require not just technical knowledge, but also philosophical reflection on what it really means to experience the world and why life matters.
Rachel, I like how you spoke about whether any aspect of behavior might be left out even if we could explain all behavior. One point that Dennett wrote that struck me is that we should not treat our own thoughts and feelings as some kind of private access to the truth. He actually uses the term "fantasy" to describe how we are misled to believe that just going inside our minds and experiencing allows us to access some internal science that has not been made valid & open to others or tested & confirmed with evidence. Dennett seems to say: yes, we all have experiences, but those experiences can also mislead us in the same way that an optical illusion misleads us.
DeleteThe takeaway for me was not that a first-person experience does not matter, of course it does, but it cannot stand on its own. We need the combination of "it feels like this" with a shared, testable approach in order to study the mind. This made me consider that the harder part is not so much debating whether there is a first-person experience or the science but understanding how to mesh both ends of the spectrum without having to pretend the other gets the whole job done on its own.
***EVERYBODY PLEASE NOTE: I REDUCED THE MINIMUM NUMBER OF SKYWRITINGS. BUT THE READINGS ARE **ALL** RELEVANT TO AN OVERALL UNDERSTANDING OF THE COURSE. SO, EVEN IF YOU DO NOT DO A SKYWRITING ON ALL OF THEM, AT LEAST FEED EACH READING YOU DO NOT READ TO CHATGPT AND ASK IT FOR A SUMMARY, SO YOU KNOW WHAT THE READING SAID — OTHERWISE YOU WILL NOT HAVE A COMPLETE GRASP OF THE COURSE TO INTEGRATE AND INTERCONNECT FOR THE FINAL EXAM.***
DeleteRachel, but Dennett does not think a complete “heterophenology” would leave anything out. (Is he right?) His heterophenomenology is part of the solution to the Easy Problem (EP) (T4). Does T4 leave something out?
It doesn’t sound like you are describing Dennett’s view, but the views of others he mentions. The Hard Problem (HP) is not the problem of describing feeling, or of explaining what it feels like to feel. It’s the problem of explaining (reverse-engineering) how and why sentient species (human and nonhuman animal) feel at all. Why is that problem “hard”? and why does solving the EP not solve the HP? (See the next reading.)
Lorena, “experience” is a WW. What we really mean is felt experience, in other words, feeling. Why do organisms not just do (or learn to do) what needs to be done, without feeling anything at all? We know the biological function of doing: what is the biological function of feeling? We also know the biological function of thinking. Thinking (cognition) could, for example, have been just computation (executing an algorithm). But it’s not. (Why not?) It feels like something to think: Why? Both seeing something and hallucinating feel like something: what for? And how? The HP is not about whether something you feel (“it’s hot in here”) is true or false but about why and how we feel anything at all. And what every sentient organism knows for sure (as Descartes’ Cogito/Sentio notes) is that they are feeling whatever they are feeling (while they are feeling).
In this paper, the author defends heterophenomenology, which is an empirical method that does not assume subjective experience is accurate when collecting data on cognition. He contrasts the third person scientific approach to studying cognition with the first person philosophical notion that there is an irreducible subjective experience to consciousness. He then concludes that these subjective experiences, upon which the foundation of firstperson science is built, are an illusion, and that heterophenomenology must use third-person methods. My favourite part of this paper was the author’s explanation of the heterophenomenological zombie twins, as it helped make the arguments in the paper more tangible.
ReplyDeleteThanks for your summary Jesse! It made it easier for me to follow the paper. It’s interesting how he wrote about learning to discredit illusions of our subjective experience. As we move forward with scientific research, we have to push past instincts that our subjective feelings know something more than what is concretely proven. Lorena also spoke to this in her comment.
DeleteThese two quotes summarize this idea well:
We’ve learned to dismiss other such intuitions in the past–the obstacles that so long prevented us from seeing the Earth as revolving around the sun.
You are not authoritative about what is happening in you, but only about what seems to be happening in you.
Aside: It’s interesting that Dennett has put Searle on “Team B” (believing that there is something more to consciousness) as Searle argued for similar points to Dennett in the video from week 3: that consciousness is a natural, biological phenomenon, and that introspection is not a sure method of study.
Dennett makes me think about how we can really study the mind. He says first person feelings are not good for science because they can be wrong, but I feel they still matter. When people say what they feel, it shows something about how the mind works from inside. Maybe we should not ignore first person experience, but use it together with real data like brain activity or behavior. That way we can see both the feeling and the reason behind it. Dennett is right that science needs proof, but I think feelings can still give small clues about what the mind really is.
ReplyDeleteRena, I totally agree! Dennett's heterophenomenology sounds devoid of bias, which in it of itself shows they are not correctly approaching innate human cognitive processes and biases. The following quote struck me:
Delete"[...] but to these verbal reports must be added all the other manifestations of belief, conviction, expectation, fear,
loathing, disgust, etc., including any and all internal conditions (e.g. brain activities, hormonal diffusion, heart rate changes, etc.) detectable by objective means."
How de we add manifestations of belief and conviction, meshing them with the raw data? How and to what extent ought we include so-called objectively measured internal conditions? How is the 'objective' measure of LH secretion integrated to vocal measurements? Dennett's approach is not inherently wrong, as I do believe that big data approaches can be informative; however, information should be integrated as to yield more information (which was not fully discussed), not just integrated for the sake of it. I believe that approaching data from the assumption that subjective experiences are wrong and that their interpretation can be rescued by 'objective', 'scientific' measures is absurd: I wonder, who is doing the scientific measurement? Well, humans. All human-used machines designed to measure some parameter have bias built into them.
In my opinion, the fallacies in Dennett's heterophenomenology are the following:
1) Their scientific method applies solely to participants, not the investigator alike.
2) No overview or guideline on how big data should be apprehended.
3) The sacralization of big data; not all information is pertinent, and no measurement is ever objective.
Camille: I understand the issues that you have found with Dennett's heterophenomenology, and believe that there is a crucial problem in fallacy number 2.
DeleteIndeed, Dennet here argues that first-person accounts of consciousness are not sufficient for the scientific study of consciouscness. Human beings being unreliable, and often being wrong regarding their experiential beliefs (as illustrated by the change blindness experiment), it seems logical that first-person accounts of consciousness cannot be sufficient in establishing a scientific account of consciousness. He therefore argues that third-person "objective and scientific" decision making is essential in this process.
I agree with you that precision of the manner in which the different components of "big data" should be apprehended remains essential, particularly as it is not obvious to me that a firt-person account (auto-phenomenological description) ought to be given the same weight or relevance as data regarding physiological measures. These sources of information are profoundly different in kind: how could we treat them in the same manner? This need for complementarity is rooted in the fundamental differences between these sources of information: these differences must therefore be accounted for in data processing.
Dennett claims that measuring expressions of belief is enough to create the participant’s heterophenomenological world is maximally inclusive to study consciousness. He says that 1st-person studies are not useful for studying consciousness because you can be mislead, believe things that are false or be unaware of some beliefs even though they are conscious experiences. In contrast, Chalmers says that 1st-person access and data are essential for the the study of consciousness as it cannot be reduced to 3rd-person data. I agree with that because everything feels like something regardless of whether you are being mislead by your senses, unaware of a past trauma, seeing it or not able to say it verbally. Isn’t what consciousness is all about? I would be curious to know what Dennett thinks about consciousness in the cases of substance influence, Alzheimer’s disease, psychosis or even paralysis. He says he measures behaviourally, viscerally and verbally, but does it really tell everything?
ReplyDeleteThis paper covers Dennett’s argument that consciousness should be studied using third-person, objective, scientific methods, through a heterophenomenology framework. He critiques the idea of “first-person science”, labelling it a fantasy based on direct private introspection, since people’s own experiences can be misleading. Conscious experiences themselves should not be treated as primary data when people hold false beliefs about what they perceive. A crucial example is apparent motion or change blindness, where people misjudge what they see, supporting heterophenomenology's focus on collecting observable data about the subject remaining neutral about them being true or false. Although some philosophers challenge Dennett for avoiding subjective experience, he argues that his method captures everything that can be scientifically studied about consciousness.
ReplyDeleteHey Emily! Great job articulating Dennett’s position well! I think the change-blindness example you used was a good example of Dennett’s overall point about introspection as unreliable; but it also highlights a deeper issue – that regardless of whether we accurately report qualia or not, the question remains as to whether brain activity itself generates felt experiences (rather than just processing of neural patterns). I think Dennett may equate prediction with explanation (meaning if we are able to use neural correlates to predict subjective reports about feelings, then we have explained them) which kind of side steps the hard problem in its entirety. Dennett’s approach doesn’t really tell us how (or even why) those processes actually generate those subjective feelings…
Delete“How does it work? We start with recorded raw data. Among these are the vocal sounds people make (what they say, in other words), but to these verbal reports must be added all the other manifestations of belief, conviction, expectation, fear, loathing, disgust, etc., including any and all internal conditions (e.g. brain activities, hormonal diffusion, heart rate changes, etc.) detectable by objective means.”
ReplyDeleteI find Dennett’s position that heterophenomenology is the only way to study consciousness to contain multiple fallacies. However, for the sake of keeping it concise, I’ll focus on one fallacy: his proposed method, which measures internal systems of cognition through external conditions, confuses correlation with causation. This method does not explain how the observed physiological or behavioural changes cause subjective experience, only that they happen alongside it. Treating the correlation between internal states and reports as an explanation risks collapsing the hard problem into data patterning. In essence, heterophenomenology might tell us when consciousness appears, but not why it feels like anything at all.
Maya, could you explicit where you understand that Dennet suggests physiological data points can be used to identify the cause of consciousness? I rather thought that he was suggesting that the medthodology should be so broad specifically because we do not know what causes consciousness, or what it is, so anything which happens alongside it must be considered as potentially yielding some information?
DeleteWhat I found most interesting is how Dennett’s heterophenomenology (his proposed method for studying consciousness) avoids relying on first-person claims by treating them as data about what seems to the subject, not as facts about what is actually happening. In other words, heterophenomenology is a third-person method that collects verbal reports and all other measurable data but stays completely neutral about whether these are true. Dennett thinks this is enough to solve the Easy Problem at the level of T4. But even a perfect T4 robot explains nothing about the Hard Problem, which is explaining how and why any of this doing is felt.
ReplyDeleteI agree with your take on Dennett’s neutrality. What really struck me in the article is how he insists heterophenomenology includes all third-person data, not just what people say, but every measurable reaction (blushes, brain activity, hesitations, etc.). He brings all of that into the “heterophenomenological world” without assuming any of it is true or false.
DeleteYou’re also right that this still doesn’t reach the Hard Problem. Dennett treats the sense that “something is missing”, what he calls the Zombic Hunch, as just another belief that needs explaining, not evidence of an inner property. But from a Type-B perspective, that move feels like it sidesteps the real question of why any of this functioning is felt. So heterophenomenology handles the Easy Problem well, but it doesn’t answer what you highlight: why doing is accompanied by experiencing.
Dennett’s heterophenomenology is not compatible with “1st person science”, which relies on introspection and takes data from introspection as truth. Heterophenomenology, on the other hand, takes any perceived experience as data, but not necessarily ‘true’ data - since the experience can either be ‘correct’ or ‘incorrect, since people can have false beliefs about their experiences. So heterophenomenology is incompatible with introspection as a scientific method, but what about empiricism? Empiricism attests that all knowledge is gained through sensory experience, but what if our senses are fallible? What ‘empirical’ data heterophenomenology accept?
ReplyDeleteDennett rejects the idea that there is a real Hard Problem, arguing that what Type-B theorists like Chalmers and Harnad treat as an unexplained inner essence is really just a misleading intuition. For him, consciousness is nothing beyond cognitive processes and the beliefs we form about them. From a Type-B stance, though, Dennett would seem to be addressing only the “easy problems,” because he denies the very phenomenon (the challenge of explaining how and why organisms feel at all, not just describing what feeling is like) that the Hard Problem is asking about. In that sense, a Type-B theorist would say Dennett never tackles the Hard Problem at all.
ReplyDeleteWhile I find Dennett’s paper interesting, I’m still unsure how the Other Minds Problem fits into his account. While I am aware that the Other Minds Problem is a philosophical idea, it is an important one to bring to the table when talking about the Hard Problem, and it is also one that Turing himself brought up (or "meant" to bring up when he brought up solipsism, according to Harnard).
ReplyDeleteDennett argues that the subject can report their own felt experience through verbal descriptions, so in that sense the Other Minds Problem doesn’t arise for the first-person perspective. But I’m struggling to see why a *third-person* mechanistic explanation of consciousness (i.e. heterophenomenology) wouldn’t still face the same challenge. From verbal reports, all we really have are descriptions *of* feelings/felt states, not direct access to those states themselves.
And, if psychology moved away from introspection because it wasn’t considered reliable, I’m unsure how heterophenomenology differs in practice. Is it due to the added components of heterophenomenology? I think the article is just harder for me to grasp.
The main issue with using T3 to model human cognition is that it doesn’t address the hard problem. While some might believe because a T3 model is grounded in experience it can model our consciousness as a whole but this model fails to address how a subjective feeling experience is created. According to Dennett, the hard problem is not actually a problem that can be solved through science. His heterophenomenolgy says that science should treat people’s reports of experience as data about beliefs and not direct windows into their experiences. This can not explain how subjective feeling arises. So a T3 model could model all of cognition without explaining why things feel like anything.
ReplyDelete“Heterophenomenology is nothing but good old 3rd-person scientific method applied to the particular phenomena of human (and animal) consciousness… bringing the data of the first person into the fold of objective science.”
ReplyDeleteI think this quote captures Dennett’s whole strategy: he wants to treat people’s reports about their experiences the same way we treat any other scientific data. He isn’t denying that people feel things, he just says that science should treat those feelings like observable claims, not as private truths we automatically trust.
I like this idea because it keeps the study of consciousness grounded. Instead of assuming that introspection gives us special access to reality, Dennett brings everything “into the fold” of public evidence: what people say, how they act, and what their bodies do. That makes the method clear and testable.
At the same time, I’m not sure this solves everything. Even if we can capture all first-person reports in a scientific way, we still don’t explain why those reports come with a feeling behind them. Dennett’s method helps with the Easy Problem, explaining what organisms do but it doesn’t really touch the Hard Problem, which asks why any of this 3rd-person activity is accompanied by a 1st-person experience at all.
So heterophenomenology gives us a strong scientific framework, but it doesn’t remove the deeper mystery of why consciousness is felt instead of just done.
While much of the debate focuses on the "Zombic Hunch," I was most struck by Dennett’s comparison of inaccessible first-person facts to "inert historical facts". He writes that these are like the fact that "some of the gold in my teeth once belonged to Julius Caesar... no possible extension of science will ever be able to say which is the truth".
ReplyDeleteThis analogy seems to do the heavy lifting in dismissing the "B Team’s" insistence on irreducible qualia. It trivializes the "Hard Problem" by suggesting that "unknowable" facts aren't profound mysteries but that they are rather just informationally empty, like the history of a gold atom. If first-person data is truly inaccessible to third-person investigation, Dennett implies it is as scientifically irrelevant as Caesar’s dental history. It pushes me to ask: why do we value an "extra ingredient" in consciousness if, by definition, it makes no verifiable difference?
Dennett argues that a complete Heterophenomenology, a full account of what people do, say, and report - would leave nothing out. But this only solves the Easy problem. It explains the functional mechanisms of perception, learning, memory and behaviour. What it does not explain is why these processes are felt rather than purely computed. A system could, in principle, perform all the same functions without having any experience at all.
ReplyDeletePTSD could make this gap obvious. We can describe all the mechanisms such as memory retrieval, threat detection, autonomic arousal, and still not explain why the traumatic memory is felt with such overwhelming emotional force. Two people could have the same behavioural triggers and neural circuitry, yet one experiences crippling flashbacks and the other does not. The “doing” is similar, what differs is the “feeling”, the lived presence of the memory.
Heterophenomenology can record the reports, but it is not the same as explaining why it is felt from the inside. That is why T4 still leaves something out. The HP is not about describing particular feelings, but about explaining how and why organisms feel anything al all, rather than functioning entirely unconsciously. Solving the Easy Problem gets a us a full model of cognition, but it does not tell us why cognition is accompanied by experience.
This reading, along with our class discussions, made me wonder about the relationship between the easy and the hard problems and the idea of degrees of freedom. How can we say that the easy problem uses all the degrees of freedom which prevents us from understanding the hard problem when we don't know the complete truth of the easy problem? Isn’t there a possibility that feelings have a strictly physiological explanation the same way that the easy problem supposedly does? My question is more why do we assume that they are two different problems when we haven’t completely explained either?
ReplyDeleteJulien, Here’s my reply. It was a little long, so instead of posting in the blog, I posted it here:
DeleteOn the “Easy Problem” (EP) vs. the “Hard Problem” (HP) of Cognitive Science
There’s also a follow-up here:
Exchange with ChatGPT On the “Easy” and “Hard” Problem of Cognitive Science.