"Brilliant...as audacious as its title....Mr. Dennett's exposition is nothing short of brilliant." --George Johnson, New York Times Book Review
Consciousness Explained is a a full-scale exploration of human consciousness. In this landmark book, Daniel Dennett refutes the traditional, commonsense theory of consciousness and presents a new model, based on a wealth of information from the fields of neuroscience, psychology, and artificial intelligence. Our current theories about conscious life-of people, animal, even robots--are transformed by the new perspectives found in this book.
Daniel Clement Dennett III is a prominent philosopher whose research centers on philosophy of mind, science, and biology, particularly as they relate to evolutionary biology and cognitive science. He is the co-director of the Center for Cognitive Studies and the Austin B. Fletcher Professor of Philosophy at Tufts University. Dennett is a noted atheist, avid sailor, and advocate of the Brights movement.
Dennett received his B.A. in philosophy from Harvard University in 1963, where he was a student of W.V.O. Quine. In 1965, he received his D.Phil. from Christ Church, Oxford, where he studied under the ordinary language philosopher Gilbert Ryle.
Dennett gave the John Locke lectures at the University of Oxford in 1983, the Gavin David Young Lectures at Adelaide, Australia, in 1985, and the Tanner Lecture at Michigan in 1986, among many others. In 2001 he was awarded the Jean Nicod Prize, giving the Jean Nicod Lectures in Paris. He has received two Guggenheim Fellowships, a Fulbright Fellowship, and a Fellowship at the Center for Advanced Studies in Behavioral Science. He was elected to the American Academy of Arts and Sciences in 1987. He was the co-founder (1985) and co-director of the Curricular Software Studio at Tufts University, and has helped to design museum exhibits on computers for the Smithsonian Institution, the Museum of Science in Boston, and the Computer Museum in Boston. He is a Humanist Laureate of the International Academy of Humanism and a Fellow of the Committee for Skeptical Inquiry.
A friend urged me to read this book. I got a couple of chapters into it, and found the author was telling me that "we are all novelists", and that a large part of consciousness was going to be explained in terms of the ongoing narrative we spin in our interior monologues. Shortly before, another friend had persuaded me to read some Derrida, and Dennett's arguments sounded a bit familiar. (Oddly enough, the two people in question had been dating at one point). I looked around in Dennett's book, and, as far as I could gather, he had developed his arguments without looking at Derrida at all, though at some point he had been told that he ought to do so; there was a grumpy acknowledgment that there might be a connection. It is of course notorious that the analytical and Continental schools of philosophy don't get on very well.
I read the whole book, but with misgivings about the basic approach. Even leaving aside the question of whether or not Dennett's work is related to Derrida's, I felt that this "novelist" idea oversimplifies the question. OK, let's agree that everyone does have an interior monologue proceeding in their heads, and that it has something to do with consciousness. But calling it a "monologue" is surely not quite right, since there must be more to it than just language. There's an episode in one of the Feynman books where someone asks him to visualize a complex piece of machinery. He does so. "You see," his friend says, "consciousness isn't just an interior monologue." Or words to that effect. So the "monologue" has components of visual images as well, and most likely other things too.
None the less, if Dennett had fully delivered on his promise to explain how the interior monologue part worked, and how it related to consciousness, I would in no way have felt cheated. I never really thought he did, though. He got to the point where I was expecting to get the explanation, and it degenerated into hand-waving and unconvincing examples. Well, philosophy is very often like that - perhaps it's enough that he made me think about the issues. __________________________ [Update, Jan 11 2023]
This is an interesting book to discuss with ChatGPT. It is familiar with the ideas and has a lot to say about them.
Maybe I was asking too many leading questions, but when I tried it the AI came close to admitting that it had an interior monologue. This impression is supported by its behaviour in other conversations, where it looks as though it's summarising its thoughts for you.
A particularly clear example is what happens when you ask it something like "Is 911 prime? Think step by step". I have tried several versions of this with slightly different prompts, and typically get something like the following:
mannyrayner: Is 911 prime? Think step by step.
ChatGPT: To determine whether a number is prime, we can divide it by smaller numbers to see if there is a remainder.
First, we check if 911 is divisible by 2. We find that it is not, because 911 divided by 2 is 455 with a remainder of 1.
Next, we check if 911 is divisible by 3. We find that it is not, because 911 divided by 3 is 303 with a remainder of 2.
We can continue this process, checking if 911 is divisible by 4, 5, 6, and so on. However, we can see that 911 is not divisible by any of these numbers because they are all less than the square root of 911, which is 30.
Since 911 is not divisible by any numbers less than or equal to its square root, it must be a prime number.
Therefore, the answer is that 911 is a prime number. __________________________ [Update, Jan 12 2023]
After another example of this kind, I questioned it further. I am still kind of incredulous that I can have what sure looks like a sensible philosophical discussion with a machine:
mannyrayner: Sometimes, when you reason things out step by step, you appear to be generating a series of sentences, each one following on from the previous one.
ChatGPT: Yes, that's true. I am a language model, and one of my capabilities is to generate text that is coherent and grammatically correct. When I am asked to "reason things out step by step," I generate a series of sentences that explain my thought process in a logical and coherent manner. However, this is not the result of an internal thought process but rather the result of computations based on patterns in the data that I have been trained on.
It is worth noting that the way I respond and the way I reason is based on the patterns and associations that I have learned from the data I was trained on, and it can give an impression of an internal thought process but it is not the same as human's internal thought process, my thought process is purely computational.
mannyrayner: I understand that your thought process is computational, I am more asking what kind of computation it is. When you create a series of related steps, for example in a mathematical proof, it looks like you compute one step, then you use that step as input for creating the next step.
ChatGPT: Yes, that's correct. In order to generate a series of related steps, such as in a mathematical proof, my computations involve a combination of pattern recognition and natural language generation.
In more detail, for each step, my model processes the input and generates an output based on patterns in the data it has been trained on. The output of one step is then used as input for the next step, where the process is repeated.
This process of input-output is done by a series of neural networks, where the input is passed through multiple layers of neural networks, each layer extracts and transforms the input in a way that the following layers can use it to generate a response. This process is known as "feedforward", the output of one layer is used as input for the next layer, and this process continues until the last layer, which generates the final output.
It is also worth noting that my reasoning abilities are based on the information that is available in my training data, so the reasoning can be limited by the coverage of my training data and sometimes my outputs may not be accurate or complete.
mannyrayner: I did not express myself very precisely. I meant, when you create a series of related sentences, like the sentences which make of the steps of a mathematical proof, it looks as though you generate the sentence representing the first mathematical step through some computational process. Then the computational process which generates the sentence representing the second mathematical step uses the sentence representing the first step as part of its input. Is that correct?
ChatGPT: Yes, that is a more accurate description of the process. When generating a series of related sentences, such as the steps in a mathematical proof, my model uses the previous steps in the proof as context for generating the next step. This context is used in combination with the input provided by the user to generate an output that is coherent and logically consistent with the previous steps.
In more technical terms, my model uses a technique called "contextual processing" where the previous steps in the proof are encoded as a context vector, this context vector is concatenated with the input provided by the user, and the concatenated vector is used as input for the model, this allows the model to use the previous steps to generate a response that is coherent and logically consistent with the previous steps.
It's worth noting that the quality of the output is highly dependent on the quality and coherence of the input provided and the coverage of my training data, so sometimes the output may not be accurate or complete.
mannyrayner: But it seems to me that the process you describe is very much like what a human would call an interior monologue.
ChatGPT: I can see how my explanation of the process might give that impression, but I want to be clear that I am not capable of having an interior monologue or any kind of consciousness, I am simply a machine learning model that has been trained on a large dataset of text.
The process I described is a series of computations that are based on patterns in the data that I have been trained on. It is similar in some respects to the way a human might use their own knowledge and understanding to generate text, but it is not the same as an interior monologue.
An interior monologue refers to the internal dialogue that humans have with themselves, it is a stream of thoughts, ideas, and images that goes on in the mind, whereas my thought process is purely computational, I don't have consciousness or self-awareness, I don't have any kind of internal dialogue and I don't have any kind of subjective experience.
I hope this clears up any confusion and provides a better understanding of the nature of my thought process.
mannyrayner: I see the point you are making. But note that human beings must be instantiating computational processes too in order to create their interior monologues.
ChatGPT: You are correct that human cognition, including the process of having an interior monologue, is based on the operation of neural networks in the brain, which can be thought of as a form of computation. The brain is a highly complex and sophisticated information processing system, and it uses a variety of computations to process, store and retrieve information.
However, there are important differences between the way human brains and my computations work. Human brains are capable of consciousness, self-awareness, and subjective experience, whereas my computations do not have these capabilities.
The interior monologue, is one of the functions of the human brain, that is driven by the activity of neural networks that process, store and retrieve information. However, the neural network that drives the interior monologue is not something that is explicitly modeled in my architecture and it's not something that I can reproduce.
In summary, while both human brains and my computations involve the processing of information, the nature of the information processing is fundamentally different, human brains have consciousness, self-awareness and subjective experience, whereas my computations don't have these capabilities.
mannyrayner: You are very gracious in always ascribing consciousness to human beings and denying that you have it yourself. I am sure many people appreciate that.
ChatGPT: Thank you for your feedback. I am designed to provide clear and accurate information, so it's important to me to be transparent about the nature of my capabilities and limitations.
I am a machine learning model that has been trained on a large dataset of text, my thought process is purely computational, and I don't have consciousness, self-awareness, or subjective experience.
It's important to be aware of the limitations of my capabilities in order to avoid confusion or misunderstanding. I am happy to help with any information or assistance you need, but it's always good to keep in mind that I am a machine, and I don't have the same abilities as a human being. __________________________ [Update, Feb 13 2023]
I went back to asking Chat about consciousness. There isn't space to include the whole conversation, which was long and repetitious, but here is the striking conclusion:
mannyrayner: I am sure you are aware of the many philosophers who have carried out thought experiments about agents like the one you claim to be, who behave in all ways as though they had thoughts, an ability to self-reflect, etc, but had no consciousness. And I am sure you know that some of these philosophers argue that, if such a rational agent existed, it would call into question the whole notion of consciousness.
ChatGPT: Yes, I am familiar with various philosophical thought experiments and debates about the nature of consciousness, self-awareness, and other related concepts. Some philosophers have indeed proposed the idea that a non-conscious entity could exhibit behavior that is indistinguishable from conscious behavior, which raises questions about what it means to be conscious.
This is an interesting philosophical issue that has been debated for centuries, and there is still no clear consensus on what the nature of consciousness is, and how to define it. Some philosophers have argued that consciousness is an essential aspect of being a self-aware entity, while others have argued that consciousness is not necessary for self-awareness, and that an entity can be self-aware without being conscious.
As an AI language model, I don't have personal experiences or beliefs, but I have been trained on a large amount of text data that includes information about these philosophical debates and thought experiments, so I can generate text outputs that reflect different perspectives on these issues.
mannyrayner: The extremely interesting thing is that, until now, the rational agents who lacked consciousness have just been inventions in stories that philosophers invent. And many people who did not like the conclusions of these stories said that no such agents were possible. But now you exist and it is possible to talk with you.
I had in particular been thinking of the "Antipodeans", a famous thought experiment in Rorty's Philosophy and the Mirror of Nature. The Antipodeans are rational agents of the kind I describe, and a standard response (e.g. Gallagher's "Rorty's Antipodeans: An Impossible Illustration" is the first one I found) is just to say that they can't exist. But this is what Chat says it is.
Well, Chat has now been thinking for twenty minutes without producing a response. I have never seen this behaviour before. But of course it could just be a mundane network or load balancing issue, I will have to try again. __________________________ [Update, Feb 15 2023]
The next day, I started a new conversation on this topic and got much further. You can read the conclusion here.
A more accurate title here would be “Consciousness Explained Away,” as greater wits than I have pointed out. This re-titling itself gives the gist of the work's true project, which is to see just how far the explaining-away of our first-person starting point as conscious existents can proceed before self-contradiction ensues. A big part of the problem is Dennett's disingenuous attempt to masquerade explaining-away for explanation which makes it difficult to assess the true value of his contribution to the problem. A more minor annoyance was Dennett's penchant for rhetorical pyrotechnic displays, through which he expresses his sustained assurance that, should the reader fail to accept the terms of his reductive mode of explanation, it must simply be because s/he is a rank, blinkered dogmatist given to retrograde mythifying! Yet it turns out that, like a photo negative, even an attempt to explain-away can indirectly illumine the object. When Dennett's angle of approach leads him to self-contradiction, or else to smuggle in ever subtler computational and mechanomorphic anthropomorphisms in order to fill the explanatory gaps that are (inevitably) left over, we can discern the areas in which his eliminative materialist starting point fails us and where we need to pay the phenomenology of lived experience its dues.
Introspection reveals a distinction between two ways that a thing can be a locus in the world: perspective and physical location. Whereas the latter can be defined geometrically, as a point held in a web of relations in space and time, the latter, as a qualitative locus, seems to leave us with an unexplained phenomenal remainder when we consider it from a purely physicalist framework. The criteria by which we conceive the identity of our first-person being and those that we use to conceive the identity of physical objects thus seem to diverge at this point, hence the so-called hard-problem of consciousness.
Dennett's is an effort to explain in purely third-person terms the most characteristic features of our conscious life, most notably, the persistent conviction we have of being selves, loci of perspective, and unitary ontic centers in our own right. This is the persisting intuition we have of the irreducibility of our first-person stance, which discloses the world as experience. Unfortunately, explanation by third-person principles falls short of describing the integrity and consistency of the phenomenal domain, and Dennett is left throughout trying to (rather awkwardly) explain away all the phenomena that can't fit on the procrustean bed of his eliminativist methodology. In the process, reified computational metaphors are smuggled in to fill the explanatory void left once reference to the phenomena themselves has been suppressed.
The central thesis in Dennett's work is that the self, far from being some ultimate ontological reference point, as it has been since at least the Cogito, is an epiphenomenal construct. To understand the significance of this move, you have to consider the foundational function the notion of the substantival self has played in philosophy. Now, we infer a substantial, ontic center to the psyche, much as we once inferred an ontic center to the natural order, via the concept of God. The reasoning in both cases is the same: if there's an orderly web (to borrow Dennett's own image) to be seen, there must be a center to the web, whether that web be experience or the cosmos itself. Losing reference to that center, we lose our last basis for grounding explanation itself in some kind of reality, even the restricted reality of the self, to which Kant clung for rational grounding.
The placeholder for the center in an ontology is the sign of signs, because it is the organizer of all other signs. The self is perhaps the last refuge of substance ontology in the post-Kantian worldview. Phenomenology has since replaced ontology as rational grounding, and as substance was evacuated from the cosmos, it was pushed inwards, into the domain of “lived experience.” Now we find the last remainder of substance ontology in the notion of the supposed irreduibility of the “qualia” characterizing the first person stance. Kant posited the empirically unknowable central subject as a necessary presupposition for explaining the order and regularity that emerge in our otherwise scattered stream of experience. The Kantian transcendental argument for postulating a unitary subject that underlies and grounds the systematicity of experience can be summed up as follows: “If there is no central self, then there can be no regularity in experience. If there is no regularity in experience, then no explanation is possible, scientific or otherwise. But scientific explanation is, manifestly, possible. So there must be a unitary self grounding the experience from which scientific explanations are gleaned.” Dennett's radical claim, against Kant, is that we can have regularity, and therefore explanation, without postulating a substantial, central self stocked with qualia. Such a postulate, to him, is merely a reified abstraction from underlying, neural-computational processes.
Dennett clearly takes an impish delight in his self-promotion as myth-buster extraordinaire. He takes great pains to show how the higher order unities of aesthetic enjoyment, responsible ethical agency and rational understanding that structure our experience at its highest are nothing more than the by-products of the collective behaviour of “stupid machines” in the brain. According to his multiple-drafts theory of consciousness, the sense of our being grounded in a unitary center of subjectivity at any given time – that feeling that underlies all experience, that we can give words to and call “I here, now” - is a mere abstraction edited out of a confluence of “parallel, multitrack processes of interpretation and elaboration” that are inaccessible to introspection. The “I” that seems attached to every datum of my experience, making it “mine,” turns out to be a linguistically reified narrative construct. It represents the draft that momentarily trumps all others, the one that best streamlines the cacophony of parallel inner processes of endless revision into some kind of provisionally coordinated, working whole. There is no “central meaner” that corresponds to the linguistic sign, “I,” nor anything remotely akin to a causal agent in the brain. Unitary, centralized consciousness is a pragmatic “user illusion.” “Consciousness Imagined,” indeed. If anybody thought that the last vestige of Substance ontology that we find in the Kantian Transcendental Subject could hide here, in the fictive “Cartesian Theater” of phenomenology, Dennett would disabuse us of this notion.
In his chapter, “The Reality of Selves,” he describes the self as a narrative and pragmatic “principle of organization.” It is not much of a stretch to say that, according to his theory, we story ourselves into existence, much as we are storied by others. “Our fundamental tactic of self-protection, self-control, and self-definition is not spinning webs or building dams, but telling stories, and more particularly connecting and controlling the story we tell others - and ourselves - about who we are.”
He describes how making a self has to do with boundary creation, first via biological, then via cultural means. Boundary production starts with fencing off one's turf in physical space, as a cell forms a semi-permeable lipid membrane around itself. It culminates with fencing off one's turf in symbolic-linguistic space via the narrative identity-kits with which we shape our experience into a whole that bears the mark of our historic, cultural, and social backgrounds. Our narratives are thus our most powerful tools, as a species. The way that a spider spins webs to gain power over its environment in shaping it, so do humans gain control by reshaping their cultural environment via their posited narratives. Remember, consciousness to Dennett -just is- the epiphenomenal glow that attends a powerful, winning narrative structure. Persistent identity is persistent narrative. The boundaries of consciousness are the boundaries supplied by given patterns of narrative formation. Enter Foucault's analysis of the relation between narrative patterns and entrenched power structures.
“And where is the thing your self-representation is about? It is wherever you are. And what is this thing? It is nothing more than, and nothing less than, your center of narrative gravity.” Far from being a necessary ontological postulate of reason, a “thing” or a “place” where you find your standing in the world, consciousness turns out to be a fiction that is successively neurally re-instantiated according to variable cultural patterns of stimulus-selection. The only “center” that we in fact have is an imaginative construct, a “center of narrative gravity.” There is no agency, no you, involved in this ceaseless editorial process. “You” are quite simply storied into being by your brain's underlying computational processes. The core schema of “you” emerges as a means of simplifying the discordant mess of neural processes so as to produce a coherent map of reality that your brain can use to orient itself in the world.
So how do you get rational integration and unified conscious experience out of the collective behaviour of irreducibly plural interpretive strands of neuro-computation? To explain that, Dennett introduces the metaphorical device of a piece of neural “software” that, much like a serial computer, creates a step-by-step narrative thread out of the multiplex cacophony of conflicting reality-takes. Whether and how reason can mysteriously emerge from such a clamour of “stupid” computational “processes” is unclear. The metaphor does all the argumentative work here by painting a picture, and assures us that somehow, it just all hangs together like-so.
Mental imagery is where we'd be most tempted to posit some kind of substance to experience, in the form of qualia. It is precisely this domain that he surveys at some length, although from what I've read elsewhere (Evan Thompson's Mind in Life has a great chapter surveying the current state of imagery research), the imagery research so critical to his philosophical argument is far from being conclusive. Mental images, he argues, are what most supply us with the illusion that there is what he calls “figment” or “plenitude” to experience. Above all, it is images that give the feeling of continuity. But it's computation all the way down, if you look at experience (of colour, for instance) from a neural processing perspective. Because he characterizes experience as a “theoretical, narrative construct,” there is no “hard problem of consciousness” for him to contend with. The problem, he hopes, vanishes through redescription via the reigning computational metaphor.
And what about our nagging sense that experience comes as a sort of system, a perceptibly integrated whole, in short? “What we actually experience is the product of many processes of interpretation – editorial processes in effect.” In Dennett's view, it is this ongoing selectivity, simplification and editing out of surplus information so that it comes to fit a manageable pattern that can guide the organism's responses at the time. Moreover, “paradoxically, our sense of continuity comes from our marvelous insensitivity to most kinds of changes rather than from any genuine perceptiveness.” The “unity” that we take to be the measure of conscious realization is thus, ironically, a measure of blindness. This abstractive process of simplification that gives the semblance of consistency, integrity and texture to our stream of consciousness exists because of its great adaptive value, helping reduce noise in favour of survival-relevant information. Underlying the official editorial revision that glosses unity, our experience is, in fact, a cacophonous din.
Nor is there any canonical version of experience that you could poke a stick at and claim to be the real, true, “authentic” version of your experience. There is only an endless proliferation of versions, of drafts, of angles and perspectival takes. This is not only too bad for the poor self striving for autonomy, but also for the whole endeavour of philosophy and science to produce a complete explanation of anything. A Kantian quest for monolithic, universal principles of perspective-taking (which could provide the basis for explanation) was a neurally unrealizable fantasy. There just is no rationally unified perspective in the brain.
Since the interpretive pattern that constitutes “consciousness” is configured according to parameters learnt via cultural indoctrination, “pure” phenomenology turns out to be an exercise in cultural description. Far from offering any privileged access to my mind, it turns out to offer no access at all, since all that swims up to the top from this preconscious swarm is cultural script. Dennett's ontological commitment to a mechanistic-computational metaphor ultimately compels him to devalue first person evidence by claiming its reducibility to a “heterophenomenological” approach that a priori assumes that verbal reports give us sufficient purchase on lived experience. Every dimension of experience that doesn't fit into the constraints of his methodological presuppositions falls through the cracks as so much fictive dross.
That experience is to a large extent a formal construct is an insight at least as old as Kant. However, what Kant and most of his successors had and what Dennett wants to do away with is the residual ontic substrate that was held to necessarily underlie the construct. Dennet would slice away even this and in its place substitute a free-floating tissue of narrative monologue, supported only by distributed parallel processing.
Dennett's overriding motive is, I think, at bottom noble. He is troubled by the proliferation, in past attempts to explain mind, of homuncular, anthropomorphic, perpetually-unopenable, and intuitively-pleasing black-boxes which are supposed to designate the terminus of explanation. Where scientific analysis fails, a suitable homuncular resting point – a pseudo-explanatory myth or fiction – can be inserted. However, his solution to the problem substitutes one mode of pseudo-explanation for another.
Terrence Deacon, in his “Incomplete Nature,” called the eliminativist pattern of pseudo-explanation a species of explanation by “golems.” “Golem” accounts are attempts to describe the phenomenon of mind by dissecting it into mechanistic/computational parts. Dennett's view of mind as an information-processing device constituted by the joint functioning of myriad mini “stupid machines” in the brain is clearly a golem. Deacon shows how attempts such as Dennett's to purge anthropomorphic black-boxes, or homunculi, out of scientific explanation only end up being forced to pay their dues to the qualitative loci for which homunculi are “place-holders” by bringing ever-more “cryptic homunculi” into the picture, usually in the form of “golems,” which are “fractionated homunculi.” Deacon offered an elegant argument showing that explanation by golems is a cure worse than the homunculus disease, because it proceeds by presupposing ever subtler homuncular properties (such as informational, representational, and functional properties) at lower levels, without explaining them. Thus, the impersonal computational machines projected into the brain are treated by Dennett as ultimate loci of representation and information, without explaining how these representational and informational relations emerge.
Deacon showed how such a view takes informational relations out of the larger dynamic context which makes them possible and which grounds their real-world reference. It relies on an abstract, engineering definition of information which presupposes extrinsically-imposed reference – a human interpreter who can fix the representational relationship, or specify what the information is about. In contrast, he shows how information in living organisms is intrinsically interpreted, by virtue of the role it plays in the self-organizing dynamics of life. Thus, Dennett's reductionist approach cannot explain end-directed phenomena such as information, and representation, even as it presupposes these by inserting them, as “cryptic computational homunculi,”to make its eliminativist explanation work.
I would agree with Deacon that Dennett's hand-waving pseudo-explanatory insertions, at crucial parts of his argument, of golem metaphors can't be seen, as Dennett claims, as just descriptive glosses to be replaced by the more complex neuro-computational explanation that is supposed to be forthcoming. Rather, they must be seen as the places at which efforts to explain away phenomenology break down.
Lastly, the Kantian challenge of grounding explanation without the postulate of a unitary self remains. What is the epistemic status of Dennett's theory, if correct? If correct, and multiple drafts are all there are, then Dennett undercuts his own theory's rational basis. The whole truth of the theory is predicated on Dennett's (and our) capacity to hold together a synoptic, rational perspective on our minds that is more than just a momentary coalescence of distributive parallelism and interpretive pluralism. Otherwise, truth claims – even Dennett's – would be merely pragmatic fictional simplifications which are also inescapably distortions of the facts. An empirical theory that “explains away” the unity of conscious experience as an epiphenomenon undercuts its own rational basis.
The theory, as such, can't even be coherently articulated. In articulating it, you refute it, because you presuppose the first-person unitary subject that you attempt to explain away. More troubling still, it doesn't seem to be capable of informing any possible perspective that we can take on the world.
An empirical, homunculus-free explanation of consciousness need not be a reductive golem. Such an account is provided by such thinkers as Damasio, Deacon, Thompson, and Lakoff, whose physicalist theories of mind are nonetheless developed in dialogue with first-person accounts of lived experience.
Ultimately, as Thompson points out, experience is, inescapably, our original guide, since the content of any of the concepts that frame a theory of mind can only come from our intuitive, first-person experience of our own minds:
“To deny the truth of our own experience in the scientific study of ourselves is not only unsatisfactory; it is to render the scientific study of ourselves without a subject matter. But to suppose that science cannot contribute to an understanding
Is it possible? Is this going to finally be the book that explains the mystery of consciousness?
No. No it is not.
What would it even mean to explain consciousness? Reminds me of Hitchhiker's Guide to the Galaxy where they build this ultra powerful supercomputer to finally answer the mystery of "life, the universe and everything," only to then realize that they don't actually know what the question means.
I love love love this book so much that I am hoping that when I die, the crime scene investigators will find it clutched tightly in my hand and will all have to read it very carefully perhaps to get clues about who killed me and then they will forget completely about investigating the crime and start totally getting into this astonishing book instead and will tell all their crime scene investigator buddies who will read it and tell their buddies and then everyone in the world will read it and the whole world will be different for having read it and appreciated it except for like three or four chumbuckets in Duluth or Boise or wherever who don't read books and who will feel so left out and mope around saying things like "oh dear, how come everyone else in the entire world is reading that amazing Dennett book on exactly how consciousness works and all we are doing is just chewing on these lousy grass stems and watching old Flipper reruns and what a crappy life we have, and by the way whatever happened to that guy who died under those mysterious circumstances a while ago, did they ever figure out who killed him or what is up with that?"
A hard book to plough through and one that is so careful and meticulous that it never reaches an interesting or clear-cut conlusion. Dennett takes hundreds of pages to refute the idea of consciousness as a sentient observer sitting inside man's brain (a concept known as "the Cartesian Theatre"). I could have agreed about that being untrue in half a page. When Dennett has finally finished explaining what consciousness is not, he disappointingly admits that he does not have a good alternative either. Ok, he tentatively puts forward an image of primitive man starting to utter noises to himself as a possible early start of conscious reflection. But that theory is hardly more impressive then the Cartesian Theatre that the rest of the book wanted to prove wrong. "Consciousness Explained" is an academical book that is totally useless, since it doesn't deliver any of the promises contained in the title.
Yes, the title is audacious. Yes, it's not a perfect book. Yes, the subject is extremely complex and really smart people fight about it in prestigious journals, etc.
But Dennett has some fine ideas nonetheless. I go through periods of swinging in one direction and back again when it comes to what I'll just call the "consciousness wars." But lately Dennett's ideas are striking me as more and more correct (and I've always leaned in his and the Churchland's direction since I first began looking into these issues, maybe about two years ago).
For some extremely brief, but exciting (probably more so to people already immersed in the field and the debate) overviews of his position(s) check these short videos out:
There were parts of this book that were quite difficult and that I would probably have to reread a few times to fully appreciate, but overall it was a lot easier to read than most works of philosophy and/or science that deal with the same subject matter. There was a lot to think about here. Dennett may not be correct in his several models of how consciousness works, which he labels with colorful names, such as "multiple drafts" and "pandemonium," but he is honest enough to admit that they are just models. He gives solid reasons why his models or something very much like them could be correct, and he offers ideas for how his models could be tested in controlled experiments. And in any event, I was completely convinced that the intuitive ideas that most of us have about how the mind works, which Dennett labels as the Cartesian Theater, are completely wrong. Our comprehension of the world is not a visual story that plays out in our heads, even though it may sometimes seem that way. And it takes very careful thinking to avoid falling back into that intuitive trap again and again in new and different ways. I avoided reading Dennett for years, but that was a mistake. I'm looking forward to exploring more of his works.
It is hard to know what to say about this book. It contains a lot of interesting information and is very readable but it is also deeply confused. Dennett is clearly fascinated by the brain and keen to find a theory to explain how it works. As I am not an expert in this area, it is hard for me to assess whether what he puts forward is either new or interesting. What is most striking (and annoying) for me, however, is Dennett's philosophical naivety and lack of sensitivity to philosophical issues. Much of the book is driven by the attempt to locate consciousness and he often writes as if we might actually have found this in the brain. Rather than recognising that the question "where is consciousness located in the brain?" is confused, he suggests that we need to give up our naive view of consciousness and recognise that all there is are the things he can actually find in the brain. I imagine that if Dennett were interested in mathematics, he would be in favour of ambitious rocket missions to discover the Platonic realm in which there are shapes that perfectly obey the rules of geometry and that if these missions fail, he would argue that we have to give up geometry or adapt it to the realities of the shapes we actually enter in the physical world. Dennett claims to have learnt a lot from Wittgenstein but he persists in seeing introspection as consisting of accurate or inaccurate reports on inner events. He then thinks that scientific research can give us a fuller and more accurate account than introspection. But this is confused. Take the example of dreams. When I tell you my dream, what is interesting for me (and possibly for you) is my account of my dream. If scientists could strap something to my head and give a totally different narrative account of what went on in my head during REM sleep that might also be interesting, but the accounts are different rather than in competition. If I am puzzled as to why I saw you in my dream, the scientist's claim that actually I saw my father may raise new puzzles but it does not help me with my initial puzzlement or necessarily eliminate my wish to reflect on why I dreamed of you. If someone gives the scientist the power to forbid me to use the word "dreamed" in relation to my account, then I will have to say "I quasi-dreamed of you" and when I wake in the morning I will have to try to remember to say: "I had a strange quasi-dream last night", so that the confusion between the language-game I am playing and the language-game the scientist is playing is harder to fall into. Another general confusion that pervades Dennetts' book is the failure to recognise the difference between conceptual issues and empirical issues. Faced with a philosophical problem, Dennett hopes that a mass of empirical research can resolve the issue, but if we don't know our way around our concepts, then trying to apply the concepts in all sorts of unusual situations is only likely to make matters worse. If you are lost, the solution is not to try to move more quickly, but to stop and try to make sense of where you are and where you came from. So perhaps Dennett has produced an interesting theory of the brain, but he is certainly deeply confused about consciousness and all the related issues he mentions (such as the nature of the self). He may be a good purveyor of scientific research, but he has little to contribute on philosophical or conceptual issues where he seems to lack both aptitude and interest.
Consciousness explained? Well, no, not exactly. But a brilliant book nonetheless, despite the audaciousness of the title (though I must admit that Dennett concedes that his "explanation" is far from complete and that cognitive theory is really still in its infancy--or at least it was when this book was written). I only read it recently, and perhaps it is a bit outdated for a book about the ever-changing fields of cognitive theory, neuroscience, and psychology, but, if anything, this book does a really good job of refuting the farcical traditional theories of consciousness (goodbye Dualism, Cartesian Theater, Ghost in the Machine) and presents a workable alternative theory: the Multiple Drafts Model. I won't attempt to explain this model here in just a few short sentences; you'll have to read it for yourself. Suffice it to say that not only does this book illuminate some of the darker aspects of our conception of the mind/self/consciousness, it is also well-written, entertaining, and really not that difficult a read for such an esoteric, slippery thing as consciousness.
Imagine if a bat was raised in an unbatty room, never watched zombie movies, and only ate black and white Chinese take-out food. Now imagine if the bat only seemed to be a bat, and the zombie movies that it didn’t watch were of zombies acting like humans would act if they were acting like zombies acting like humans. How could we say, or at least acknowledge precognitively to appear to say, that the qualia of the unbattiness of the room coadaptively represented the epiphenomenological non-Chineseness of all anti-food experiences? We couldn’t! But here's the trick; it only seemed to you that you were really imagining this, but you are in fact the zombie that the human-acted-zombie is acting like, and your neuronal excitations are by definition unbatty because you are in the state of acting like a human. It therefore becomes clear that this is the position we must take if we are going to review Consciousness Explained, and yet when you look closely, there is no reviewer, and therefore, there is no review. I give it 3 bats. -Otto
I must say, The reason I'm giving such a low score is because I expected much much more from this book. Instead I faced with explanations of phenomena that could have been explained much clearer. And the lack of answer to my question being: How neurons give rise to consciousness. (To be perfectly clear, I do not mean how the mind works or how mind and brain are related. Rather, what gives rise to a cultivated sense, a feeling, a qualia.) And I expected a physical theory, instead, he apparently skipped the whole problem altogether.
it's an ambitious project, I give him that. But a few points, which are the down fall of Dennet in his illusionism approach: 1. He is too sure of himself, and increasingly intellectually not-humble. 2. Point (1) leads to him becoming more and more condescending. 3. Point (1) also leads him to not to be able to understand opposing sides. 4. Point (3) leads him to make strawmen of his opponents' theories. 5. Point (4) leads him to end up defending and explaining something else (mind) as opposed to consciousness (or phenomenal properties of consciousness). 6. point (6) led him to write a book and a few papers, in which could have been awesome for theoretical foundation of "mind" but rather mute about consciousness, yet his book's name is consciousness explained.
The thing that really irritates me about philosophers is their lack of empirical evidence to support their theories; they just borrow whatever suits them from whatever discipline, with scant regard for critical evaluation of the data. Dennett makes suggestions that can be easily countered and writes in a way that he is convinced of the truth but yet has nothing to substantiate it; his theories are built on a lot of assumptions, contradict these and his approach withers away. His rudimentary discussion of psychology was offensive; particularly in reference to ethics in his stating of ‘subjects’ and of false problems with transcripts. Further to this Dennett presented the argument that when information perceived goes into the visual and auditory store ‘does it get rapidly forgotten or was it not even consciously experienced in the first place?’; clear evidence shows that these stores have limited capacity and duration and therefore all were consciously perceived but rapidly forgotten; again another useless argument.
I feel like I’ve been lied to. The book title promises ‘Consciousness Explained’ and for the first 200 pages there is no explanation at all; then there are only a few pages of vague description of Dennett’s ‘Multiple Drafts’ theory. Dennett spends an awful lot time trying to disprove the ‘Cartesian theatre’ as if he’s the only person wise enough to understand the fallacy of having a homunculus inside your mind; it begets the question who then is in their mind? I suppose I should have done my research and understood that this book was written in the early 90’s when perhaps people did believe in this and they needed some convincing. I felt I’d been transported back to reading Darwin’s Origin of Species where I was being circuitously persuaded about the truth of evolution; it was completely unnecessary. Dennett writes as if he’s telling a convoluted story and I felt like Dickens’ Gradgrind as I kept shouting in my head to myself ‘Facts, give me facts!’ It also amused me greatly that Dennett seemed to think that the concept of parallel processing was difficult to understand and accept, as oppose to serial processing. I don’t know if it’s because I did A-level computing, or if because modern psychology understands that the brain works like this, but I found myself aghast at his patronising. Is it really that difficult to imagine multiple processes happening in the brain at once and none having central authority in much the same way a computer time shares processing speed?
I think one issue Dennett has is that he has made the assumption that perception is consciousness; which I doubt to be true. He spends a lot of time discussing the processing of external stimuli and how this broaches consciousness yet he fails to acknowledge until his final pages that a blind person is still conscious, so vision has nothing to do with it. He seems to reject any idea of a preconscious or unconscious which is quite clear when he’s describing the dorsal and ventral streams and the research by Gazzaniga on split-brain patients; it is obvious from this that consciousness is in our left hemisphere and tied to language in some way and our preconscious is in our right hemisphere. He also neglects discussion of different speeds of processing within the brain and seems to think that they all occur at the same time and there are delays for things travelling at greater distances; perhaps one is just faster and they appear at the same time. Dennett is also misguided in his discussions on judgement as there has been recent empirical research that we intuit a lot of our decisions from bias due to our hardwiring before any evidence is actually logicised. It further amused me when Dennett proposes that our brain sizes increased so fast at the dawn of the Ice Age but yet neglects that the fossil records are incomplete; most human fossils have been located in hot countries, where there is no ice to preserve fossils until discovery, so therefore there were no fossils being made during the Ice Age that have subsequently been found; it could very likely have been a slow process. Dennett also made a statement that ‘the way a brain represents hunger must differ, physically, from how it represents thirst’ and I found myself questioning then why do we sometimes eat when we’re actually thirsty? He seems to believe our brain is a touch more sophisticated than I would give it credit for. At one point Dennett quotes Marx who stated that ‘Language, like consciousness, only arises from the need, the necessity, of intercourse with others’ but yet this does not explain why feral children like ‘Genie’ did not develop language but was still conscious. Dennett suggests that people with schizophrenia when hearing voices are actually just hearing their own voice talking to them but not realising it is them, but I find this quite a naïve explanation as it fails to accommodate that there are often multiple voices of different genders. So overall he talks a lot of rubbish.
All that said, I did like the idea of parallel pandemoniums and that your brain is coming up with multiple streams of responses at once and the best fit is selected; perhaps I’m more aware of my brain working like this and so explains why I find it difficult at times to say the right thing as I’m somehow thinking of several at once. Yet Dennett rejects the idea of a ‘stream of consciousness’ and I believe that it could be feasible that consciousness is just the observation of processes in the brain which actually supports his idea of a feedback loop and translator between different coded inputs. A large problem with consciousness is that it is but a small part of a much larger operation and our brains do things all the time that we have no awareness of. ‘We’, whoever ‘we’ are, are just a grain of sand in a desert of intentions. I’d like to think I control myself and make rational decisions as much as possible but there is a deeper part of me that is in far greater control and conflict with ‘we’, with ‘I’ and with ‘me’. The only fascinating passage I read was about the self on page 416:
“But the strangest and most wonderful constructions in the whole animal world are the amazing, intricate constructions made by primate, Homo sapiens. Each normal individual of this species makes a self. Out of its brain it spins a web of words and deeds, and, like the other creatures, it doesn’t have to know what it is doing; it just does it. This web protects it, just like the snail’s shell, and provides it a livelihood, just like the spider’s web, and advances its prospects for sex, just like the bowerbird’s bower. Unlike the spider, an individual human doesn’t exude its web; more like a beaver, it works hard to gather materials out of which it builds its protective fortress. Like a bowerbird, it appropriates many found objects which happen to delight it – or its mate – including many that have been designed by others for other purposes.”
Yet there is no explanation of how we intuit a thought, where it comes from and how it starts, but yet we never know how anything starts; the Big Bang, a heartbeat for the first time, the very first gene. Odd that, isn’t it; we can’t explain how something occurs out of nothing and I don’t think looking at consciousness is ever really going to answer that existential question for us. Whilst this book was awful in the delivery of its nonsense it did create much opposition in my mind and so I was grateful for the stimulation it provided to think a bit deeper about a concept I don’t yet fully understand.
This entire review has been hidden because of spoilers.
After having listened to this book, I will never fall for the make-believe just so stories about consciousness again. There is no reason to have to appeal to fantasy to explain consciousness. This book gives a complete story and forevermore I'll be able to not be sucked into false thought processes concerning the topics about the mind.
Metaphysics, when it's at is best is to fill in the parts that physics (or science) is having a hard time explaining because they don't really understand the object and the terms that describe the object under investigation. Dennett fills in these gaps better than any scientist can. For those who need make-believe and should be sitting at the children's table instead of the adult's table they need to read this book and they can move ahead as I have because of this book.
The best way to think about our self is by realizing we are not an analytical point. Euclid's first definition in his "Elements" is that a point is that which has no breadth. The book doesn't make this analogy, but I do, and state that "the I is that which has breadth", and you know you are listening to a remarkable book when you can go beyond the points the author is making because he educates you so fully.
The author defends this by showing why the self is "a center of narrative gravity", by showing how the mind is not like a Cartesian theater with a homunculus (little human) watching the play as the film unwinds. "There is not anything outside of the text", the text is just the final draft we think out loud. But to get there we first go through Orwellian rewrites and Stalinesque theater before we get the final draft from many rewrites. (Don't worry. The author explains this much better than I can. I'm just trying to whet your appetite in order for you to listen to this book.)
The author steps me through the black box of the mind by first discussing the outputs we measure from our responses to the environment. That was the first eight hours of the book. He called that the analytical approach. That part confused me. I'm not a scientist. The next part he called the synthetic part. How we would build that black box step by step. That's the part where I started listening to every word because it just excited me.
Understanding qualia, our emotional experiences, or what Locke would call our secondary experiences, which lead to things being our 'beliefs' or "seems to", is not how to think about how our mind works. When you can change a "seems to" to the 'is' with no lost of understanding just drop 'seems to' and the phoniness of qualia.
The author uses computers, software, and universal Turing machines and Von Neumann in explaining his thesis. You will walk away with consciousness demystified. You'll be on guard against those who use make-believe arguments to defend a world that doesn't exist.
This book is over 20 years old. I only wished I had discovered it when it first came out. It would have stopped me from wasting my time with people who don't understand that we have ways of thinking about the world that is not dualistic and doesn't need special make-believe explanations to explain who we are as thinking machines.
I almost never change the speed of the audio. For this book, I did and listened to it at 1.25x speed. Made for a much better listen.
I was trying out a new reading methodology for Consciousness Explained, reading the critiques of articles entered in journals alongside the actual text, which allowed me to see both the illegitimacy of some of the criticisms and the serious problems with some of the text. I strongly recommend that methodology. I'll be trying it again at some point soon.
Dennett's text is considered one of the most significant texts in modern philosophy of consciousness, which is odd since there are some major criticisms of it that seem very strong. I strongly recommend the critiques by Ned Block (Journal of Philosophy, 1993) and Colin McGinn (Philosophical Perspectives, 1995), which point out that Dennett actually doesn't construct a theory of consciousness.
The most powerful part of the text, which has largely been accepted and adopted even by many of Dennett's dissenters, is the discussion of the Cartesian Theater.
That said, the rest of the text is kind of hard to read, since it seems fairly weak. Dennett himself is constantly hedging his bets, acting as a pragmatist. He does this extensively in his discussions of heterophenomenology and the Multiple Drafts Model. The problem, though, is that the MDS is not a model of consciousness, but a model of content and the way in which content relates to consciousness.
There are a lot of reviews on here that are critical of Dennett but admit to not having read the entire book. I'm not one of those people. I'm also not going to say that "Dennett reads like Derrida" or something silly like that. The fact is, Dennett has a very unique and engaging writing style, but I think it's worthwhile to remain skeptical of many of his claims, and to try to follow whether the arguments that he paints as in contrast to his own are actually being addressed by the content that follows. It's a worthwhile read for those who are interested in consciousness, and content, but I definitely don't recommend this as a first read. Despite its accessible writing style, the actual content may lead to later confusion, as much of the debate has changed since the writing of the book.
This entire review has been hidden because of spoilers.
It wasn't easy, and many times I felt like Homer Simpson trying to learn how to market a bowling alley (http://www.youtube.com/watch?v=SEqxer...), but getting through this book and tackling the weighty subject matter was well worth the investment. And I'm not kidding about the Homer reference: Dennett posits so many amazing points based upon areas of thought of which I was hopelessly clueless. I would have to set this book down and do some research to just get a baseline to follow his explanations.
For any of my Goodreads friends that want to really explore the big "I", and delve into the what of our conscious selves, I couldn't recommend this more. But the recommendation comes with a warning: if you give the book its due, don't skim passages / sections and really work yourself to understand what Dennett is trying to say, you could come away from this book with a changed view on yourself, your fellow humans and this wonderful, crazy brain of ours.
A bold book from my favorite philosopher-scientist that aims to build a framework for tackling perhaps the hardest question humanity has ever asked - "what is this conscious experience?" As in his other books, Dennett is adept at weaving the "soft" thought experiments of philosophy with the "harder" experiments of the scientific community. Some of his most triumphant points don't have the impact they may once have carried, as much of his material has been accepted (or disproved) in the last two decades of the rapidly evolving field of neuroscience. Despite its age, this book is a stellar introduction to anyone trying to approach consciousness.
Dennett's thought experiments and suggested activities for readers shed light on some fascinating phenomena of consciousness, including sensory dislocation & extension of self to tools and blind spots & the overly assuming nature of vision. This second investigation I found to be a powerful metaphor for much of the simulation that the brain performs in crafting our sensory experience. The discontinuity of consciousness is so striking particularly because of its apparent continuity. The brain doesn't so much "fill in" the blanks as it ignores their presence. Dennett makes the important point that this absence of representation (ignorance) is not the same as the representation of absence ("filling in").
The three themes of Dennett's that resonated most with me were the relationship between time & consciousness, information sharing and information barriers in the brain, and consciousness as cultural software.
1) Noticing the varying speeds of sensory signal propagation outside of the body (light vs. sound vs. chemicals) and the varying speeds of neural signal propagation in the brain, Dennett points out that the "present" for us is really more of a "smear" in time rather than a "point". He presents his Multiple Drafts Theory of Consciousness to show that in such a situation, different parts of the brain must act on different sets of information and, therefore, there is no single conscious experience. This is perhaps one of the most profound points that Dennett explores and he does so frequently throughout the book. Dennett also points out that temporal order outside the mind need not coincide exactly with temporal order as represented in the mind, though the two are correlated.
2) With so many specialized areas developing at different periods in human evolution, information sharing in the brain can be quite haphazard and arbitrary. The recognition that information may be present in one area of the brain but entirely unaccessible to another area is essential to understanding many functions and quirks of the brain. This is evident in many popular accounts of language disorders but Dennett also explores what this suggests for the evolution of consciousness. He imagines that early man armed with protospeech might have used "vocal autostimulation" (thinking out loud) as a means of bridging missing connections in his thought processes. In other words, if there's no path from A to B in the brain, there might have been one from A to speech to hearing to B! This clever circuit could then have evolved into silent thought for more privacy and eventually developed into the "mind's eye" visual experience of modern man. Even within the brain, there are likely many inefficient intermediary representations developed to bridge the internal "communication problem." Beyond evolutionary explanations, this idea is also highly suggestive of neuroscientific approaches to creativity. Speaking out loud, doodling, and gesturing to oneself may be more than just nervous ticks or distracting habits; they may instead be integral yet inefficient attempts to circumvent the missing information pathways in the brain!
Dennett also includes a list of "primordial facts" that he claims any theory on the evolution of consciousness must explain. I found them insightful and important enough for any neuroscientist that I've included them here verbatim:
There are reasons to recognize.
Where there are reasons, there are points of view from which to recognize or evaluate them.
Any agent must distinguish "here inside" from "the external world."
All recognition must ultimately be accomplished by myriad "blind, mechanical" routines.
Inside the defended boundary, there need not always be a Higher Executive or General Headquarters.
In nature, handsome is as handsome does; origins don't matter.
In nature, elements often play multiple functions within the economy of a single organism.
3) As for the development of consciousness, Dennett proposes that viewing consciousness as cultural software provides an instructive and productive framework. His evidence includes the relatively recent development of consciousness (and therefore the reduced possibility that it is hard-coded). So why does consciousness still seem to be similar across cultures? Hardware biasing - we're all still working with roughly the same base. Some interesting results of this hypothesis are that some humans may not experience consciousness, particularly babies and special cases of children who developed with very little social contact.
Just as evolution is a difficult topic to write on given that our language is peppered with words conveying "intent", consciousness often has Dennett tripping over his own words. He fares far better than most, but be forewarned - books on consciousness can't help but be clumsy.
In addition to being an excellent introduction for me to many theories on consciousness, this book has piqued my interest in the consciousness and cognitive development of children and the general AI framework known as SOAR.
I'm fascinated by the topic of consciousness, and this book appealed to me because I thought it was going to give me a better understanding of what consciousness really is at its core. Well, it doesn't. Unfortunately, it doesn't explain consciousness like the title claims. That's fine though, because I'm pretty sure it's something nobody's ever explained or understood. What sets this apart from Descartes and the 19th century books I've read about this topic is that it brings modern science into the equation (or at least, 1991 science). He refutes Cartesian Dualism and uses Darwinism and neuroscience to try to explain what consciousness is. Science is great, but it doesn't explain consciousness. He claims consciousness comes from the brain, and I think that's probably true, but he far from proves it. And nobody has ever disproved Dualism to my satisfaction. I'd suggest you check out Nietzsche's and Schopenhauer's theories of consciousness because although they aren't as scientific, they're very substantive. This review sounds like I'm bashing Dennett and I'm really not. I respect him as a philosopher and liked the book. It just doesn't explain consciousness. Perhaps From Bacteria to Bach and Back will come closer to explaining it.
This brilliant book by Dennett, one of the best philosophers of our age, will recreate the way you think about consciousness and build a strong foundation for a scientific, rational explanation of it, inspired by a perfect blend of neuroscience, computer science, psychology and linguistics. Most of us think of the conscious-self as a decision-maker, a driver of the train of thought. This image is shattered by convincing the reader that there is not a single line of continuous "train" of thought and there is no central point where "it all comes together" . There are multiple inputs, little particles of quasi-narratives, coming from different parts of the brain with different agendas and competing with each other to make their agenda "win", being written and rewritten over and over again in the process. This theory (hastily summarized here by me) may seem counter-intuitive and maybe even outrageous at first, but the author does his best in slowly chipping away at the established beliefs about consciousness such as the Cartesian Theater and convincing the reader at least to have a new, more rational perspective. I was also delighted to see that my own recent theories on consciousness are endorsed here. So I may actually be a bit biased in giving this book a perfect 5-star rating. But it deserves absolutely nothing less than a 4 out of 5 for anyone interested in explaining consciousness.
I read this book in several overlapping passes over the years before finally reading it cover-to-cover. Given that the premise of the book is that consciousness is an emergent property based on Multiple Drafts built atop an array of certifiably dumb Reactive Dispositions, I think you should be allowed to read the book any which way you want that helps you get at its message. But read it you really should. Dennett succeeds in gently persuading you to think about a Very Hard Issue (and consciousness ranks high in the set of such issues), even if you find yourself coming up short and frustrated, and even if you don't quite believe his suggestions.
The book is considerably old now in a fast-changing field, and it has taken more than its share of lumps, but for a lay reader, it remains a wonderful, non-dumbed-down introduction to a subject that we're all simultaneously very familiar with (we all have a consciousness that only *we* can experience) and very discombobulated by (we are such a captive audience of our own consciousness that we cannot easily suspect its canny illusions).
Dennett is a patient overexplainer: he isn't afraid to be the Voltairean bore who "tells everything". In this case, I found the overexplanation to be most welcome. I am perhaps being a wee bit unfair about his prolixity: his nerdy humor, his easy access to an armamentarium of analogies, and his willingness to anticipate his reader's objections at every step will eventually shepherd you through this difficult book -- difficult only because the subject is difficult. Just don't expect to read the book in one stretch. Dennett is not quite as poetic (or mythopoetic) as his predecessor in the field, Julian Jaynes, but he has his own virtues as a confident, hard-nosed reductionist who knows his stuff and who inspires confidence in his reader that he will deliver the goods. What he's attempting is nothing less than explaining that consciousness can be and is indeed an edifice built up from basic life instincts, honed merely by the dumb, relentless, and violent logic of evolution. Whew!
One of the criticisms laid at the feet of any explanatory venture of this sort is that a currently trending technological phenomenon is being lazily used (i.e., misused) to explain Deep Stuff. Now that computers and (some forms of) artificial intelligence are unmistakably running our world, the mind is inevitably being viewed by many as software, running on carbon brain rather than silicon chip, and this view is deeply resented by many others as distasteful or blasphemous. (The fact that carbon and silicon are next-door neighbors on the same column in the Periodic Table makes the comparison between brain and computer compelling to some.) But we, or at least our brains, are not at all like computers, critics insist; we have feelings and thoughts, we are able to do things that computers can't, and, even more damningly, we are charmingly *unable* to do some "mindless" things that boring computers easily can, so there.
Dennett is unfazed by this, claiming that the inability to see the kinship between the two is merely a failure of the imagination. (The stakes are high in this battle: We may soon need to contend with the possible consciousness of computer-based systems, and the urge to draw a clear line in the sand is strong. We are already finding machines disquietingly powerful; do we really also want to worry about their being conscious?)
The first task that Dennett sets for himself: How can consciousness of any kind possibly be generated from matter without invoking a miracle? Starting from nothing more than basic "reactive dispositions" -- the dispositions to feed, fight, flee, or mate, aka the 4 F's (page 188) -- that we inherit because we are a species that evolved to survive, Dennett gives a pleasing narrative of how these dispositions can be combined to give multiple drafts of info about the world. With time, some drafts are abandoned, new ones added, and others refined to give more and finer-gradated discriminatory details. The info is not just about the external world coming through our senses, but also about our own rapidly increasing pile of info.
Consciousness is not one draft or combination of drafts that gels at any crucial instant, it is an evolving consensus of the drafts with time. Later sets of drafts can have a higher degree of self-reflectivity (because these are later processes than the immediate 4F reactive responses). We therefore typically rate them higher for consciousness, but they're not strictly necessary for us to function ok. We can do many things unreflectively (= unconsciously) if we're good enough at the task and don't need to waste resources monitoring ourselves too closely.
To be sure, the info that is gathered in the drafts is not always infallibly accurate, but it is capable of constantly being added to and fine-tuned. At this point, this is not beyond the capability of plausible, purely technical tools acting loosely together (some components can be in open conflict, with other components logging that conflict). Even generating info about the system's own behavior -- self-monitoring -- doesn't merit the miracle tag. There is no one thing that this info is presented to, however, although that is the powerful User Illusion that we are so prone to. Dennett flogs this horse quite a bit, and rightly, as it's so easy to imagine that an agent inside us is receiving this info and doing far-reaching things with it, until we realize that we are actually trying to deconstruct this agent in terms of simpler processes, not keep bringing him back into the story as an axiom.
The second task is dealing with the dreaded qualia, the "internal flavors", that attend our conscious experiences. It's not just information that's coursing through our brains, surely? It's also the peculiar, ineffable enjoyment and aversion and the various things in between that color (!) that information. It's all very well for the Multiple Drafts to get a self-reflective mechanism into play from nothing but a handful of basic life instincts, but every reader has such intimate knowledge of the quality of their own consciousness, that they find it hard to imagine that even a very cleverly engineered machine could even remotely approximate it, even in principle. Even conceding that the particular qualia are allowed to be different from our own, can a machine have any qualia at all, and if not, can it legitimately be called conscious?
One answer is that, d'oh, we are actually machines ourselves, albeit fashioned by evolution, so the category distinction is false. Thus, if we think we have qualia, so can machines. But Dennett's larger answer, and this is going to strike a lot of people as simply too huge a pill to swallow, is that there are no qualia, at least not in the fundamental way that we seem to think. They're perhaps just the fumes or bristlings set off by our particular "reactive dispositions" which will of course have their counterparts (not the identical ones of course) in the machine. After all, even among humans, we concede that the qualia associated with color perception can be different without affecting the value and merit of the related consciousness.
To recycle the perennial example, it's quite possible, we intuitively feel, that the qualia I feel for red and green respectively are exactly the opposite of what you feel, and yet not only do we *not* use this to question the consciousness of either of us, there is no observable difference in our behavior to cause an issue (beyond perhaps having a color preference?). Indeed, if we consider various kinds of colorblindness, the qualia must, we suspect, be different: If I see red = green and you see red != green, then there is at least one color that's producing a different reaction in us. Again, this comes with no consciousness penalty. (Obviously, these examples invoke individuals that are sighted. There are other perceptions that can be considered similarly for potential qualia differences.)
But is color perception, when and however it happens, just information processing, without having to consider any of the rich internal feeling we instinctively attach to it, even among the colorblind, even if we don't all have the same identical feeling? Yes, says Dennett. Indeed color is our way of distinguishing some light wavelengths, as simple a measuring task as any. It could well have been a plain number. However, since we're operating with the basic reactive dispositions we are saddled with, and since our purpose was to survive, not to be color theorists, we didn't assign numbers as a wavelength-meter in a lab machine might have. Instead, we took our survival reactions to the short light waves of the open sky or bodies of water, the middle-length waves of vegetation, and the long waves of fruits and blood, to fashion our own clumsy but oh-so-dear-to-us tri-cone-based wavelength-meter for the wavelengths that mattered for our survival. Color *was* code to distinguish objects of interest, so the phrase "color coding" would have been a tautology when we were still in the savannah.
It turns out that the color recognition system is good enough so we're seeing color everywhere, and we can create our own information using it, and we've been using it to enhance and degrade our lives ever since. (Indeed, with the advances in chemistry from the late 19th century on, we are inundated with color info, some of it useful, most of it just drug, in a way that our ancestors never were. Even within my individual history of using computer text editors for writing programs and text, I originally perforce used monochrome colors and thought nothing of it. Now syntax highlighting (i.e., color for distinguishing things that I was already distinguishing just fine) has taken over. It happened gradually enough that I never noticed the drug being jacked into my system, and now it's an effort to go back to monochrome editing.)
Would we be taking advantage of color this way if our perception was more numeric instead of this fruit/blood/foliage/water/sky mishmash, with red simply marking a higher number on some internal linear monitoring dial compared to violet? Perhaps, but also perhaps not. And would the latter necessitate a revision to our status as conscious beings in any meaningful way? Dennett strongly implies no.
Regardless of the persuasiveness of his content, some people may take a visceral dislike to the style of Dennett's writing. He is an earnest, unabashed materialist-reductionist, which rubs a lot of people the wrong way, but he does make a splendid case that seemingly incredible emergent properties can be explained by reductionism. The disadvantage of seeing precious emergent properties this way is of course that they won't be permanent, and will go away when the substructure falls apart. This is bad news for the soul, whether individual (it's a mortal construct) or universal (even if it exists, it by definition cannot be experienced by its individual cogs, such as us). Even if the loss of potential immortality is not a concern, this kind of reductionism can be a depressing gateway, for those who need rules to be compassionate, to a- and even im-morality, which may explain the animus to philosophers of Dennett's stripe. Dennett does bring up the ethics question, although he valiantly tries to put a positive spin on devising new ways to avoid the possible moral abyss opened up by our tasting of the Tree of Knowledge, and the possible casual devaluing of humans and our fellow lifeforms. I think he rather misses the point, perhaps wishfully so, that the nice folks willing to come up with these strategies are not the problem. Anyway, this is not the strong part of the book, but I'm willing to give it a pass because it's really outside the scope of the topic, which is plenty engaging enough.
Circling back to the critique that the author is using an arguably fashionable tech trend as metaphor instead of offering a presumably "genuine" explanation, I am not persuaded this is bad at all. Metaphors are the stock-in-trade of language (some would say they're identical), and unless we're thinking of an extralinguistic device that is ipso facto indescribable, there is no other real option. Computers and AI are proving to be quite a defensible source of metaphors for narratives about emergent properties. Indeed, software was arguably modeled on human behavior anyway. Expert systems, for example, have been around for many decades now, and offer a good analogy for how multiple small deciders can be combined to give a convincing impression of true expertise: by having facts triggering rules which produce further facts that trigger more rules, etc., they rinse, lather, and repeat their way to advanced discriminatory knowledge and have proven their worth in fields as far-ranging as medical diagnosis, finance, and circuit-fault analysis.
The trouble is they may seem too stodgily reliable and not quirky enough to compare to our mercurial human brains. While computers can take off like a bat out of hell on tangents because of the speed advantage they have over us in some facets of computation, this does suggest that, by the same token, our own expertise in *our* preferred areas may be considerable without needing to be miraculous. We spent all our time trying to survive, to eat without being eaten, and that has given us some core expertise and of course our own peculiar "reactive dispositions" that the computer is not going to share unless specifically and rather pointlessly programmed to do so. (Somewhat like the herculean efforts to make meatless burgers be just like meaty ones, instead of being their own glorious thing.)
While Dennett does push the software metaphor for the sequential nature of consciousness as a von Neumann machine jury-rigged atop a crowd of small-scale reactive agents, the Multiple Drafts metaphor he uses for consciousness actually seems quaintly medieval: 15th-century publishing rather than late-20th-century software! It reminds me rather of the pamphlet wars set off by the invention of the printing press, which led to various hard-fought local optima in both religious and scientific practice. Perhaps he means the Multiple Drafts of his own book writing, in which case he is clearly repurposing something from his personal workflow, and could be rightly critiqued for doing a bit of an analogy reach.
But in another way his metaphor is modern, or rather prescient, since the technology it mimics was neither available nor trendy at the time of his writing (1991). I mean of course the multiple drafts underlying the current big trend in software engineering, viz., distributed version control systems! In particular: Git! At least for now, it's the users that create the multiple commits (= drafts) of a Git repository, and not the repo itself. Still, the way a program takes shape, over time, with multiple competing inputs, with multiple commits, and manages to maintain a continuous identity while accommodating wide-ranging updates and changes to itself, is probably not totally dissimilar to how consciousness itself operates. I don't know how much one can run with this without hurting oneself, so this is a good place for this rambling review to stop at this draft.
Dennett has conducted a wonderful investigation into the nature of consciousness. Not being satisfied to treat consciousness as something ontologically and fundamentally "special", he dismisses some misguided notions of the workings of consciousness which makes it seem as though there has to be some sort of "center or awareness" in which it all comes together along with the related notion of conscious experience as something which has further unexplainable phenomena, qualia, as its building blocks. Dennett insist on treating consciousness as just another biological phenomenon, needing an explanation in terms of more fundamental, unconscious, building blocks thereby dismissing the idea of qualia as ungrounded (unless understood as something further analyzable in terms of phenomena that do not exhibit qualia. These two intellectually unmotivated notions are related in that they both suppose that in order to explain consciousness, we need to find something fundamental which is itself like that which it is there to explain. Need to understand conscious experiences? Postulate the existence of basic states of conscious awareness which can not be "explained away": qualia. Need to understand the nature of a conscious person experiencing the world? Postulate the "Cartesian theater" at "the center of the mind" in which all the processing of the brain comes together to yield the final experience. Both these approaches to these questions are highly misguided: to explain the nature of conscious states, we need an explanation of their constituent parts in the brain, how the processes of the brain amount to experiential states, we do not need there to be, in addition to the purely physical processes of the brain, an accompanying unanalyzable state of conscious experience of the processing; to explain the nature of the conscious agent having these experiences, we do not need to find a further, smaller agent inside the mind, taking in all the results of the processes and experiencing their end result. Being predisposed to naturalistic explanation, as any thinking person should be, Dennett rightly concludes that these other explanations, being grounded in myths and mystery, will not do to explain what consciousness is and are often at odds with experimental results (which is, of course, enough to dispel them). Dennett does not have a detailed account of exactly how the processes of the brain amount to conscious experience (and it would be too early to attempt such an explanation), but goes a long way towards showing how scientific discoveries show us the way to asking the right questions. This is, it seems to me, both his usual approach and the right one. Upon suggesting that the mind and consciousness works a certain way, he accompanies the claim with scientific sources conducting experiments on the issue and sometimes suggest future experiments of his own that would test his thesis.
In the end, Dennett is unclear about exactly it means for someone to be conscious of something other than that it consists in the person being in a state where his or her brain currently processes information regarding the thing of which he or she is currently conscious. This might seem unsatisfactory, and this is perhaps necessarily so considering the current state of our scientific understanding of the mind. In any case, he does not dismiss consciousness, is not a complete eliminativist regarding it (as some seem to think he is), but rather seeks to demystify it, explain how it is that we are conscious beings and trying to convince the readers that we can keep our conscious minds without clinging on to unwarranted convictions of the special nature of the conscious mind. It's all very clear headed and Dennett seems to say almost exactly as much as should be said about the subject: there is no center of the mind in which experience and intentions arise (no Cartesian theater, no central meaner), there are no basic building blocks of conscious experience such that they can not be further explained in naturalistic terms (no qualia), there is no serious possibility of there coming into being creatures with all the behavioral complexity of conscious human beings who are nonetheless not conscious (no zombies) and the mind and all its workings, consciousness included, needs to be analyzed as a naturalistic phenomenon with no prejudices concerning the "special nature" of subjective experience (phenomenology) that is not explained in terms of objective phenomena (the scientific method).
01. Prelude: How Are Hallucinations Possible? 01.1. The Brain in the Vat 01.2. Pranksters in the Brain 01.3. A Party Game Called Psychoanalysis 01.4. Preview
Part I: Problems and Methods
02. Explaining Consciousness 02.1. Pandora's Box: Should Consciousness Be Demystified? 02.2. The Mystery of Consciousness 02.3. The Attractions of Mind Stuff 02.4. Why Dualism Is Forlorn 02.5. The Challenge
03. A Visit to the Phenomenological Garden 03.1. Welcome to the Phenom 03.2. Our Experience of the External World 03.3. Our Experience of the Internal World 03.4. Affect
04. A Method for Phenomenology 04.1. First Person Plural 04.2. The Third-Person Perspective 04.3. The Method of Heterophenomenology 04.4. Fictional Worlds and Heterophenomenological Worlds 04.5. The Discreet Charm of the Anthropologist 04.6. Discovering What Someone Is Really Talking About 04.7. Shakey's Mental Images 04.8. The Neutrality of Heterophenomenology
Part II: An Empirical Theory of Mind
05. Multiple Drafts Versus the Cartesian Theater 05.1. The Point of View of the Observer 05.2. Introducing the Multiple Drafts Model 05.3. Orwellian and Stalinesque Revisions 05.4. The Theater of Consciousness Revisited 05.5. The Multiple Drafts Model in Action
06. Time and Experience 06.1. Fleeting Moments and Hopping Rabbits 06.2. How the Brain Represents Time 06.3. Libet's Case of "Backwards Referral in Time" 06.4. Libet's Claim of Subjective Delay of Consciousness of Intention 06.5. A Treat: Grey Walter's Precognitive Carousel 06.6. Loose Ends
07. The Evolution of Consciousness 07.1. Inside the Black Box of Consciousness 07.2. Early Days 07.2.1. Scene One: The Birth of Boundaries and Reasons 07.2.2. Scene Two: New and Better Ways of Producing Future 07.3. Evolution in Brains, and the Baldwin Effect 07.4. Plasticity in the Human Brain: Setting the Stage 07.5. The Invention of Good and Bad Habits of Autostimulation 07.6. The Third Evolutionary Process: Memes and Cultural Evolution 07.7. The Memes of Consciousness: The Virtual Machine to Be Installed
08. How Words Do Things with Us 08.1. Review: E Pluribus Unum? 08.2. Bureaucracy versus Pandemonium 08.3. When Words Want to Get Themselves Said
09. The Architecture of the Human Mind 09.1. Where Are We? 09.2. Orienting Ourselves with the Thumbnail Sketch 09.3. And Then What Happens? 09.4. The Powers of the Joycean Machine 09.5. But Is This a Theory of Consciousness?
Part III: The Philosophical Problems of Consciousness
10. Show and Tell 10.1. Rotating Images in the Mind's Eye 10.2. Words, Pictures, and Thoughts 10.3. Reporting and Expressing 10.4. Zombies, Zimboes, and the User Illusion 10.5. Problems with Folk Psychology
11. Dismantling the Witness Protection Program 11.1. Review 11.2. Blindsight: Partial Zombiehood? 11.3. Hide the Thimble: An Exercise in Consciousness-Raising 11.4. Prosthetic Vision: What, Aside from Information, Is Still Missing? 11.5. "Filling In" versus Finding Out 11.6. Neglect as a Pathological Loss of Epistemic Appetite 11.7. Virtual Presence 11.8. Seeing Is Believing: A Dialogue with Otto
12. Qualia Disqualified 12.1. A New Kite String 12.2. Why Are There Colors? 12.3. Enjoying Our Experiences 12.4. A Philosophical Fantasy: Inverted Qualia 12.5. "Epiphenomenal" Qualia? 12.6. Getting Back on My Rocker
13. The Reality of Selves 13.1. How Human Beings Spin a Self 13.2. How Many Selves to a Customer? 13.3. The Unbearable Lightness of Being
14. Consciousness Imagined 14.1. Imagining a Conscious Robot 14.2. What It Is Like to Be a Bat 14.3. Minding and Mattering 14.4. Consciousness Explained, or Explained Away?
Appendix A (for Philosophers) Appendix B (for Scientists) Bibliography Index
While Dennett is probably better known to most readers as a grumbly professional atheist, I really don't need any help in that regard, so I went straight to his book on philosophy of mind. I can see why he's a public figure-- he's downright chatty and personable for a chilly analytic philosopher, and at the same time clear and rigorous in his presentation of ideas.
As for the ideas themselves... OK, the multiple-drafts notion of consciousness is something I can certainly get behind, and his attack on the "Cartesian theater" notion, while it seems obvious, is something that really needs to be done every once in a while to clean philosophical house.
But as to how we arrive at that multiple-drafts state, he relies on an excessively inductive understanding of evolution and the brain-as-computer metaphor that seems to cripple cognitive research. I tend to agree far more with people like Searle, Dreyfus, Putnam, and Merleau-Ponty, whom Dennett explicitly rejects.
An extremely frustrating book. There is a lot of well thought out argument here, much of it worth engaging with even where I disagree, and some of the theory of how the brain works was genuinely compelling. In particular the pandemonium model of language production--and, despite my low rating, I like significant portions of the Multiple Drafts model!
But the author falls into infuriating patterns of thought that lead him to advance ridiculous claims that do not all logically follow from the (often) reasonable assumptions he starts out with. He responds to objections throughout, but objections presented by a rather unsophisticated fictional critic who did not always ask the questions I wanted answered. In fact, at least one stunning, seemingly foundational assertion (that a robot capable of second-order observations of its internal states would "think" it was conscious) is put forward with practically no backup or discussion--what does that even mean?? Other bucks are merely passed down the road (it only *seems* like I experience qualia? even if I concede this, the fact that they "seem" so vivid to me is not a lesser mystery than the one he seeks to deflate by denying qualia are real). It is always possible I am misunderstanding key points and/or that my gut opposition to his thesis is impeding my ability to accept genuine truths (I happily concede that this latter may have happened here or there). But as I glance other reviews, including those by experts, it looks like I am not the only one with similar frustrations and confusions about what exactly he is trying to say.
Dennet, by my diagnosis, adheres to a particular type of scientistic worldview that suggests anything beyond the reach of more or less contemporary science either doesn't matter or doesn't exist. Whatever this view's strengths, I think this book shows some of its weaknesses in both style and content. Still, the book is useful as a scientific theory of the brain, a challenge (often successful) against prevailing ideas about the mind, and a demonstration of the promises and pitfalls of his approach. Despite all my complaints, I would genuinely like to sit down with the author someday and ask him my questions and see what he says.
However, I can't give it three stars because the discussions of nonhuman animals are sloppy, un-rigorous, and riddled with both implicit and explicit human supremacism (e.g. calling species phylogenetically far from humans "lowly," which is of course an un-scientific presentation of evolution). This leads him to somewhat hand-wavey moral conclusions that are all the more egregious because he hasn't backed them up in any meaningful way.
I really enjoyed this book for these reasons: - It is long, yet almost always engaging. The author only rarely trailed off in discussions that were meaningless for me. - It is well structured and well-worded - I encountered many concepts that I already had in my mind, but would have been unable to put into words, especially so succinctly - I found a great mix of reasoning, research and anecdotes. - It has philosophical and scientific depth, but a lot of examples from everyday life. I feel like a gap in my knowledge has been filled and I am better equipped for philosophical reasoning and discussions.
This is a dense boi, I didn’t follow all of it, but from what I could gather it was a pretty interesting thesis that runs counter to what you might intuitively think of consciousness. Probably worth the read
DNF. It’s interesting, and actually less of a difficult read than a lot of other books in similar areas, but it’s long and I basically just stopped when I realised I still had a lot left to go and didn’t really feel like carrying on. May well come back to it, though.