Consciousness Explained (1991)
by Daniel C. Dennett (1942-)
I first read this book several years ago, but for some reason didn’t report on it. After reading other books by Dennett, I decided to reread this one.
My most lasting impression from the first reading is that the title overstates the content. Dennett doesn’t actually explain consciousness. This impression remains on rereading it.
What Dennett does is give a sketch that might be useful as an approach to explaining consciousness, and then gives many reasons for rejecting more intuitive approaches that prevent many people from even attending to his explanation of his sketch of an approach to explaining consciousness. This is a worthwhile effort, and an important service; but the title might get in the way for some people. Some of the more interesting (to me) parts and points:
Dennett bases his approach on the results of experiments by a broad range of cognitive scientists: psychologists, neurologists, computer scientists. He is using data and language that were unavailable to earlier generations of philosophers, and so must spend effort (seemingly an inordinately large amount of it) tearing down conceptions left over from previous, intuitive approaches.
In section 4.3 he calls his approach heterophenomenology, a term he seems to have coined to contrast with earlier philosophers’ technical terms. It seems to mean using mixed phenomena to investigate the working of the mind. This is necessary because the mind is not available for direct examination. A key item in the mix is the verbal (or other behavioral) indication of a subject as to what is happening in his mind. Part of the verbal behavior includes understanding the instructions of the investigator, such as “Push the left button in such-and-such a case, the right button in this other case.” Another part is the transcript of the subjects’ responses, correlated with the other aspects of the investigator’s set-up (the ‘cases’) and the other behavior or reactions exhibited by the subjects, often captured by videotape or recordings of physiological states.
The experimental approach results in a sound recording, interpreted by stenographers as a text, then interpreted by the investigator as a sequence of speech acts. Typically, other recorded data in the absence of such speech acts can be about the subjects’ bodies or brains, but not about their minds.
As a result of this approach, the investigation of consciousness is largely about verbal behavior. This mundane fact is often overlooked, and other phenomena are often included in a discussion of consciousness, relying on intuition or prejudice, and leading to spurious, untestable conclusions.
Dennett compares the interpretation of texts from experiments with the interpretation of fiction, such as the world of Sherlock Holmes. Separate from the literal text, readers can agree on characteristics of Holmes’s London. Similarly, investigators can agree on aspects of the mind-activities that led from certain conditions to certain verbal reports.
In section 5.5 he introduces his Multiple Drafts model of consciousness. Briefly, he assumes the brain contains numerous agents or homunculi (which he calls the Pandemonium architecture, after Oliver Selfridge), each capable of making sense of some restricted aspect of the sensory stimuli entering into and internal state of the brain. Each agent has some influence on other agents, including those that drive the body’s verbalization (internal and external). At any moment, many agents are connected loosely to the verbalization facility, and many different verbalizations are being promoted. Only one stream can actually get through the entire pipeline, but the conflicting influences of multiple sources can sometimes be heard in “slips of the tongue”. The multiple ways of interpreting the world are only evident as “drafts” of the history of the mind. As these are mulled over, and compared with subsequent events, they are revised to make a coherent history, the conscious memory of the past. He compares the revisions that occur with the Stalinesque approach of creating false historical events (e.g. show trials) that “explain” past events, and the Orwellian approach of simply destroying evidence and rewriting history. Both approaches can have similar results, narrative histories that fit certain facts, but might be false in different (hidden) ways.
In section 6.2 Dennett describes the brain’s representation of time. The brain doesn’t have a precise clock with which to timestamp mental events. Instead, the brain likely uses the “content-sensitive setting”, similar to the slap of the slateboard at the beginning of a movie scene. The redundancy in relating several different aspects of a content-full situation allows reference by some agent to some aspect of an event, without actually recording the time; and noticing one thing while there is still a trace of activity of another can serve to keep the memories of them ordered in time.
In section 6.5 he describes “Grey Walter’s Precognitive Carousel.” This was an experiment (purportedly, the evidence is not absolutely perfect) performed in the 1960s on patients who had had electrodes implanted in their brains (for other purposes). Walter arranged for patients to watch a slide show, with their finger on the button to advance the slide carousel, while his equipment monitored their electrodes. The patients were told to advance the slides any time they wanted. However, the button was not actually connected to the projector; instead Walter’s equipment detected a burst of activity known to precede the action of pressing the button, and advanced the slides. Rather than being unnoticed, the patients reported that the slides advance before they pressed the button, but so close that they could not stop pressing – some said that they were afraid that their actual button press was going to cause the slide to advance twice.
In section 7.2 he describes “the birth of boundaries and reasons”, a stage in evolution:
In the beginning, there were no reasons; there were only causes. Nothing had a purpose, nothing had so much as a function; there was no teleology in the world at all. The explanation for this is simple: There was nothing that had interests. But after millennia there happened to emerge simple replicators. While they had no inkling of their interests, and perhaps properly speaking had no interests, we, peering back from our godlike vantage point at their early days, can nonarbitrarily assign them certain interests – generated by their defining “interest” in self-replication. …
As soon as something gets into the business of self-preservation, boundaries become important, for if you are setting out to preserve yourself, you don’t want to squander effort trying to preserve the whole world: you draw the line. You become, in a word, selfish. This primordial form of selfishness … is one of the marks of life. Where one bit of granite ends and the next bit begins is a matter of slight moment; the fracture boundary may be real enough, but nothing works to protect the territory, to push back the frontier or retreat. “Me against the world” – this distinction between everything on the inside of a closed boundary and everything in the external world – is at the heart of all biological processes …
Thus are laid the foundation stones. We can now explain the following primordial facts:
1. There are reasons to recognize.
2. Where there are reasons, there are points of view from which to recognize or evaluate them.
3. Any agent must distinguish “here inside” from the “external world.”
4. All recognition must ultimately be accomplished by myriad “blind, mechanical” routines.
5. Inside the defended boundary, there need not always be a Higher Executive or General Headquarters.
6. In nature, handsome is as handsome does; origins don’t matter.
7. In nature, elements often play multiple functions within the economy of a single organism.
One of his points is that the brain, based on design capable of changing with experience, redesigns itself.
For our purposes, let’s just say that one way or another, the plastic brain is capable of reorganizing itself adaptively in response to the particular novelties encountered in the organism’s environment, and the process by which the brain does this is almost certainly a mechanical process strongly analogous to natural selection. This is the first new medium of evolution: postnatal design-fixing in individual brains. The candidates for selection are various brain structures that control or influence behaviors, and the selection is accomplished by one or another mechanical weeding-out process that is itself genetically installed in the nervous system.
Amazingly, this capability, itself a product of genetic evolution by natural selection, not only gives the organisms who have it an edge over their hard-wired cousins who cannot redesign themselves, but also reflects back on the process of genetic evolution and speeds it up.
In section 7.4 Dennett makes the point that genetic evolution of the human brain was essentially complete by 10,000 years ago. All of the development since that time, all human cultural achievements, is a result of the brains of humans redesigning themselves, and passing on some of the redesign culturally, i.e., via memes.
Dennett uses the word represent to indicate something more complex than, say a sunflower representing the position of the sun in the sky, or other “hard-wired” biological structures that permit an organism to react appropriately to its environment. He wants to reserve the word for “post-natal redesign” of a plastic brain. He allows non-human brains to do this, such as a zebra representing where a lion is when it stops watching for a moment (and vice versa). But he lists some of the things an adult human brain can represent:
1. the position of its body and limbs
2. a spot of red light
3. a degree of hunger
4. a degree of thirst
5. the smell of a fine old red burgundy
6. the smell of a fine old red burgundy as the smell of Chambertin 1971
9. the square root of the largest prime number less than 20
10. the concept of a nickel-plated combination corkscrew and staple-remover
It’s pretty certain that no animal can represent items 6 to 10 in that list, and also that a lot of experience is required before an infant brain could do so. He says of representing:
Plasticity makes learning possible, but it is all the better if somewhere out there in the environment there is something to learn that is already the result of a prior design process, so that each of us does not have to reinvent the wheel. Cultural evolution, and transmission of its products, is the second new medium of evolution, and it depends on phenotypic activity in much the same way phenotypic plasticity depends on genetic variation. We human beings have used our plasticity not just to learn, but to learn how to learn better, and then we’ve learned better how to learn better how to learn better, and so forth. We have also learned how to make the fruits of this learning available to novices. We somehow install an already invented and largely “debugged” system of habits in the partly unstructured brain.
Section 7.7 opens with some epigraphs. I found this one particularly apt:
“The large brain, like large government, may not be able to do simple things in a simple way.” – Donald Hebb, 1958.
Also this one:
“The most powerful drive in the ascent of man is his pleasure in his own skill. He loves to do what he does well and, having done it well, he loves to do it better.” – Jacob Bronowski, 1973
In this section Dennett begins to treat the brain as a computer. This thread of his approach seems very weak to me, and not really essential. Still, here’s what he says:
Human consciousness is itself a huge complex of memes (or more exactly, meme-effects in brains) that can best be understood as the operation of a “von Neumanesque” virtual machine implemented in the parallel architecture of a brain that was not designed for any such activities. The powers of this virtual machine vastly enhance the underlying powers of the organic hardware on which it runs, but at the same time many of its most curious features, and especially its limitations, can be explained as the byproducts of the kludges that make possible this curious but effective reuse of an existing organ for novel purposes.
The mention of memes is gratuitous in this passage. The italicized jargon is used deliberately to emphasize that this conception couldn’t have been made until those memes had been created. I agree that consciousness is a serial form of activity, but don’t see the need to introduce the von Neuman architecture, or to find a strained sense in which to say it is implemented on the brain’s parallel architecture. In the end, though, it doesn’t matter much.
Further into the section, he mentions Plato’s metaphor of the birds, where he imagines a man who has caught a lot of wild birds and placed them in his aviary. In the sense that they are contained in his structure, and no one else’s, he possesses them. The hard part is getting a particular bird to come to your hand when you want it. Similarly with learning. We can store many things in our memories, but it requires particular habits of mind to make them come reliably when needed. “Learning to reason is, in effect, learning knowledge-retrieval strategies. That is where habits of mind come in. We have already seen in crude outline how such general habits of mind as talking-to-yourself or diagramming-to-yourself might happen to tease the right morsels of information to the surface (surface of what?). But more specific habits of mind, refinements and elaborations of specific ways of talking to yourself, can improve your chances even further.”
Chapter 8 has the following quote from Helen Keller, that illustrates the connection between consciousness and language (and is very zen-like or taoist):
Before my teacher came to me, I did not know that I am. I lived in a world that was a no-world. I cannot hope to describe adequately that unconscious, yet conscious time of nothingness. . . . Since I had no power of thought, I did not compare one mental state with another. – Helen Keller, 1908.
In section 8.2, Bureaucracy Versus Pandemonium, he says:
One of the skeletons in the closet of contemporary linguistics is that it has lavished attention on hearing but largely ignored speaking, which one might say was roughly half of language, and the most important half at that. Although there are many detailed theories and models of language perception, and of the comprehension of heard utterances (the paths from phonology, through syntax, to semantics and pragmatics), no one – not Noam Chomsky, and not any of his rivals or followers – has had anything very substantial (right or wrong) to say about systems of language production. It is as if all theories of art were theories of art appreciation, with never a word about the artists who created it – as if all art consisted of objets trouvés appreciated by dealers and collectors.
It is not hard to see why this is so. Utterances are readily found objects with which to begin a process. It is really quite clear what the raw material or input to the perception and comprehension systems is: wave forms of certain sorts in the air, or strings of marks on various plane surfaces. And although there is considerable fog obscuring the controversies about just what the end product of the comprehension process is, at least this deep disagreement comes at the end of the process being studied, not the beginning. A race with a clear starting line can at least be rationally begun, even if no one is quite sure where it is going to end. Is the “output” or “product” of speech comprehension a decoding or translation of the input into a new representation – a sentence of Mentalese, perhaps, or a picture-in-the-head – or is it a set of deep structures, or some still unimagined entity? Linguists can decide to postpone an answer to that stumper while they work on the more peripheral parts of the process.
With speech production, on the other hand, since no one has yet worked out any clear and agreed-upon description of what initiates the process that eventually yields a full-fledged utterance, it is hard even to get started on a theory.
This is followed with a few examples of some very preliminary attempts.
Further on, he quotes E. M. Forster: “How do I know what I think until I see what I say?” His point is that consciousness is largely (if not entirely) talking to ourselves and out loud. He has this quote from the Clark’s The Life of Bertrand Russell:
It was late before the two guests left and Russell was alone with Lady Ottoline. They sat talking over the fire until four in the morning. Russell, recording the event a few days later, wrote, “I did not know I loved you till I heard myself telling you so – for one instant I thought ‘Good God, what have I said?’ and then I knew it was the truth.”
Chapter 9 is titled The Architecture of the Human Mind. Dennett says the hard part is over, and starts to consolidate his view. (There are over 200 pages left.) He provides a thumbnail sketch of his theory “so far:”
There is no single, definitive “stream of consciousness,” because there is no central Headquarters, no Cartesian Theater where “it all comes together” for the perusal of a Central Meaner. Instead of such a single stream (however wide), there are multiple channels in which specialist circuits try, in parallel pandemoniums, to do their various things, creating Multiple Drafts as they go. Most of these fragmentary drafts of “narrative” play short-lived roles in the modulation of current activity but some get promoted to further functional roles, in swift succession, by the activity of a virtual machine in the brain. The seriality of this machine (its “von Neumanesque” character) is not a “hard-wired” design feature, but rather the upshot of a succession of coalitions of these specialists.
The basic specialists are part of our animal heritage. They were not developed to perform peculiarly human actions, such as reading and writing, but ducking, predator-avoiding, face-recognizing, grasping, throwing, berry-picking, and other essential tasks. They are often opportunistically enlisted in new roles, for which their native talents more or less suit them. The result is not bedlam only because the trends that are imposed on all this activity are themselves the product of design. Some of this design is innate, and is shared with other animals. But it is augmented, and sometimes even overwhelmed in importance, by microhabits of thought that are developed in the individual, partly idiosyncratic results of self-exploration and partly the redesigned gifts of culture. Thousands of memes, mostly borne by language, but also by wordless “images” and other data structures, take up residence in an individual brain, shaping its tendencies and thereby turning it into a mind.
Section 10.3, Reporting and Expressing, is in Part 4, The Philosophical Problems of Consciousness. As such, it is quite philosophical. Here’s a sample:
What happens when we speak? At the heart of our everyday conception of this there is a truism: Provided we’re not lying or insincere, we say what we think. To put it more elaborately, we express one of our beliefs or thoughts. Suppose, for instance, you see the cat anxiously waiting by the refrigerator and you say, “The cat wants his supper.” This expresses your belief that the cat wants his supper. In expressing your belief, you are reporting what you take to be a fact about the cat. In this case you’re reporting the cat’s desire to be fed. It’s important to note that you’re not reporting your belief, or expressing the cat’s desire. The cat is expressing his desire by standing anxiously by the refrigerator, and you, noticing this, use it as the basis – the evidence – for your report. There are many ways of expressing a mental state (such as a desire), but only one way of reporting one, and that is by uttering a speech act (oral or written or otherwise signaled).
One of the most interesting ways of expressing a mental state is by reporting another mental state. In the example, you report the cat’s desire, thus expressing your own belief about the cat’s desire. Your behavior is evidence not only that the cat has the desire but also that you believe that the cat has the desire. You might, however, have given us evidence of your belief in some other way – perhaps by silently getting up from your chair and preparing supper for the cat. This would have expressed the same belief without reporting anything. Or you might have just sat in your chair and rolled your eyes, unintentionally expressing your exasperation with the cat’s desire just when you had gotten comfortable in your chair. Expressing a mental state, deliberately or not, is just doing something that provides good evidence for, or makes manifest, that state to another observer – a mind reader, if you like. Reporting a mental state, in contrast, is a more sophisticated activity, always intentional, and involving language.
Dennett then goes on about folk psychology and higher-order mental states (such as beliefs about beliefs). The discussion requires reading at least twice (in my experience), but I’m not sure it is crucial to the underlying theory. It is more like a convention adopted by philosophers when talking among themselves; I’ve seen this before in Dennett’s work.
The title of section 13.1, How Human Beings Spin a Self, is an allusion to the non-conscious way that spiders spin a web. In the course of it, he invokes this bit of fundamental data about humans:
Our fundamental tactic of self-protection, self-control, and self-definition is not spinning webs or building dams, but telling stories, and more particularly concocting and controlling the story we tell others – and ourselves – about who we are. And just as spiders don’t have to think, consciously and deliberately, about how to spin their webs, and just as beavers, unlike professional human engineers, do not consciously and deliberately plan the structures they build, we (unlike professional storytellers) do not consciously and deliberately figure out what narratives to tell and how to tell them. Our tales are spun, but for the most part we don’t spin them; they spin us. Our human consciousness, and our narrative selfhood, is their product, not their source.
These strings or streams of narrative issue forth as if from a single source – not just in the obvious physical sense of flowing from one mouth, or one pencil or pen, but in a more subtle sense: their effect on any audience is to encourage them to (try to) posit a unified agent whose words they are, about whom they are: in short, to posit a center of narrative gravity. Physicists appreciate the enormous simplification you get when you posit a center of gravity for an object, a single point relative to which all gravitational forces may be calculated. We heterophenomenologists appreciate the enormous simplification you get when you posit a center of narrative gravity for a narrative-spinning human body. Like the biological self, this psychological or narrative self is yet another abstraction, not a thing in the brain, but still a remarkably robust and almost tangible attractor of properties, the “owner of record” of whatever items and features are lying about unclaimed. Who owns your car? You do. Who owns your clothes? You do. Then who owns your body? You do! When you say “This is my body.” You certainly aren’t taken as saying “This body owns itself.” But what can you be saying, then? If what you say is neither a bizarre and pointless tautology (this body is its own owner, or something like that) nor the claim that you are an immaterial soul or ghost puppeteer who owns and operates this body the way you own and operate your car, what else could you mean?
There’s more, but the point drifts away. Dennett’s “explanation” of consciousness is finished, but not the end of the story. He points out that many people claim they can’t imagine a conscious robot, when in fact they have done so many times. For various reasons, however, many people don’t want to admit the possibility of consciousness outside humans. They would rather deny the power of their own imaginations than admit to even entertaining the idea. Eventually someone will have to implement consciousness to prove the possibility. However, it will take a long time, decades or centuries.
I think Dennett is on the right track, but it will likely have little practical effect in cognitive science. On the other hand, the notion that the mind is a collection of memes, self-designed and self-designing, has potential for other purposes. It should be possible to introduce a large number of people to the notions of memetics, and then to the effects that memes have on them and the people they know. Conscious selection of memes could be a new wave in the new form of evolution that Dawkins, Dennett and others have posited.