Download Understanding Consciousness: A Review of Daniel Dennett's 'Consciousness Explained' and more Study notes Physics in PDF only on Docsity!
DENNETT'S CONSCIOUSNESS
EXPLAINED : ITS CRITICS, AND
THE CONTROVERSY OVER THE
“TRUE NATURE” OF CONSCIOUSNESS
BENJAMIN NEWMAN
Swarthmore College
Are zombies possible? They’re not just possible, they’re actual. We’re all zombies. (Dennett Consciousness Explained, 406)
It would be an act of desperate intellectual dishonesty to quote this assertion out of context! (footnote to the above ).
IN CONSCIOUSNESS E XPLAINED , Daniel Dennett sets forth an extended sketch of what may, or may not, be an explanation of consciousness. The controversy regarding what, exactly, Dennett does explain, and what, if anything, he omits, arises both from profound and genuine disagreement with his views on the part of many contemporary philosophers and from ambiguities in the definitions of some key terms (e.g. “consciousness” and “explanation”). Between the fundamentally conflicting worldviews and behind the conceptual ambiguities, is there a real philosophical controversy to be found, or just a lot of empty rhetoric? I believe that there is a real and meaningful question lurking in the interplay between Dennett and his critics, and it will be my goal in this paper to bring that question to light and to frame it in such a way that it is clear what is at stake. What is at stake? Dennett’s approach to the problem of consciousness is firmly reductive and materialistic. He believes that the physical facts about the brain (and, to a lesser extent, about the context in which it finds itself) are sufficient to account for all of the facts about the mind, consciousness included. Some of Dennett’s critics would have ended that last sentence with the words “consciousness excluded,” and they do have a point. Dennett’s account could be considered an eliminative one, a case of explaining consciousness away. He himself is somewhat equivocal on this point. If we understand “consciousness” to mean some special, ineffable, all-or-nothing property, we will find nothing like that in Dennett’s explanation. If, on the other hand, we take “consciousness” to mean a sophisticated sensitivity to the world around us, a sense of self, and the having of a train of thought, Dennett’s account does include these
6 BENJAMIN NEWMAN
things, and offers compelling answers to the questions of how they might be realized in the brain. So, which is it to be? There is a fundamental disagreement between Dennett and his critics regarding what it is that must be explained. It will not do, Dennett (“Living on the Edge”, 147) reminds us, to appeal to our shared intuition of what we mean when we say that one or another entity is or is not conscious. It may be uncontroversial to interpret such an attribution as meaning simply that the entity in question has whatever it is we have when we are conscious, but it is also obviously circular—Dennett (ibid.) argues that the question of whether or not something is conscious has “no clear pre-theoretical meaning,” and that therefore we ought to develop a theory first, and see afterwards whether or not the theory itself sheds any light on such questions of definition. This is exactly what Dennett has done, drawing his conclusions on what consciousness is from his theory, rather than plugging them in beforehand. It is thus not surprising that his conclusions differ strikingly from the preconceived notions of what consciousness must be on which most philosophers rely. The key difference between Dennett and his critics, as far as can be discerned in this preliminary analysis, seems to be a matter of definitions. It is not that Dennett holds to one particular definition of consciousness while his critics hold to one or several others. Rather, Dennett is willing to postpone the question of what consciousness is until the question of how consciousness works has been settled. It is to this latter question that Consciousness Explained really offers an answer, and it is to this answer that we now turn in greater detail.
Dennett’s Explanation of Consciousness: Discoursing on Methods
Since the key disagreement highlighted by this paper is methodological, it is fitting that Dennett begins his explanation with a clear statement of his methodological choices. His first such choice is an absolute rejection of dualism. How is this decision a question of methodology? Dennett notes that most celebrated dualists do not attempt any explanation of how the nonphysical mind (or nonphysical mental properties of the brain) is supposed to work—they do not seem compelled to give an account of the mind’s essential nature in a way that sheds any light on its function. While dualism does leave open the possibility of detailed empirical study of the mind, few avowed dualists are engaged in such research, and no dualist has a research program on the question of how the mind relates to the body—most declare quite openly that they think the question cannot be answered. While it might turn out that the dualists are right on this score, Dennett is perfectly justified in adopting the working assumption that his project is
8 BENJAMIN NEWMAN
must be explained, and an explanation of them would take the form of a “drafts” theory of the mind. Our sense of a unified stream of consciousness, he claims, is dependent on the time scale at which we view ourselves. At a higher level of magnification, that stream breaks up into many parallel streams of content-fixing events, each carried out by one of the many specialized subsystems that make up the brain. Taken together, these content fixations form something like a narrative, except that this narrative has no canonical form—there are always multiple working drafts, all subject to constant editorial revision, stretching into the unwritten future. The only way a piece of content can become part of any “canonical” narrative is in response to a probe— a direct question to the effect of “What are you thinking of right now?”—and the particular content that emerges in response to such a probe is extremely sensitive to the probe’s nature and timing. What we catch in the “stream of consciousness” depends a great deal on when we go fishing and how we bait the hook. Since the contents of consciousness are, in the sense just articulated, indeterminate if we never bother to check them, the existence of anything like a unified stream of consciousness grows out of the habit of constantly checking, or self-probing—actually several habits, including, but not limited to, such familiar tricks as talking-to-oneself and its silent descendant, private deliberative thought. How do we acquire such habits? Some of them are built in, such as our insatiable curiosity, an innate endowment of most primates, while others are learned, but the greatest contributions, including self-talk, come from language and culture. Not only is language itself one of our most important habits of mind, but it makes possible the learning of such habits by deliberate training and cultural transmission. The ancient Greeks had amazing powers of memory—their storytellers had to be able to recite epics like the Odyssey from memory, something we would find very difficult. Their brains were no different from ours, so why are our memories so different? They used an advanced system of learned mnemonics, which we no longer teach each other because we no longer need them—when we want to remember something as long as a book, we just write it down. Thus, changes in culture can lead to changes in the mental landscape, by affecting both the necessity and the availability of mental habits. At the level of these higher-level, self-probing activities, there finally is one, canonical inner narrative, although it lapses occasionally (whenever we forget to check), and it is meaningless to ask it for fine temporal distinctions shorter than the time scale of the activities which generate it. Thus, a serial, one-thing-at-a-time information process emerges from the activities of a massively parallel information process,
DENNETT'S CONSCIOUSNESS EXPLAINED 9
effectively creating a “virtual machine”—computer science lingo for a computer of one type implemented as a program on a computer of another type. This virtual machine greatly extends the capabilities of the brain by providing the organization required to embark on long-term projects (like philosophy term-papers), but it is not very efficient—that’s why deliberate pondering, the kind of thought that resembles talking-to-oneself, is so slow compared to more hardwired mental activities like deciding to swat a mosquito. Besides being inefficient, the “Joycean machine” is also unreliable—easily distracted and sidetracked, often falling into ruts or repeating itself, and even repeating itself. If these pitfalls are to be avoided, the machine must have some way of regulating its own behavior. One solution is to make use of the already existing perceptual equipment, which has proven so very effective at keeping tabs on complex, unpredictable features of the environment. The system will just have one more complex, unpredictable, vitally important thing to keep track of itself. In such a system, most of the mechanisms will be ad hoc jury-rigs, but most (including many of the jury-rigs) will also be wonderfully sophisticated. Over time, regulatory systems will gradually adapt to the quirks of the systems they regulate, ultimately giving the virtual machine as a whole an effortless familiarity with its own idiosyncrasies. Explicitly keeping track of its own internal states as well as the world around it, organizing and directing the activities of the multiple parallel processes out of which it emerges, and intimately familiar with the quirks of its own activities in a private way that no similar but distinct system could ever be, this virtual machine could account for all of the activities of a human mind. “But is it conscious?” we ask. “YES,” Dennett declares boldly, “my theory is a theory of consciousness. Anyone or anything that has such a virtual machine as its control system is conscious in the fullest sense ... because it has such a virtual machine” (281). This is the essence of Dennett’s explanation, which he further develops by using it to work through some of the thornier philosophical problems that have surrounded consciousness. Although it is on his answers to these problems that Dennett has so often been taken to task by his critics, I will not be organizing my critique of the opposition around the problems themselves, which include zombies, qualia, and the self. This is because the problems overlap, and many of the same (or deeply similar) arguments are used across all of them. For this reason, the next section is organized around the forms of argument used by Dennett’s critics, and it is within this context that we will encounter the problems themselves.
The Other Side: What’s Wrong with Dualism?
DENNETT'S CONSCIOUSNESS EXPLAINED 11
dualist flag will be waved, if not taken up outright, by many of Dennett’s other critics. However, most of them do not aim directly at Dennett’s rejection of dualism, as Foster does. While claiming that there is something inherent in the nature of mind which makes it invulnerable to materialist explanations like Dennett’s, they seem to feel that it is not necessary resorting to argument in establishing these claims about the nature of the mind. Rather, they are content to let the matter rest upon our shared intuitions about the mind, as if all of the important facts were simply obvious, and everyone could simply agree on them. Dennett draws the reader’s attention to this in the introduction to his chapter on methodology (67), calling it the “first-person plural presumption”, and objecting to it on the grounds that there is still a great deal of controversy, even among those who claim that we can all agree on the nature of consciousness as approached through introspection. His answer to this, as mentioned before, is to deny that introspection is ever simply a matter of looking inside and seeing what’s there—all introspection is construction, dependent on theory and subject to categorical as well as particular errors. His critics, of course, will have none of this. John Searle says that it is not difficult to give a common-sense definition of consciousness, one that, not being analytic, will not “aim to analyze the underlying essence of [the] phenomenon,” and to his credit, he does give such a definition: simply put, consciousness is what we have when we are not, in the vernacular sense, unconscious. As Searle admits, this definition does not cut to the heart of the matter, but from it he draws the following conclusions, entirely without argument: consciousness must be all-or-nothing, present or absent. It is an “inner, first-person, qualitative phenomenon.” Searle offers neither evidence nor argument for any of these claimed features of consciousness, nor does he carefully define any of these adjectives. As he freely admits, he has not provided a definition which is pertinent to the analysis of the phenomenon in question, yet he proceeds to render a great deal of analysis from it. On what does he base his analysis? On what grounds does he claim to know what he is talking about when he uses the word “consciousness”? Searle seems to be claiming that we can uncritically pull important philosophical results directly from our shared, commonsense intuitions—in other words, that all of the important features of consciousness are simply obvious, to everyone. Searle’s conviction that the nature of the mind is obvious is demonstrated admirably in his most celebrated thought experiment, the Chinese Room (“Minds, Brains, and Programs”). Searle grants to his opponents that we have a computer program that passes the Turing Test (in Chinese). He then implements this program as a set of lookup
12 BENJAMIN NEWMAN
tables and ledgers, whose operations he will carry out by hand. He refers to the program itself as a “set of instructions” which he is “given”, perhaps suggesting a relatively modest packet of papers with easily followed step-by-step instructions in English. As claimed, the program passes the Turing Test just as surely under Searle’s implementation as before, so we might be tempted to say that it understands Chinese; but Searle protests. He himself doesn’t understand Chinese—he never did—and there is obviously nowhere else in the system as he has described it for the understanding to reside. Lest we claim that somehow it is the whole system which understands, Searle proposes to swallow the system. Simply put, he will memorize the program. It is at this point where Dennett and I both cry foul. Searle has pulled a fast one on his readers, for the program which he asked them to imagine is in all likelihood quite complicated, and now he asks them to imagine that he has got it all memorized. If we really go to all the trouble of imagining a program that passes the Turing Test, implemented on paper, we should probably imagine a large library, not a few small paper packets. Yet, despite this cheat of imagination, Searle claims that it is still obvious that the system does not understand, allowing a number of controversial premises to slip by without argument. While careful to separate such notions as understanding, intentionality, and consciousness, this pattern is typical of his reasoning on all three topics. Another important philosopher who relies on the “obviousness” of his key premises is Thomas Nagel. In “What Is it Like to Be a Bat?” he claims that for an organism to be conscious, there must be “something it is like to be that organism—something it is like for the organism.” He does not offer any argument for this position, nor does he bother to explain what he means by it in any detail. He does, however, go to some lengths to show that no materialist theory can accommodate this feature of consciousness—thus roundly refuting all (or all current) materialist theories, if we all agree on what he claims without argument or evidence, presumably on the basis that it is just obvious. After all, Nagel is quite willing to “assume we all believe that bats have experience” (393) before using them as a case study. Apparently he also assumes we all agree with his definition of what it means to be conscious. John Foster, whom we have already encountered in defense of dualism, makes a similar claim in his response to Dennett’s demand that a dualist give some account of the nature of the nonphysical mind. “I do not myself see,” he writes, “why the dualist needs to admit that there is anything more to the nature of the mind than what introspection can reveal” (1993, p.29). Dennett’s answer has already been noted—that simply positing a mind with all
14 BENJAMIN NEWMAN
would make the additional point in response to Fellows & O’Hear that circularity is to be hoped for in identifying the self as the interpreter of its own narrative, for if a self is a narrative, then it must, on pain of nonexistence, be circularly self-interpreting.
You Can’t Tell by Looking
In this section I will address the related problems of qualia and zombies. It’s often difficult to tell which of these problems an argument addresses, and this is not surprising, since their gist is the same—the traditional philosopher proposes that there are features of consciousness over and above the functional interplay of mental contents; thus, these features might be changed, or removed entirely, without affecting this functional interplay, and thus without affecting an agent’s outward behavior in any way. Dennett and I believe that this notion is incoherent, but many of his critics still cling to it in one or both of its most popular forms—qualia and zombies. Among the features of consciousness which materialists are most often accused of omitting, the raw, qualitative aspects, or qualia, of our states of awareness take pride of place. Almost every one of Dennett’s critics accuses him of failing to do justice to qualia, and, to credit those critics, Dennett says outright that there are no such qualities ( Consciousness Explained 372). By this he means that mental states do not have intrinsic, ineffable properties distinct from the functional roles they play, proposing instead that these functional dispositions are sufficient to account for all of the claimed facts about qualia, including their supposed ineffability. Many philosophers also claim that they can imagine “zombies”, creatures that possess all of the functional attributes Dennett imagines—to the point of being physically or at least outwardly indistinguishable from conscious human beings—but without genuine consciousness. That the two claims are closely related is best indicated by the fact that qualia head the list of what zombies don’t have. To Dennett’s suggestion that we are zombies, and our consciousness as distinct from our functional underpinnings is an illusion, they ask who, if not some conscious self, is deceived by this illusion, echoing the arguments already noted regarding regress. However, my response to the issue of qualia and zombies will be different. It turns on the notion of evidence. What does it mean to have evidence for something? Presumably, by “evidence for P” we mean information that, ceteris paribus, makes the truth of P more likely—thus, information whose availability to us is causally related to the truth of P. Also, when we say that we make some statement based on evidence for P, we mean that our utterance is caused by the availability of this information. Our definition of
DENNETT'S CONSCIOUSNESS EXPLAINED 15
evidence supports the following counterfactual: ceteris paribus, the removal of P causes the removal of evidence for P in at least some circumstances. Otherwise, the relation between the evidence for P and the probability of P does not hold. Thus, if we have made some utterance based on evidence for P, we cannot escape the counterfactual “if not for P, I would not have made that utterance.” To deny the counterfactual is to deny that any of the causes of the utterance can truly be called evidence for P. Now, let the proposition P be “We are really conscious, and not just zombies who act exactly like conscious people in every way.” The counterfactual is clearly false in this case, for by hypothesis a zombie-philosopher will behave just like his conscious counterpart, right down to assertions about being conscious and not just a zombie. Thus, it is not the case that replacing a philosopher with a zombie could ever cause him to express different opinions about his own consciousness. Whatever is the cause of philosophers’ views regarding the existence of consciousness in any sense that would make zombies coherently possible, it cannot meaningfully be called evidence. The same argument can be applied to statements about qualia. The above argument is related to Richard Rorty's “Holism, Intrinsicality, and the Ambition of Transcendence” that makes an even stronger case against the proponents of intrinsic properties, asking whether the notion of intrinsic, non-relational properties is coherent at all. He suggests that not just selves but all objects are, so far as they figure in our discourse, nothing over and above centers of narrative (or at least descriptive) gravity. In other words, there is nothing essential to an object X beyond its being that of which most X-sentences are true. How could we assert otherwise? To do so would simply be to add another X-sentence to the list. If Dennett took this metaphysical thesis (attributed to Wittgenstein) as far as Rorty does, he wouldn’t even have to call his heterophenomenological entities “mere fictions,” for if everything is just a center of descriptive gravity, the whole fact-fiction, real-intentional distinction dissolves entirely. Unfortunately (in Rorty’s view), Dennett does not go this far—he still clings to the possibility of ontology as distinct from epistemology. However, Dennett does go far enough to be able to claim that non-relational properties make no sense. His critics, Rorty says, disagree with him on this deep, metaphilosophical level—they (Rorty singles out Searle and Nagel) are all believers in intrinsic properties. Certainly, Searle and Nagel say as much themselves. However, Rorty also chides Dennett for claiming heterophenomenology as neutral ground in this debate, for there is no neutral ground—to claim that the aspects of a phenomenon that can be talked about are all of the important aspects is to beg the question
DENNETT'S CONSCIOUSNESS EXPLAINED 17
describe non-mental phenomena before attempting to explain them because we describe them and explain them from the same, third-person point of view, and usually in much the same terms. With mental phenomena this is not the case, since the terms which our language gives us for describing consciousness carry with them many presuppositions, embodied both in their definitions and in their usage. That it seems linguistically awkward to speak of an entity as “almost conscious” is one example. To open our investigation of consciousness by describing the phenomenon to be explained in such loaded terms is to beg many of the most important questions, perhaps without even realizing it. For example, in admitting a “commonsense” definition of consciousness, Searle begs the question of whether or not consciousness is an all-or-nothing phenomenon. He does so by adopting a language in which the sentence “This entity is a little bit conscious” sounds wrong because it is ungrammatical. Of course, question-begging is still possible even if such linguistic traps are avoided—Dennett admits to begging the question against dualism, after all—but beginning one’s investigation by defining consciousness in any terms whatsoever, particularly in vernacular terms, is likely to result in at least some accidental question-begging. Most philosophers will agree that this is sloppy at best, and at worst forecloses many options and leads to ill-founded conclusions. If we are going to beg important questions, let us at least do so on purpose. Of course, Dennett cannot help starting with a few metaphilosophical assumptions, and thus begging a few important questions (as Rorty notes), but he does his best to avoid this by not starting with any definition of consciousness. Instead, he focuses on the intersubjectively verifiable phenomena we normally associate with consciousness, developing a coherent theory that can explain these phenomena, and only when the model is complete does he see if it contains any hooks on which we can hang our prior notions about consciousness. Here are the hooks he chooses: a mental content is conscious if and only if it is available to be used in subsequent processing leading to its expression in overt (especially verbal) behavior. Since this criterion is not often well-defined at the time when a particular content is fixed, the fact of any content’s status as conscious or not is indeed contingent on future cognition—there is no fact of the matter about which mental contents are conscious at any given moment. The qualities of a mental content, or its qualia, are exactly the sum total of all reactive dispositions that are inherent in the organism as a result of the presence of that particular content, and the pleasant or aversive nature of these qualities is due to the presence among these reactive dispositions of very ancient, primordial pursuit
18 BENJAMIN NEWMAN
and avoidance reactions. The self is an explanatory posit, a center of narrative gravity created by an organism (in this case, a human one) in the ongoing process of narrative-spinning that is higher cognition. The self is not, however, an illusion, for there are true facts that are most easily stated in terms of selves—for example, the very fact that some particular set of narratives, besides having all issued as speech sounds from the same orifice of the same human body, also cohere logically as autobiographical of a single entity. As Dennett argues throughout Consciousness Explained, these identifications are not at odds with any facts about consciousness that can be supported by argument from intersubjectively uncontroversial premises. In this paper, I have tried to identify and call into question some of the premises underlying the overwhelmingly negative critical response to Consciousness Explained from within the philosophical community. These premises range from the viability of dualism to the existence of intrinsic properties, but by far the most common and pervasive is the simple claim that we already know what consciousness is, that we already possess a description of it, and that the job of an explanation is to use the data of empirical science to justify this preexisting description. It is thus enlightening that so many of Dennett’s critics gleefully despair of ever finding such an explanation. In Dennett’s view, and my own, that a particular description of a phenomenon guarantees that it will never be explained disqualifies that description, unless it can be argued conclusively that all descriptions guarantee this. Given the absence of such an argument (for none of Dennett’s critics consider alternative descriptions of consciousness), Dennett is justified in proposing that an explanation will require a completely new description, but to adopt such a new description means letting go of some of our most cherished intuitions. It is this step that Dennett is willing to take, in defiance of his critics. Some critics claim that it is more noble, more heroic, and truer to human nature to hold out for some special essence that makes us conscious and human. Dennett asks, is it not more heroic to let go of our preconceived notions, and to dare to accept the consequences of our honest inquiry into the nature of consciousness? And I answer, Yes, it is.
REFERENCES
“Symposium: Daniel C. Dennett’s Consciousness Explained.” Inquiry 36 (1993): 3-160. Dahlbom, B. Ed. Dennett and His Critics. Cambridge: Blackwell Publishers,
Dennett, D. Consciousness Explained. Boston: Little, Brown & Co., 1991. -----. “Living on the Edge.” Inquiry 36 (1993).
