What Is Dennett’s Theory a Theory of?


Ned Block




            In Consciousness Explained and  some papers written before and since, Dan Dennett expounds what he says is a theory of consciousness.  But there is a real puzzle as to what the theory is about.   There are a number of distinct phenomena that ‘consciousness’ is used by Dennett and others to denote.  If the theory is about some of them, it is false; if it is about others, it is banal.


            A convenient locus of discussion is  provided by Dennett’s claim that consciousness is a cultural construction.   He theorizes that “human consciousness (1) is too recent an innovation to be hard-wired into the innate machinery, (2) is largely the product of cultural evolution that gets imparted to brains in early training”.[1]   Often, Dennett puts the point in terms of  memes.  Memes are ideas such as the idea of the wheel or the calendar or the alphabet; but not all ideas are memes.  Memes are cultural units, the smallest cultural units that replicate themselves reliably.  In these terms then, Dennett’s claim is that “Human consciousness is itself  a huge complex of memes” (210).   The claim is sometimes qualified (as in the “largely” above).  I think the idea is that consciousness is the software that runs on genetically determined hardware.  The software is the product of cultural evolution, but it would not run without hardware that is the product of biological evolution.


            I claim that consciousness is a mongrel notion, one that picks out a conglomeration of very different sorts of mental properties.  Dennett gives us little clue as to which one or ones, which of the “consciousnesses” is supposed to be a cultural construction.  Now this would be little more than a quibble if his claims about consciousness were plausible and novel proposals about one or more  “consciousnesses”, one or more “elements” of the mongrel. OK, so he doesn’t tell us exactly which consciousness the claim is about, but we can figure it out for ourselves. As far as I can see, there is no kind of consciousness that is both plausibly and non-trivially a cultural construction, a collection of memes.(But perhaps Dennett will prove me wrong in his reply.)  For some kinds of consciousness, the idea that consciousness is a cultural construction is a non-starter.  For others, there is an empirical issue, but the cultural construction claim seems likely to be false, and Dennett does not defend it.  For others, it is utterly banal--certainly not the exciting new thesis Dennett presents it as.   So  my challenge for Dennett will be to provide us with a notion of consciousness on which his claim  is both true and interesting.  Of course, I wouldn’t be bothering with all this if I thought Dennett had an answer.  What I really think is that Dennett is using the mongrel concept of “consciousness” the way Aristotle used the concept of “velocity”, sometimes meaning instantaneous velocity, sometimes meaning average velocity, without seeing the distinction.[2]  I think Dennett has confused himself and others by applying an unanalyzed notion of “consciousness”, conflating theses that are exciting and false with others that are boring and true. I won’t be arguing for this directly, but it is the upshot of what I will have to say.


            My procedure will be to go through the major elements of the mongrel briefly, with an eye to filling in and justifying the claim that what Dennett says is not true or not novel.   I should say at the outset that I do not intend to be presupposing any controversial views about whether the inverted spectrum hypothesis makes sense, whether there can be “absent qualia” (that is, whether there can be creatures functionally identical to us, such that there is nothing it is like to be them) and the like.  What I have to say here is supposed to be independent of such issues.


Phenomenal Consciousness.  Phenomenal consciousness is experience.  Phenomenal conscious properties are the experiential properties of sensations, feelings and perceptions, for example, what it is like to experience pain, what it is like to see, to hear and to smell.   Thoughts, desires and emotions also have phenomenal characters, though these characters do not serve to individuate the thoughts, desires and emotions.  Phenomenal properties are often representational.  For example, what it is like to see something as a refrigerator is different from what it is like to see the same thing from the same angle as a big white thing of unknown purpose and design.  And there is a representational commonality to what it is like to hear a sound as coming from the right and what it is like to see something as coming from the right.  I believe that there is a difference in these experiences that is not representational, a difference that inheres in non-representational features of the modalities; but I will not assume this in what follows.  I also think that phenomenal consciousness is not characterizable in functional or intentional or cognitive terms, but again I will not assume this here.


            There was a time when Dennett was an out and out eliminativist about phenomenal content, but his views have changed.  He now offers a  theory of it, though he cautions us that his views of what phenomenal consciousness is are at variance with a picture of it that has a strong hold on our intuitions. I hope it is just obvious to virtually everyone that the fact that things look, sound and smell more or less the way they do to us is a basic biological feature of people , not a cultural construction that our children have to learn as they grow up.   To be sure, cultural constructions have a big impact on the way things look, sound and smell to us.   As I said, phenomenal consciousness is often representational, and the representational aspects and phenomenal aspects of phenomenal consciousness often interact.  To use Dennett’s wonderful example, suppose we discovered a lost Bach cantata whose first seven notes turn out by an ugly coincidence to be identical to the first seven notes of `Rudolph the Red Nosed Reindeer.’  We wouldn’t be able to hear  the cantata the way the Leipzigers of Bach’s day would have heard it.  So culture certainly has an impact on phenomenal consciousness.  But we have to distinguish between the idea that culture has an impact on phenomenal consciousness and the idea that phenomenal consciousness as a whole is a cultural construction.  Culture has a big impact on feet too.  People who have spent their lives going barefoot in the Himalayas have feet that are different in a variety of ways from people who have worn narrow pointy high-heeled shoes for eight hours a day every day.  Though culture has an impact on feet, feet are not a cultural construction.  So the impact of culture on phenomenal consciousness does not give us a reason to take seriously the hypothesis that phenomenal consciousness was invented in the course of the development of human culture or that children slowly develop the experience of seeing, hearing and eating as they internalize the culture.  Indeed, children acquire the culture by seeing and hearing (and using other senses) and not the other way around. We should not take seriously the question of whether Helen Keller had her first experience of eating or smelling or feeling at the age of seven when she started learning language. We should not take seriously the idea that each of us would have been a zombie if not for specific cultural injections when we were growing up.   We should not take seriously such questions as whether there was a time in human history in which people biologically just like us used their eyes and ears, ate, drank and had sex, but there was nothing it was like for them to do these things.[3] And a view that says that such questions should be taken seriously should be rejected on that basis.


            Though almost everyone believes in phenomenal consciousness, some hold a deflationary or reductionist view of it, identifying it with a functional or intentional or cognitive notion. Mightn’t such views of phenomenal consciousness make the thesis that phenomenal consciousness is a cultural construction more intelligible?    The best way to answer this question, I think, is to examine the other consciousnesses, the other elements of the mongrel.  They are the best candidates for a deflationist or a reductionist to identify with phenomenal consciousness.


Access-consciousness.  Let us say that a state is access-conscious if its content is poised for free use in controlling thought and action.  More specifically, a state  with a certain content is access-conscious if, in virtue of one’s having the state, a representation which has that content is (1) poised to be used freely as a premise in reasoning, according to the capabilities of the reasoner, (2) poised to be used freely for control of action.  In the case of language-using organisms such as ourselves, a major symptom of access-consciousness would be reportability.  But reportability is not necessary. My intent in framing the notion is to make it applicable to lower animals in virtue of their ability to use perceptual contents in guiding their actions.


            In my view, this is the notion of consciousness that functionalists should want to identify with phenomenal consciousness. We needn’t worry about whether access-consciousness is really distinct from phenomenal consciousness, since the question at hand is whether either of them could be a cultural construction.  I am dealing with these questions separately, but I am giving the same answer to both, so if I am wrong about their distinctness it won’t matter to my argument.


            Access-consciousness is a tricky notion which I have spelled out in some detail elsewhere.[4]  I will briefly make two comments about it.  First, the reader may wonder what the “in virtue of” is doing in the definition.  It is there in part because there are syndromes such as blindsight in which the content of a perceptual state is available to the perceiver only when he is prompted and hears himself guess what he is seeing.  In blindsight, there are “blind” areas in the visual field where the person claims not to see stimuli, but the patient’s guesses about certain features of the stimuli are often highly accurate.  But that doesn’t count as access-consciousness because the blindsight patient is not in a position to reason about those contents simply in virtue of having them.  A second issue has to do with the fact that the paradigm phenomenally conscious states are sensations, whereas the paradigm access-conscious states are thoughts, beliefs and desires, states with representational content expressed by “that” clauses.  There are a number of ways of seeing the access-consciousness of sensations such as pain.  Pains are often (some have argued always) representational, and so these representational contents are candidates for what is inferentially promiscuous, etc., when a pain is access-conscious.  Alternatively, we could take the access-conscious content of pain to consist in the content that one has a pain or a state with a certain phenomenal content.[5]


            Now to the point of this excursion into access-consciousness: could access-consciousness be a cultural construction?  Could there have been a time when humans who are biologically the same as us never had the contents of  their perceptions and thoughts poised for free use in reasoning or in rational control of action?  Could there be a human culture in which the people don’t have access-consciousness? Would each of us have failed to be access-conscious but for specific cultural injections?  Did Helen Keller become access-conscious at age seven?  Once asked, the answers are obvious.  Dogs have access-consciousness in virtue of their abilities to use perceptual contents in guiding their actions. Without access-consciousness, why would thought and perception ever have evolved in the first place? The discovery that access-consciousness is anything other than a basic biological feature of people would be breathtakingly amazing, on a par with the discovery that housecats are space aliens.  Anyone who claimed such a thing would have to marshal a kind of evidence that Dennett makes no attempt to provide.  (Of course, to say that access-consciousness is a basic biological feature of people is not to say that it is literally present at birth.  Teeth and pubic hair are biological but not present at birth.)


             Access-consciousness is as close as we get to the official view of consciousness of Consciousness Explained, and also in Dennett’s later writings.  In a recent reply to critics, Dennett  sums up his current formulation of the theory, saying “Consciousness is cerebral celebrity--nothing more and nothing less.  Those contents are conscious that persevere, that monopolize resources long enough to achieve certain typical and “symptomatic” effects--on memory, on the control of behavior, and so forth.”[6]   The official theory of Consciousness Explained is the Multiple Drafts Theory,  the view that there are distinct parallel tracks of representation that vie for access to reasoning, verbalization and behavior.  This seems more a theory of access-consciousness than any of the other elements of the mongrel.  But surely it is nothing other than a biological fact about people--not a cultural construction-- that some brain representations persevere enough to affect memory, control behavior, etc.   Of course, our concept of cerebral celebrity is a cultural construction, but cerebral celebrity itself is not.   No one should confuse a concept with what it is a concept of.  Now we have reached a conundrum of interpretation: the closest thing we have to an official  concept of consciousness in Dennett’s recent work is not a concept of something that can be taken seriously as a cultural construction.  In his reply, I hope Dennett tells us how, according to him, cerebral celebrity could be a cultural construction. In the meantime, I will search for another kind of consciousness that he could have in mind.


            I said that the concept of consciousness is a mongrel concept. Our use of a single word reflects our tendency to see the elements of  the mongrel as wrapped together.  In particular, we think of conscious qualities as given, as completely present with nothing hidden. To see phenomenal consciousness as completely present is to see it as entirely accessible.  These are ideas about consciousness, but they are ideas that affect phenomenal consciousness itself, what it is like to be us, just as in Dennett’s example what it is like to hear the imaginary Bach cantata would be influenced by the idea we have of the Christmas ditty.  Our theories of phenomenal consciousness do influence phenomenal consciousness itself to some extent.  Our experience might be somewhat different in a culture in which a different view of phenomenal consciousness was prevalent.  But we should not allow such interactions to make us lose sight of the main effect.  True, culture modulates cerebral celebrity, but it does not create it.  We must not conflate cultural influence with cultural creation.


            It should be noted that our  theories, even wildly false theories, about many things, not just consciousness itself, can influence our experience.  For example, we sometimes think of seeing as a process in which something emanates from the eyes.  We talk of moving our gaze sometimes as if it were a beam of light.  And we sometimes talk of seeing through a dirty window as if our gaze could to some extent penetrate it.  These notions were parts of theories of vision in ancient times, and even now appear in childrens’ theories.[7] Perhaps these ideas affect our phenomenal consciousness--or perhaps it is the other way around.


Monitoring consciousness  The idea of consciousness as some sort of internal monitoring takes many forms.  One form is one that Dennett discusses in Consciousness Explained: higher order thought.  In Rosenthal’s version[8], to say that a state is conscious in this sense is to say that the state is accompanied by a thought to the effect that one is in that state.  Perhaps in another time  or culture people were or are much less introspective than they are  here and now, but would anyone claim that there was a time or place when people genetically like us (and who are not shell-shocked or starving to death) had children who had no capacity to think or say  something on the order of “Carry me, my leg hurts”?  To be able to think or say this involves being able to think that one’s leg hurts, and that is to think a higher order thought in the relevant sense.   I won’t say that it isn’t possible that some wise child of our species discovered  that she could get Mom to carry her by talking  (and thinking) about her pain, but it would take weird and wonderful discoveries to convince me that this is a theoretical option to be taken seriously.  Dennett does not give any hint of the kind of weird and wonderful discoveries that would be needed.  So we have to doubt that this is what he means.  (Though it should be noted that Dennett makes a number of very favorable remarks about this idea of consciousness in Consciousness Explained.)


Self-consciousness  There are a number of closely-connected notions of self-consciousness clustered around the notion of the ability to think about oneself.   Let us begin with a minimal notion of self-consciousness, one that requires thinking about oneself, but not in any particular way.   Certainly this very minimal kind of self-consciousness is unlikely to be a cultural construction.  Consider deception.   Deception involves thinking about getting others to think that one believes something other than what one actually believes.  This involves the minimal self-consciousness just mentioned.   There is pretty good evidence that higher primates practice deception, so it seems unlikely that humans had to invent it.[9]  Further, some higher primates, notably chimps, show other signs of self-consciousness in the minimal sense. Some primates show signs of exploring their bodies in mirrors while other primates, and humans below age one and a half do not.  Gallup[10]  and others have painted bright spots on the foreheads and ears of anesthetized primates, watching what happened.  Chimps between the ages of 7 and 15 usually show surprise on looking at their mirror images, then touch the spot, attempting to wipe off the mark.  This is known as the mark test.  Non-primates  never do this. Human babies don’t pass the mark test until the middle of their second year.  This is now a well-established phenomenon replicated numerous times, though there are raging controversies.[11]   As far as I can see, the controversies have little to do with chimp self-consciousness in the minimal sense.  Some of the controversy is about whether chimps have a “theory of mind”, but that is not required for minimal self-consciousness.  Anyway, there is independent evidence from the literature on genetic defects that humans have an innate module dedicated to understanding the minds of other humans, a module which is therefor most unlikely to be a cultural construction.  Autistic people appear to lack that module, even when they are otherwise cognitively normal, and there is another syndrome, a chromosomal abnormality, Williams Syndrome, in which the patients have the mind-module even when they are terribly sub-normal in other cognitive respects.[12]  Carey, et.al. mention a story that illustrates the lack of theoretical understanding characteristic of Williams Syndrome.  A young adult woman with Williams Syndrome had read a number of vampire novels.  When asked why vampires bite necks, she was very puzzled, and eventually answered that they must have “an inordinate fondness for necks”.  She had no idea that vampires are supposed to consume blood.  This sort of evidence for genetic mental modules certainly puts a heavy burden of proof on anyone who claims that anything so basic as consciousness in any of the senses discussed so far is a cultural construction.


             Another controversy is over whether chimps really realize that they are seeing themselves in the mirror.  Perhaps a chimp who is a subject of the experiment thinks she is seeing another chimp with a dot on his forehead and that makes the subject chimp wonder whether she has a dot on her forehead too.  Maybe so, but the ability to wonder whether I have a dot on my forehead presupposes that I have minimal self-consciousness.  Another objection to these experiments is that perhaps they mainly test understanding of mirrors. But plausibly understanding mirrors involves having some idea that it is oneself that one is seeing in the mirror.


            The most fascinating result I’ve heard of in this area in recent years is unpublished work by Marc Hauser on the cotton-top tamarin, a small monkey which has a large white tuft on the top of its head.  Monkeys had never been observed to pass orthodox versions of the mark test.  Hauser thought that perhaps the inconsistent responses shown by chimps and other higher primates had to do with the lack of salience of the dots, so he died the cotton-top tuft outrageous electric colors, flamingo pink, chartreuse, etc..  The finding is that the cottontops passed the mark test.  Normally, they don’t look in the mirror much, and rarely longer than 1-3 seconds at a time.  Hauser observed long stares on the part of the monkeys with died tufts of 30-45 seconds, and a three-fold increase in touching their tufts.  Further, it seems unlikely that the monkeys thought they were looking at other monkeys, since staring in this species is a threat, and these monkeys were staring peacefully, something they do not normally do.  Hauser has run all sorts of controls, e.g. painting the mirror instead of the monkey,  checking what happens when a monkey sees another monkey with a died tuft, checking the effect of the smell and the feel of the die, and the result stands (Hauser, personal communication).


            Another experiment (mentioned by Dennett, p. 428) is that a chimp can learn to get bananas via a hole in its cage by watching its arm on a closed circuit TV whose camera is some distance away.[13]  Though there is strong evidence that chimps (and maybe monkeys) are self-conscious in the minimal sense, given the controversies in the field, I will draw a weaker conclusion, namely that it is up to anyone who claims that humans are not, as a biological matter, self-conscious in the minimal sense to debunk this evidence. In the absence of such debunking, we are entitled to suppose that it is false that self-consciousness in the minimal sense is a cultural construction in humans.  (I will ignore the possibility that self-consciousness is an independent creation of  monkey, chimp and human culture.)  The idea that minimal self-consciousness is a cultural construction  is certainly more of a genuine possibility than the options canvassed earlier, but  it is nonetheless a poor empirical bet.


            But haven’t I given up my case for confusion by admitting that this is an empirical question which could come out either way?  No, Dennett will not get off the hook so easily.  To be sure, he sees his theory of consciousness as empirical, but empirical in a diffuse way, not something that could be refuted by experiments on cottontop tamarins.  Dennett is well aware of the work on the mark test and chooses not to mention it, at least not in anything I’ve seen.  He clearly does not see his theory of consciousness as depending on these specific empirical results, since he mentions few of them; indeed, to the extent that he does mention this type of work, he appears to have something like the same view that I have expressed.  He seems to discuss the Menzel experiment so as to support the idea that chimps are self-conscious in something like the way that we are.  He describes the result as “a decidedly non-trivial bit of self-recognition.” (428)


            Where are we?  I have argued that if Dennett’s theory is about phenomenal consciousness or access-consciousness, it is obviously false and if it is about minimal self-consciousness, it is less obviously false but still false.


            There is a notion of self-consciousness that  is a better candidate for what Dennett has in mind than minimal self-consciousness.  He does have a chapter on the self in which he paints the self as a fiction, a fiction invented in human history.  Now I should say right off that I have long been sympathetic  to something like this idea, though I prefer a more conservative version, viz., that the self is much more fragmented than we like to think.  This is an idea that has been around for many years, one that grows ever more prominent as the evidence mounts up.  The first really impressive case for it by a philosopher was Thomas Nagel’s famous paper on split brains and the unity of consciousness.[14]  Nagel argued that the fragmentation observed in split brain patients exists to some degree in normal people, and in the light of it our concept of the self crumbles.  This sort of idea has been widened and elaborated for many years now by many psychologists and neuropsychologists, notably by Gazzaniga and his colleagues[15]  Gazzaniga tries to explain many ubiquitous cognitive phenomena in terms of the relations among “sub-selves”, especially the efforts of some “sub-selves” to rationalize the behavior of other “sub-selves”. [16]


            Now here is the relevance to Dennett.  I have been talking about a rather minimal notion of self-consciousness, one that it seems that chimps and human toddlers and maybe monkeys have.  This is a very unintellectual notion of self-consciousness, because it is very relaxed about the notion of the self.  In particular, this weak notion of self-consciousness does not require any conception of the self as being or as not being a federation of sub-selves.  But we are free to frame a more intellectual notion of  the self that does presuppose that we are not such a federation.   I think it is the self in this sense, the nonfederal sense, that Dennett thinks is a fiction.  And that sense of the self gives rise to a conception of self-consciousness, namely thinking about oneself in some way that is incompatible with being a federation.  Let us call this sense of self-consciousness NONFEDERAL-SELF-consciousness.  I spell it this way to remind the reader that the emendation attaches to the concept of the self: and only derivatively to the concept of consciousness--it is self consciousness that involves thinking of the self in a certain sophisticated way, a way that Dennett thinks (and I agree) is probably wrong.  Note that unity and nonfederation are distinct, since unity is compatible with both federation and nonfederation.  The U.S.A. is a unity as well as a federation.  One could think of oneself as a unity without having the conceptual equipment to think of oneself as either a federation or not a federation.  I would guess (and its just a guess) that members of our species have always thought of themselves as a unity, but only in recorded history have thought of themselves as NONFEDERAL, or, in the case of Dennett, Nagel, et. al. (including myself) as FEDERAL.


            So perhaps what Dennett means when he says that consciousness is a cultural construction is that NONFEDERAL-SELF-consciousness is a cultural construction because the nonfederal self is a cultural construction.  But that would make the claim a total banality.  It is no surprise at all that the ability to think of oneself in a very sophisticated way is a product of culture.  You can’t think of yourself as falling under a sophisticated concept without  having the sophisticated concept. We could call thinking of oneself as chairman CHAIRMAN-SELF-consciousness.  CHAIRMAN-SELF-consciousness involves thinking of oneself as the person who guides the Department, the person who has the keys, etc.   And all could agree that CHAIRMAN-SELF-consciousness is a cultural construction because the concept of a chairman is a cultural construction.  But that is no news.  Don’t get me wrong.  I’m not saying that the idea that selves are not federations is a banality.  On the contrary, I think it is an interestingly false thesis.  What I am talking about is the claim that it requires culture to think of one’s self as nonfederal (or as federal)--that’s the banality.  To put it slightly differently, Dennett’s claim that we are federations, that we have federal selves, is a very interesting and profound idea that I agree with.  What is banal is that it takes culture to think of oneself using such an interesting and intellectual concept (or its negation).  The thesis about the self is interesting, the thesis about self-consciousness is banal.


            However, I find nothing in the texts to justify the idea that what Dennett means is that NONFEDERAL-SELF-consciousness is a cultural construction.[17]   So I can’t claim to have figured out what Dennett means yet.  But there is an important piece of the puzzle that  I haven’t  introduced yet.


            Before I get to that piece, I want to guard against one source of misunderstanding.  I mentioned earlier that we think of phenomenal consciousness as wrapped together with access-consciousness  in thinking of phenomenal consciousness as accessible.  And this way of thinking--which may be a cultural product--influences phenomenal consciousness itself.  We think of phenomenal consciousness as having nothing hidden about it, and experience might perhaps be somewhat different in a different culture in which phenomenal consciousness was not thought about in this way.  I cautioned against the mistake of jumping to the conclusion that if this is right it shows that consciousness is a cultural construction.  To confuse being influenced by culture with being created by culture would be a serious error, one that I am not attributing to Dennett.  Consciousness and feet may both be influenced by culture, I concluded, but neither is created by it.  What I am leading up to is that a similar point can be made about the relation between phenomenal consciousness and self-consciousness.  There is a “me”-ness to phenomenal consciousness that may come in part from culture; this aspect comes out in part in the way we describe phenomenal consciousness as “before the mind”.   Whether or not the “me”-ness of phenomenal consciousness is in part cultural, the ideology of unity of the self mentioned earlier gives us reason to think there is an influence of culture on the way many or most of us experience the world.[18]  But once again, though there may be a cultural influence on phenomenal consciousness here, this is no reason to postulate that phenomenal consciousness is a cultural creation.  Surely, in any culture that allows the material and psychological necessities of life, people genetically like us will have experiences much like ours: there will be something it is like for them to see and hear and smell things that is much like what it is like for us to do these things.


            I mentioned a new piece of the puzzle.  Here it is: In the early part of Consciousness Explained (p. 24), Dennett tells us that consciousness is like love and money.  He thinks that you can’t love without having the concept of love, and (more plausibly), that there wouldn’t be any money unless some people had the concept of money.   (In another work, soon to be mentioned and quoted from, he includes right and wrong in the list of things that don’t exist without their concepts.)   According to Dennett, you can’t have consciousness unless you have the concept of consciousness.  This is certainly a wild-sounding view (and he concedes this).  Its incompatibility with common ideas is exemplified by the fact that we are inclined to think that animals are conscious (in at least the phenomenal, access and minimal-self senses) but don’t have the concept of consciousness.   I don’t know what Dennett’s argument for this claim is or what kind of consciousness he has in mind, but it does seem closely connected with the idea that consciousness is a cultural construction.   Here is a line of reasoning that connects them.  Suppose that Dennett is right that we can’t be conscious without having the concept of consciousness.  And suppose further that the concept is a cultural construction.  Then consciousness itself requires a cultural construction and could for that reason be said to be a cultural construction.  Since there is a close connection between the claim that consciousness requires its own concept and the claim that consciousness is a cultural construction, we should consider what kind of consciousness it is supposed to be that you can’t have without having a concept of it.  For whatever kind of consciousness it is that requires its own concept will no doubt be the Holy Grail, the kind of consciousness we have been seeking that is a cultural construction (and interestingly so).


            Dennett credits Julian Jaynes as one of the sources of the idea that consciousness is a cultural construction (p. 259).  Now we are in luck because Dennett has written a long review of Jaynes’ book, Origins of Consciousness in the Breakdown of the Bicameral Mind  which links this view to the idea that consciousness requires its own concept, a view which  Dennett also credits to Jaynes[19].  What kind of consciousness is it that Jaynes is supposed to think requires its own concept?  Dennett criticizes my review of Jaynes[20] for misconstruing a revolutionary proposal as a simple blunder..


In a review of Jaynes’s book some years ago, Ned Block (1981) said the whole book made one great crashing mistake, what we sometimes call a “use mention” error: confusing a phenomenon with either the name of the phenomenon or the concept of the phenomenon.  Block claimed that even if everything that Jaynes said about historical events were correct, all he would have shown was not that consciousness arrived in 1400 B. C., but that the concept of consciousness arrived in 1400 B.C.  People were conscious long before they had the concept of consciousness, Block declared, in the same way that there was gravity long before Newton ever hit upon the concept of gravity [A discussion of morality follows]...Right and wrong, however, are parts of morality, a peculiar phenomenon that can’t predate a certain set of concepts, including the concepts of right and wrong.  The phenomenon is created in part by the arrival on the scene of a certain set of concepts... Now I take Jaynes to be making a similarly exciting and striking move with regard to consciousness.   To put it really somewhat paradoxically, you can’t have consciousness until you have the concept of consciousness. (p 152) [Note: though Dennett calls this a paradoxical way of putting it, he says this repeatedly and does not put it any other way.]



            Jaynes has a very concrete version of Dennett’s hypothesis that consciousness is a cultural construction, namely that it was invented in Europe by the ancient Greeks around 1400 B.C.  We don’t need to get into the issue of what Jaynes actually meant by `consciousness’. For my purposes, the issue is what Dennett takes Jaynes to mean, because Dennett himself endorses the idea that consciousness is a cultural construction in this sense.  Here is what he says.



Perhaps this is an autobiographical confession: I am rather fond of his way of using these terms; [`consciousness’, `mind’ and other mental terms] I rather like his way of carving up consciousness.  It is in fact very similar to the way that I independently decided to carve up consciousness some years ago.


So what then is the project?  The project is, in one sense, very simple and very familiar.  It is bridging what he calls the “awesome chasm” between mere inert matter and the inwardness, as he puts it, of a conscious being.  Consider the awesome chasm between a brick and a bricklayer.  There isn’t, in Thomas Nagel’s (1974) famous phrase, anything that it is like to be a brick.  But there is something that it is like to be a bricklayer, and we want to know what the conditions were under which there happened to come to be entities that it was like something to be in this rather special sense.  That is the story, the developmental, evolutionary, historical story, that Jaynes sets out to tell. (149)


            So it looks like the kind of consciousness that requires its own concept and  is a cultural construction is after all  phenomenal consciousness.  W. V. Quine tells me that he asked Jaynes what it was like to be a person before consciousness was invented.  Jaynes replied, Quine says, that what it was like to be them was no different from what it is like to be a table or a chair.  The passage just quoted suggests that Dennett would agree.


            So we are back to square one.  I’ve been going through concepts of consciousness one by one looking for a concept of consciousness in which Dennett’s thesis escapes both falsity and banality, and phenomenal consciousness is the first concept of consciousness I tried.  If phenomenal consciousness is not reducible to one of the other consciousnesses, then the claim that phenomenal consciousness requires its own concept and is a cultural construction is obviously false for reasons I gave.  But if Dennett does favor one of these reductions, we have every right to ask: “Which one?”  And if the answer is one of the consciousnesses I have covered, the claim is false or banal.


            Perhaps Dennett will say that he will have no part of my distinctions, that they impose a grid on the phenomena that doesn’t sit well with his way of thinking of things.  But this is no defense.  Consider randomness. The concept can be and is used in two very different ways.  Sometimes we say a particular sequence is random if it is produced by a random process, even if the sequence itself consists of eighteen consecutive sevens.  Other times what we mean is that it is of a type that one would expect to be produced by a random process, that is, it has no obvious pattern. Suppose someone makes a claim that is false on one concept of randomness and banal on the other.  It would be of no use at all for the offender  to defend himself by saying that he didn’t find the distinction congenial.  Given the fact that on one way of cutting things up, his thesis is trivial or banal, it is up to him to give some precise way of thinking about randomness that disarms the objection.  He must show how his thesis can be neither false nor banal, and to do this he will have to make his notion of randomness precise in a way that allows us to see that the criticism is wrong.


            The application of the analogy to Dennett is straightforward.  I have argued that on one grid that we can impose on the phenomena, his claim is either false or banal.  He does not have the option of simply saying he doesn’t like the distinctions.  He will have to find a way of making more precise what he is talking about under the heading of “consciousness” in a way that rebuts the charge of falsity or banality.  It is no good just refusing to make distinctions at all, since anyone can see that ‘conscious’ is highly ambiguous, and my argument puts the burden of proof on him.


            In another publication written about the same time as this paper, I have made a shorter version of some of these points about Dennett and Dennett has replied.[21]  Here is what I see as his main point:


Although Block discusses my theory of consciousness at some length, his discussion always leans on the presupposition that his putative distinction is in place. My theory of consciousness is stranded, he concludes, between being trivially false (if a theory of P-consciousness), non-trivially false (if a theory of “just” A-consciousness) and banal if a theory of “a highly sophisticated version of self-consciousness.” But since I not only decline to draw any such distinction, but argue at length against any such distinction, Block’s critique is simply question-begging. I may be wrong to deny the distinction, but this could not be shown by proclaiming the distinction, ignoring the grounds I have given for denying it, and then showing what a hash can then be made of ideas I have expressed in other terms, with other presuppositions. If Block thinks his distinction is too obvious to need further defense, he has missed the whole point of my radical alternative. This is a fundamental weakness in the strategy Block employs, and it vitiates his discoveries of “fallacies” in the thinking of other theorists as well. Those of us who are not impressed by his candidate distinction are free to run the implication in the other direction: since our reasoning is not fallacious after all, his distinction must be bogus.

            First of all, though Dennett has some complaints against the phenomenal consciousness/access-consciousness distinction, he never mentions any problem about the notions of access-consciousness, monitoring consciousness or self-consciousness, nor  does he impugn the distinctions among these things.  Oversimplifying (see below), Dennett wishes to treat phenomenal consciousness as a type of access-consciousness.  But the argument I gave can run on just monitoring consciousness, self-consciousness and access consciousness of various sorts.   Supposing that phenomenal consciousness just is a type of access consciousness, what then is Dennett’s theory about?  If it is about access consciousness, Dennett will run into the problem mentioned earlier that it is obviously a biological fact about people and not a cultural construction that some brain representations persevere enough to affect memory, control behavior and the like. Since this is Dennett’s favored way of describing access, it is not easy to understand how seeing phenomenal consciousness as a type of access consciousness  is supposed to avoid the problem.  If there is some novel form of access that his theory is about, it is surprising that he has not told us in any of his many publications on this topic, including his reply to a version of the criticism of this paper.


            Secondly, Dennett does not reject the phenomenal consciousness/access consciousness distinction.  Far from it--he reconstructs it.  His idea is that phenomenally conscious contents are richer in information and more accessible than the level required for access-consciousness.  Thus, he says, I am “inflating differences in degree into imaginary differences in kind.”  I  believe I can show that this reconstruction will not do--see my reply in BBS. For present purposes, lets suppose Dennett is right: the difference is one of degree.  Which degree, then, does his thesis apply to?  Or does it apply to monitoring or self-consciousness?  My criticism does not depend on taking the distinction to be one of kind rather than of degree.


            Thirdly, Dennett contrasts the informational paucity of the perceptual contents of the blindsight patient with the informational richness of normal vision.  Some classic blindsight studies involve prompting the blindsight patient to guess as between an ‘X’ and an ‘O’ or as between a horizontal and a vertical line.  Normal perceptual contents are much richer, representing colors and shapes that are a small subset of a vast number of possibilities. In normal vision, we can “come to know, swiftly and effortlessly, that there was a bright orange Times Roman italic ‘X’ about two inches high, on a blue-green background, with a pale gray smudge on the upper right arm, almost touching the intersection? (That’s a sample of the sort of richness of content normally to be gleaned from the sighted field, after all.)”  Supposing that Dennett is right that phenomenal consciousness contents are just contents that are particularly rich in information and accessibility, is it phenomenally conscious contents that are cultural constructions and require their own concepts?  It is hard to take seriously the idea that the human capacity to see and access rich displays of colors and shapes is a cultural construction that requires its own concept.  Indeed, there is a great deal of evidence that culture does not even influence these perceptual contents.  For example, in cultures which have only two or three color words, the people make all the same perceptual distinctions that we do.  Further, they recognize the same colors as focal that we do even if their languages do not separate out those colors.[22]   In a fascinating series of studies, Eleanor Rosch showed that the Dani,  a New Guinea tribe that has only two color words, nonetheless remember and represent colors in many respects just as we do.  For example, they learned words for focal colors much more easily than words for non-focal colors (e.g. blue as opposed to greenish blue).  When asked to learn words for oddball color categories covering focal colors plus adjacent non-focal colors, some subjects wanted to quit the study.[23]  There is reason to think that many aspects of color and shape perception are genetically coded features of the visual system, and not a product of culture or something that requires any concept of consciousness.[24]


            So I leave the reader with a quandary, one that  I hope Dennett will now resolve, since he gets the last word.  Consciousness is a mongrel notion: there are a number of very different  concepts of consciousness.   On some of these, notably phenomenal consciousness, access-consciousness and monitoring consciousness, the idea that consciousness is a cultural construction is hard to take seriously.  If it is minimal self-consciousness that is meant, it is an empirical issue where available evidence goes against the cultural construction idea.  (But if that was what Dennett meant, you would think he would  have commented negatively on that evidence; instead his limited comment is positive.)  If it is  a sophisticated self-consciousness that is meant (NONFEDERAL-SELF-consciousness), then the thesis is true but utterly banal, because it is no surprise that the ability to apply a sophisticated concept to oneself requires a cultural construction.  I don’t claim to have covered all the options.  But I have covered enough options to make it fair to ask for an answer:  What kind of consciousness is it that requires its own concept and is a cultural construction?[25]

[1]Consciousness Explained,  Little Brown, 1991, p. 219. References that are not attributed  to another work are to this book.

[2]T.S. Kuhn, “A Function for Thought Experiments” in Melanges Alexandre Koyre Vol 1.  Hermann 307-334, 1964

[3]This last point could be rebutted by the claim that througout human evolution there was a culture that created phenomenal consciousness (apparently contrary to Julian Jaynes’ view to be discussed later).  If we allow ourselves to take the view that phenomenal consciousness is a cultural construction seriously we will have to take this issue seriously.  My point, however, is that we should not take this question seriously.  It is a poor question that will just mislead us.

[4]”On a Confusion about a Function of Consciousness”, The Behavioral and Brain Sciences 18, 2, 1995, 227-247See also my reply to my critics in the same volume.   This paper is reprinted in The Nature of Consciousness: Philosophical Debates, N.Block, O. Flanagan, G. Guzeldere (eds), MIT Press, 1995.

[5]One problem with the first of these suggestions is that perhaps the representational content of pain is non-conceptualized, and if so, it would be too primitive to play a role in inference.  After all, dogs can have pains, and it is reasonable to doubt that dogs have the relevant concepts.  In response to an earlier version of this distinction, Davies and Humphreys have made a suggestion which I can adapt.  (See the introduction to their Consciousness, Blackwell, 1993.)  A state with non-conceptualized content is access-conscious if, in virtue of one’s having the state, its content would be inferentially promiscuous and poised for rational control of action and speech if the subject were to have had the concepts required for that content to be a conceptualized content.

[6]”The Message is: There is No Medium” Philosophy and Phenomenological Research vol. LIII, 4, December 1993, p. 929.

[7]These theories are known as extramission theories.

[8]”Two Concepts of Consciousness”.  Philosophical Studies 49: 329-359

[9] Merlin Donald’s Origins of the Modern Mind: Three Stages in the Evolution of Culture and Cognition, Harvard, 1991 is often thought to be rather  critical about the evidence for the conceptual capacities of chimps compared to humans.  It is interesting in this regard to find Donald replying to six critics who criticize  him for this by admitting that there is impressive evidence for ape deception.  See his reply to critics in Behavioral and Brain Sciences 16:4, p. 777, 1993.   Merlin says he is especially impressed with data on chimps’ capacities, including some that indicate “sense of `self’” in Alexander Marshak’s  “Correct Data Base: Wrong Model?” p. 767 in the same journal issue.  Mitchell and Miles (in the same issue) provide further data supporting this conclusion.

[10]Gallup, G.  “Self-awareness and the emergence of mind in primates”  American Journal of Primatology 1982, 2: 237-248.  The most complete and up-to-date survey on the mark test as of my writing this is D. Povinelli’s “What Chimpanzees Know about the Mind”, in Behavioral Diversity in Chimpanzees, Harvard University Press, 1994.  See also R. W. Mitchell, “Mental models of mirror self-recognition: Two theories” in New Ideas in Psychology 11, 1993; 295-332; R. W. Mitchell, “The Evolution of Primate Cognition: Simulation, Self-Knowledge and Knowledge of Other Minds”.  In Hominid Culture in Primate Perspective, D. Quiatt , J. Itani (eds), University Press of Colorado, 1993.

[11]Self-Awareness in Animals and Humans, S. T. Parker, et. Al., eds, Cambridge University Press, 1994?.  See also Voll 11, No 3 of New Ideas in Psychology, 1993.  R. W. Mitchell’s paper (“Mental Models of Mirror-Self-Recognition: Two Theories” draws fire from Gallup and Povinelli, De Lannoy, Anderson and Byrne”, and there is a reply by Mitchell.  I think one gets a pretty good idea of what the controversies are like from this exchange.

[12]Carey, S., Johnson, S. & Levine, K. “Two Separable Knowledge Acquisition Systems: Evidence from Williams Syndrome.” Tager-Flusberg, H., Sullivan, K. & Zaitchik, D. “Social Cognitive Abilities in Young Children with Williams Syndrome”.  These papers were presented at the Sixth International Professional Conference of the Williams Syndrome Association., July 1994.

[13]Menzel, E. W., Savage-Rumbaugh, E.S. and Lawson, J.  “Chimpanzee (Pan troglodytes) Spatial Problem Solving with the Use of Mirrors and Televised Equivalents of Mirrors.”  Journal of Comparative Psychology 99, 211-217: 1985.

[14]”Brain Bisection and the Unity of Consciousness,” Synthese 1971

[15]M. Gazzaniga and J. E. LeDoux, The Integrated Mind, Plenum: New York, 1978; M. Gazzaniga, The Social Brain, Basic Books: New York: 1985.  See also Marvin Minsky’s Society of Minds. Simon and Schuster, 1985.

[16]For a detailed account of the difference between phenomenal consciousness and self-consciousness, and of why it is self-consciousness that matters to morality, see Stephen White, “What is it Like to be a Homunculus” Pacific Philosophical Quarterly 68: 148-174, 1987.

[17]Further, if that was what Dennett meant, wouldn’t he have advanced his theory of the self as a fiction in the course of presenting the theory of consciousness?  Instead, the theory of consciousness (including consciousness as a cultural construction) is presented in Part II of the book (mainly chapters 7-9), and the theory of the self  is given in Part III at the end, in the next to the last chapter of the whole book, Chapter 13.

[18]I am indebted to an unpublished paper on the self by Stephen White.

[19]”Julian Jaynes’ Software Archeology”, Canadian Psychology 1986, 27: 2, p. 149-154.

[20]Cognition and Brain Theory,IV, 1, 1981, pp. 81-83.

[21]My paper is the one in The Behavioral and Brain Sciences 18, 2, June 1995 mentioned earlier.  Dennett’s reply is “The Path Not Taken” in the same issue, p 252--253.  ‘P-consciousness’ and ‘A-consciousness’ are the terms used in that paper for phenomenal and access consciousness.

[22] B. Berlin and P. Kay, Basic Color Terms: Their Universality and Evolution, University of California Press, 1969

[23]”On the internal structure of perceptual and semantic categories” in T. E. Moore, ed. Cognitive Development and the Acquisition of Language, Academic Press, 1973, p. 111-144.

[24]I can’t resist commenting on Dennett’s suggestion that phenomenal consciousness  can be characterized in part in terms of informational richness.  (I won’t comment on the accessibility part of the theory.)


                Weiskrantz (1988) notes that his patient DB had better acuity in some areas of the blind field (in some circumstances) than in his sighted field.  And there is an obvious way to increase the superiority of the blind field--namely by decreasing the richness of the deliverances of the sighted field.  Suppose a Mad Scientist kidnaps a blindsight patient and damages the sighted part of the visual system.  Many blind people are unable to do much more than distinguish light and dark, so we can imagine the Mad Scientist injuring a blindsight patient by so damaging his sighted field.  In the sighted field, he experiences the difference between light and dark.


                But don’t we still have an informational superiority of the blind field?  Dennett describes the informational content of blindsight as “vanishingly” small.  In his book, he emphasizes the cases in which the blindsight patient is given a forced choice, e.g. an ‘X; or an ‘O’.  But blindsight patients can exhibit contents that are far more informational than that.  In Pöppel, et.al.’s famous paper,  the first human blindsight study, the patients were asked to move their eyes in the direction of the stimulus, which they could do.  (Pöppel, E., Held, R., Frost, D. “Residual visual functions after brain wounds involving the central visual pathways in man”.  Nature 243, 2295-2296, 1973.)  So we could have a blindsight patient whose blind field discriminations involved distinguishing among  a number of  different directions, and who could not make that many discriminations in his sighted field. In the light of this point, no one should maintain that high informational content is the essence of or necessary for  experience.  Further, blindsight patients can catch a ball thrown in the blind field, and shape their hand to grasp an object presented in the blind field.  These are cases of far more than binary information, and more, I would guess, than some cases of near total blindness of the sort described.  Further, there are other blindsight-like phenomena in which subjects have rich informational contents without phenomenal or access consciousness of it.  Prosopagnosia is a neurological impairment in which subjects cannot recognize faces, even the faces of their friends and family.  Bauer (“Autonomic recognition: a neuropsychological application of the guilty knowledge test.”  Neuropsychologica 22: 457-469, 1984) showed patients photographs of people they had seen many times, e.g. John Wayne, and went through a number of names, noting a polygraph blip when the right name came up.  Other experiments have shown that many prosopagnosics have information about the faces that they cannot consciously recognize in either  the  phenomenal or access senses.  What’s the informational value of seeing that it is John Wayne?  Not vanishingly small.  Compare the richness of this content with that of say a salty taste (while holding your nose so there is no smell).  It is not at all clear that the experience of tasting without smelling has more informational value than the prosopagnosics non-experiential appreciation that he is seeing John Wayne.


[25]I am grateful to Susan Carey, Chris Hill,  Paul Horwich and Stephen White for comments on a previous draft.