Topics in Cognitive Science 1 (2009) 231–236

Copyright  2009 Cognitive Science Society, Inc. All rights reserved.

ISSN: 1756-8757 print / 1756-8765 online
DOI: 10.1111/j.1756-8765.2009.01015.x

The Part of Cognitive Science That Is Philosophy

Daniel C. Dennett
Center for Cognitive Studies, Tufts University
Received 31 January 2009; received in revised form 11 February 2009; accepted 11 February 2009

Abstract
There is much good work for philosophers to do in cognitive science if they adopt the constructive
attitude that prevails in science, work toward testable hypotheses, and take on the task of clarifying
the relationship between the scientific concepts and the everyday concepts with which we conduct
our moral lives.

Keywords: Philosophy in cognitive science; Conceptual confusion in cognitive science; Philosophy

and hypothesis-generation; Qualia; Neural correlates of consciousness; Libet

I like Andrew Brook’s distinction between philosophy in cognitive science and philoso-
phy of cognitive science. In principle, you can do one without the other, and some work in
the field can be quite clearly given just one of these labels, but in general I think that good
philosophical contributions to cognitive science must work on both projects at once: it is a
mistake to think, for instance, that philosophy of cognitive science is ‘‘just for philoso-
phers’’ or ‘‘just for the lay public’’ a collection of antidotes to faulty inferences about the
import of contemporary scientific work. Cognitive scientists themselves are often just as
much in the grip of the sorts of misapprehensions and confusions as outsiders succumb to,
and being down in the trenches sometimes makes them even more susceptible.
My parade case would be the large literature by scientists in the wake of Benjamin
Libet’s work on (what he thought was) the consciousness of intention. Here the scientists’
‘‘lay’’ intuitions have misled them into dubious public pronouncements about the import of
the work (for quotes from Michael Gazzaniga, William Calvin, and V. S. Ramachandran,
see my Freedom Evolves, Dennett [2003], pp. 230–31) and also distorted the analysis of
some ingenious follow-up experiments (e.g., Haggard, 2008; Haggard & Eimer, 1999; Lau
& Passingham, 2007; Soon, Brass, Heinze, & Haynes, 2008). All of these experiments rely
on subjects making a most unnatural judgment of simultaneity the import of which is not
carefully analyzed, because of the presumption that the right question to ask is, When does

Correspondence should be sent to Daniel C. Dennett, Center for Cognitive Studies, Tufts University,
Medford, MA 02155. E-mail: daniel.dennett@tufts.edu
232 D. C. Dennett ⁄ Topics in Cognitive Science 1 (2009)

the subject become aware of the intention to act? This leads inexorably to the ‘‘discovery’’
that subjects are shockingly tardy in getting the news about what their bodies are about to
do. This interpretation tacitly presupposes that until a subject can say that he or she has a
particular intention, the subject is in no position to endorse, evaluate, review, or modify the
intention in question. This creates the illusion of an ominous temporal bottleneck, with the
Conscious Agent impatiently waiting (in the Cartesian Theater) for news from the rest of
the brain about what projects are underway. I must add that the literature on the topic by
philosophers includes some that is equally ill considered. However, since better work is on
the way, there is no need to dwell on past confusions. Here the two varieties distinguished
by Brook strike me as merging into a single enterprise: there are philosophical problems
with the cognitive science itself, not just philosophical problems about how to understand or
interpret cognitive science.
There is no dearth of reasons why philosophers are regarded askance (at best) by many in
the scientific community, and they are familiar enough so that I will just acknowledge them
in passing. Philosophers often comically misjudge their competence to evaluate concepts,
arguments, theories with which they have only the most passing acquaintance, and they are
prone to coming up with what might be called bumblebee deductions—proving from ‘‘first
principles’’ that bumblebees cannot possibly fly. I do not exempt myself from this verdict.
For instance, I used to claim that any theory that had to postulate ‘‘grandmother neurons’’
was dead in the water, but that is because I did not understand the possibilities of ‘‘coarse
coding’’ and redundancy, which do indeed make it possible to rehabilitate some models that
have elements that look an awful lot like grandmother neurons (e.g., Quiroga, Reddy,
Kreiman, Koch, & Fried, 2005). I once dismissed any theory that ‘‘replaced the little man in
the brain with a committee’’ as conceptually bankrupt—until I realized that this was indeed
a path, perhaps the royal road, to getting rid of the little man altogether. So live by the sword,
die by the sword. If philosophers are going to make arguments with conclusions about
fundamental constraints, possibilities, and impossibilities of theories or models of mind, they
had better be ready to be shown they have been tricked by their own oversimplifications.
Still, unrepentant, I think we philosophers can make just such contributions; at the very
least we can sharpen and clarify the alternatives, rendering the suspect claims clear enough
to be refuted—which is definitely progress. Wolfgang Pauli famously put down another
physicist’s model as ‘‘not even wrong’’—it was too murky to be a candidate for truth or
falsehood—and philosophers can sometimes help bring ideas into focus that otherwise
would not be refutable or confirmable but that can still influence the direction of research by
providing a sort of prevailing fog that deters people from entering certain regions of likely
theory. One of the reasons cognitive science is such a land of plenty for philosophers is that
so many of its questions—not just the grand bird’s-eye view questions but quite proximal,
in-the-lab-now questions—are still ill thought out, prematurely precipitated into forms that
deserve critical reevaluation. If philosophy is, as my bumper sticker slogan has it, what
you’re doing until you figure out just what questions to ask, then there is a lot of philosophy
to be done by cognitive scientists these days.
That raises the question: Can’t they do it themselves better than carpet-bagging philoso-
phers can? Not necessarily. Many of them certainly think so, but I relish the cases where
 D. C. Dennett ⁄ Topics in Cognitive Science 1 (2009) 233

once they try it, they come to appreciate just how tricky the issues are and come asking for
help. For instance, one of the side benefits for philosophers of the flood of books on con-
sciousness by scientists in the last 15 years or so is the number of them whose authors have
tied themselves in rather embarrassing knots (an example to which we will return).
One of the prime lessons I have learned from scientists is that they have a justified impa-
tience, even contempt, for the sorts of purely destructive criticism, the piling on and hooting,
that philosophers often try to import from the home discipline. Philosophers can sit in the
trees sniping away merrily—but that is not constructive. The background assumption of phi-
losophers always ought to be: if a problem is worth our attention, it is because really smart
people find it difficult; we should humbly try to help them sort it all out, asking, not telling,
being tentative, not preemptive, in our criticisms. I cringe when I see young philosophers
doing a smarty-pants demolition number in front of scientists, a talk that would go down
like honey in a room full of philosophers but merely makes the scientists shake their heads
in dismay.
If philosophers aren’t to do the sort of refutation by counter-example and reductio ad
absurdum that their training prepares them for, what is the alternative? To take on the bur-
den of working out an implication or two that is actually testable, if not today, in the near
future. I have found that if you can help scientists design experiments, they take you more
seriously than they otherwise would do. Conversations with David Premack at a meeting in
Minnesota back in 1975 led to his experiments with the chimpanzee Sarah on deception,
and then my ‘‘Beliefs about beliefs’’ commentary on Premack and Woodruff in Behavioral
and Brain Sciences in Dennett (1978) helped to spawn the experimental industry in false
belief tasks. A few years later, at the Dahlem conference on animal intelligence in 1982,
I proposed my ideas about the intentional stance to a room full of ethologists and primatolo-
gists and was politely asked if this scheme of mine would help them design experiments.
I sat down at lunch with several of them—Robert Seyfarth, Carolyn Ristau, Peter Marler,
Sue Savage-Rumbaugh, Hans Kummer, as best I recall—and in an hour or two we designed
a variety of experiments that they subsequently went off and did, with good results. (See
Dennett [1983] for an early progress report.) More recently, the ‘‘Appendix for Scientists’’
in Dennett’s Consciousness Explained (1991) provided a variety of experimental
predictions that have borne fruit: change blindness most prominently but also experiments
on ‘‘filling in’’ and color phi. Soon there will be experimental results on the ‘‘precognitive
carrousel’’ effect discussed in that book as well, now that noninvasive brain-scanning
technologies permit short-latency tapping of motor cortex activity.1 And, make no mistake,
the experimental work that has resulted has shown how the phenomena in question are more
complicated than I had supposed when I first proposed the experiments. That is where the
real progress lies, of course, not in the straight confirmation or refutation of a relatively
simple ‘‘theory.’’
Communication with scientists is sometimes made difficult because scientists overuse
the principle of charity. They hear about qualia, for instance, and decide that they know
what these philosophers’ qualia are (and of course qualia exist—anybody who has ever had
a toothache knows that!), and when they learn about the controversies about qualia, they
think they can solve them, or at least point to their solutions, with a few observations drawn
234 D. C. Dennett ⁄ Topics in Cognitive Science 1 (2009)

from recent experimental work. In recent work on consciousness (to return to that topic), a
moving example of this is the late Jeffrey Gray, who in 2004 posthumously published the
book, Consciousness: Creeping up on the Hard Problem, which is full of valuable and
important ideas but misses by a fairly wide margin the philosophers’ problems it aspires to
tackle. Gray had simply assumed—and why not?—that philosophers had something impor-
tant in mind, something testable, a difference that could make a difference, when they talked
about the Hard Problem, or the problem of absent qualia and functionalism. He refused to
believe that the issues that engaged the philosophers were as footling as in fact they are. (I
tried on several occasions to show him, but he just laughed: ‘‘Oh no, you can’t be serious,
Dan!’’) So he valiantly charged ahead, doing his experiments, demonstrating the role and
nature of what he called qualia, and thus creeping up on what he thought the Hard Problem
was. It has now been 5 years since his book was published, and I can locate just one philo-
sophical reaction to it (Kriegel, 2007). Why have philosophers ignored the book? Because it
manifestly did not address their hard problem and was not about the qualia they took
seriously. As Kriegel says, putting it mildly, ‘‘. . . scientists use the label ‘Hard Problem’
more loosely than philosophers’’ (p. 97).
Another area of philosophical naiveté in the cognitive science of consciousness concerns
the quest for the Neural Correlate of Consciousness (NCC). It has seemed obvious to quite a
few scientists aspiring to solve the mystery of consciousness that there has to be an NCC,
the necessary and sufficient conditions, characterized in terms of locatable neural activity,
for conscious experiences. How indeed could there not be one, if materialism is true? Con-
sider a parallel question: suppose evolutionary biologists set out to discover the Biological
Correlate of Speciation. Suppose they identify a hundred thousand instances in the past
where population divergence has occurred. Such divergence is, if not always necessary
(sympatric speciation does seem to occur), at least conditionally necessary—a condition
without which no speciation would have occurred in most of the instances studied. But prob-
ably only a handful of those hundred thousand divergences would have led, in the fullness
of time, to speciation. In the other instances, the populations reunited before speciation
could occur, or one population died out, leaving no descendants. Evolution is about the
amplification of effects that almost never happen.
What, these biologists ask, does the subset that proves eventually to have been speciation
events have in common? The answer may be: Nothing! There may be absolutely no distin-
guishing features of those divergences that eventuate in speciation and those that do not,
since whether speciation occurs depends on the happenstances of hundreds or thousands of
generations that come into existence after the divergence is complete. Similarly, it can be
the case that whether a particular contentful event—a discrimination, a binding, a categori-
zation at time t—was conscious can depend on how its descendant effects fared over sec-
onds, not milliseconds (Dennett & Akins, 2008).
There need not be anything mysterian or anti-scientific about the observation that the
quest for the NCC is probably a wild goose chase. If the processes that elevate contents to
consciousness are like the processes that elevate lineage divergences into speciation events,
we would be looking in the wrong place at the wrong time for the hallmarks of conscious-
ness by homing in on the brain processes that occurred at the time of content fixation—at
 D. C. Dennett ⁄ Topics in Cognitive Science 1 (2009) 235

the time, in other words, when the relevant content just began to make a difference in the
subject’s cognitive life.
Scientists will probably always be impatient with philosophers’ scruples about language,
and with the philosophers’ insistence that paths be carefully built from the everyday terms
(‘‘belief,’’ ‘‘dream,’’ ‘‘image,’’ etc.) to their neuroscientific counterparts and then back to
the everyday terms, but philosophers should not cave in to scientists’ hyperpragmatic elim-
inativism. (Besides, they do not live by what they say. They continue to help themselves
uncritically to all the connotations and implications of the everyday terms until they come
to grief on some quandary or paradox.) It is worth remembering that the main reason every-
body—really, just about everybody—is fascinated with, and troubled by, work in cognitive
science is that it so manifestly promises or threatens to introduce alien substitutes for the
everyday terms in which we conduct our moral lives. Will we still have free will? Will we
still be conscious, thinking agents who might be held responsible? Does suffering really
exist? It is because we truly need good, philosophically sound, scientific answers to these
questions and not to any substitutes, that philosophers have a very substantial job to do in
the ongoing progress of cognitive science.

Note

1. My account in Consciousness Explained (Dennett, 1991, pp. 167–168) of Gray

Walter’s reported experiments with chronically planted electrodes has been met with
skepticism by several readers, and digging in the archives has so far failed to yield a
written, let alone published, version of Gray Walter’s talk to the Osler Society of
Oxford that I attended in 1963 or 1964. It has been suggested by some that he was
making it all up, to dazzle the students. Perhaps he was, but the claim he made was
eminently plausible and the experiments he described can now be replicated without
boring holes in subjects’ skulls.

References

Dennett, D. C. (1978). Beliefs about beliefs (commentary on Premack et al.). Behavioral and Brain Sciences, 1,
568–570.
Dennett, D. C. (1983). Intentional systems in cognitive ethology: The ‘Panglossian Paradigm’ defended. Behav-
ioral and Brain Sciences, 6, 343–390.
Dennett, D. C. (1991). Consciousness explained. Boston: Little Brown.
Dennett, D. C. (2003). Freedom evolves. New York: Viking Penguin.
Dennett, D. C., & Akins, K. (2008). The multiple drafts model. Available at http://www.scholarpedia.org.
Gray, Jeffrey. (2004). Consciousness: Creeping up on the hard problem. Oxford, England: Oxford University
Press.
Haggard, P. (2008). Human volition: Towards a neuroscience of will. Nature Reviews. Neuroscience, 9, 934–946.
Haggard, P., & Eimer, M. (1999). On the relation between brain potentials and the awareness of voluntary move-
ments. Experimental brain research ⁄ Experimentelle Hirnforschung ⁄ Expe´rimentation ce´re´brale, 126(1),
128–133.
236 D. C. Dennett ⁄ Topics in Cognitive Science 1 (2009)

Kriegel, Uriah. (2007). Gray matters: Functionalism, intentionalism, and the search for the NCC in Jeffrey
Gray’s work. Journal of Consciousness Studies, 14, 96–116.
Lau, H. C., & Passingham, R. E. (2007). Unconscious activation of the cognitive control system in the human
prefrontal cortex. The Journal of neuroscience: The official journal of the Society for Neuroscience, 27(21),
5805–5811.
Quiroga, R., Reddy, L., Kreiman, G., Koch, C., & Fried, I. (2005). Invariant visual representation by single
neurons in the human brain. Nature, 435, 1102–1107.
Soon, C. S., Brass, M., Heinze, H., & Haynes, J. (2008). Unconscious determinants of free decisions in the
human brain. Nature Neuroscience, 11(5), 543–545.