Front. Philos.

China 2018, 13(3): 332–348

DOI 10.3868/s030-007-018-0026-1
 
 
SPECIAL THEME

Giulio Tononi, Owen Flanagan

Philosophy and Science Dialogue: Consciousness

Abstract This is a dialogue between a philosopher and a scientist about the
scientific explanation of consciousness. What is consciousness? Does it admit of
scientific explanation? If so, what must a scientific theory of consciousness be like
in order to provide us with a satisfying explanation of its explanandum? And what
types of entities might such a theory acknowledge as being conscious?
Philosopher Owen Flanagan and scientist Giulio Tononi weigh in on these issues
during an exchange about the nature and scientific explanation of consciousness.

Keywords consciousness, integrated information theory, phenomenology, hard

problem of consciousness, scientific explanation

What do we mean by our uses of the word “consciousness”? What do we mean

when we say that a given entity is conscious?

FLANAGAN: “What do we mean by ‘consciousness’?” That’s a different

question than the one we refer to when we talk about consciousness. This goes
back to the old meaning-reference issue in philosophy and it’s actually relevant to
this case because “consciousness” is not a scientific term in the first instance. In
classical Chinese, the term “心” represents heart-mind and it refers to something in
the vicinity of what we today call mind. There’s an interesting question in the
western tradition whether what we called mind from the beginning of, say, Greek
philosophy was the conscious mind or just the mind including things like what we
 

Giulio Tononi ( )
School of Medicine, Department of Psychiatry, University of Wisconsin–Madison, Madison,
WI 53719, USA
E-mail: giulio.tononi@gmail.com

Owen Flanagan ( )
Department of Philosophy, Duke University, Durham, NC 27708, USA
E-mail: ojf@duke.edu
Philosophy and Science Dialogue: Consciousness 333

now call the computational unconscious or the Freudian unconscious. It looks like
the conscious and the unconscious were woven together at the beginning. By the
time we get to the 17th century and Descartes, the word “consciousness” means
something about the first-person perspective. When Descartes says, “I think,
therefore I am,” what he means by “think” is that he perceives, he doubts, he feels,
etc. It encompasses all the mental states that seem to have phenomenal character or
“what it is like”-ness from the first-person point of view. That’s what
“consciousness” typically means.
What “consciousness” refers to is an entirely different question. Back in ancient
Greece, we have the classic example. If you wake up in the morning and say,
“Look, the morning star,” someone during that time might have informed you that
the morning star is called “Phosphorus.” I might, then, while writing poetry, say
that your eyes are like Phosphorus where I use “Phosphorus” to mean the morning
star. In the evening, I look up at the sky and see the evening star which is called
“Hesperus.” I then write a different poem that says your eyes are like Hesperus, the
evening star. An expert astronomer then comes along and says, “Even though
‘Phosphorus’ means morning star and ‘Hesperus’ means evening star, we’ve
discovered that they are one and same heavenly object. In fact, they are not stars at
all, they both refer to the planet Venus.” This is a nice case because the meanings
of the terms are different and though they seem to refer to two different things, as it
turns out, they refer to one and the same thing. That has lessons for what we are
talking about here.
What we are phenomenally aware of and what “consciousness” has historically
meant is the first-person experience. What “consciousness” refers to and how
consciousness is actually realized is a little bit like the evening star/morning star,
that’s up to experts (like astronomers and scientists) to tell us. That’s not for any
old person to figure out on his or her own.
At the end of the day, there’s a problem, and here’s why the poetry example is
important. If I say in my poem that your eyes are like the evening star and someone
says you could shorten the poem and simply say your eyes are like Venus, then one
could respond that that won’t work because that changes the meaning of the poem.
It changes the meaning even if they refer to the same thing. So, the million-dollar
question is: given that we know what consciousness is first―personally in an
intuitive way (like the way we know about the evening star/morning star and the
way we know water is wet)―how do we learn about its deep structure? That
requires scientific theory.

TONONI: I agree with everything. I will simply add a couple things. First, it is
always useful to remind ourselves that consciousness refers to “subjective
experience”—“what it is like to be.” Having experiences can happen in dreams
just as much as when we are awake. And even the experiences of reflecting hard
334 Giulio Tononi, Owen Flanagan

on the mind-body problem is just another experience. Since consciousness refers

to any experience, we shouldn’t artificially limit what it encompasses.
Second, while we can give examples of this common-sense definition―what
goes away when you fall into dreamless sleep, what goes away when I hit you on
the head, etc.—what we need to do is to slowly try to substitute these
common-sense definitions with scientific definitions. We can perform the same
trick that we did with the evening star and the morning star and say, as Integrated
Information Theory (IIT)1 tries to do, the following: an experience is identical
with a conceptual structure which is a maximum of intrinsic cause-effect power of
a certain form. This identification provides a mathematical formalism to calculate
this form and, consequently, a testable scientific identity. Is it true that when I have
an experience there is a conceptual structure there, precisely defined in
mathematical terms? If it turns out that this is true time and again then we begin to
trust the theory. We then extrapolate beyond the cases that work. If, in this process,
we find that it works in some cases but there are other cases where the theory falls
apart then we are forced to change or dismiss the theory.
Third, when we study consciousness and try to come up with a scientific
understanding, we don’t only have to explain that there is consciousness and how
much there is, but also explain why it is that particular kind of consciousness (e.g.
why it is itchy and not ticklish). So, the qualitative and quantitative elements of
consciousness are always intertwined. In neurology, we often talk not only about
the level (or amount) of consciousness but also the content of consciousness. We
test the subject along both lines. Is the subject fully alert (fully conscious)? Is the
subject in stupor (somewhere in between)? Is the subject unconscious? That’s the
quantity. The content is what in particular the subject is conscious of. What is the
subject experiencing? The subject might say that she is experiencing this or that. If
the subject is a poet or a writer she might fill a hundred pages to describe what she
was experiencing in that moment. And the subject still won’t succeed very well.
So, there is a distinction between the level and the content.
If you take IIT, however, the two are inextricably connected. You can say the
level is φ, the amount of integrated information, and the content is the form of the
conceptual structures. A nice feature of some scientific theories, assuming it’s
roughly correct, is that they begin tying together things that were previously
thought to be disparate. This is what we have come to expect from many areas of
science.

A worry might be that in the consciousness case the meaning/reference

                                                             
1
For more on this see Tononi (2016).
Philosophy and Science Dialogue: Consciousness 335

distinction dissolves. This may be what David Chalmers (and others) are after
when they attempt to make a distinction between the Hesperus-Phosphorus-
Venus case and the consciousness-brain case. It seems that what we mean is what
we’re referring to. Once you bring in the science it almost seems like we’re
changing the topic. Can you help us make progress in this area? Can
consciousness be scientifically explained?

FLANAGAN: David Chalmers is a good friend of mine, but I don’t think he

holds the view that you suggested. Galen Strawson has a wonderful paper in The
Times Literary Supplement (2015) saying this problem isn’t all of a sudden
interesting in 1996 when Chalmers wrote his book (1996). This is one of the
oldest problems in history—the difficulty of explaining how consciousness is
possible in a “material” world. It’s a really old philosophical problem.
In fact, it’s also a universal problem. It’s a problem often found in philosophical
theologies because almost all cultures believe in life after bodily death. Some
cultures have a non-karmic eschatology that doesn’t involve survival after bodily
death and going to a spiritual place. Every time a non-karmic eschatology meets
something like Hinduism, Buddhism, Jainism or Christianity, however, it gets
smashed. The theories that say you survive bodily death always win. People
simply find theories that have consciousness continuing beyond bodily death
extremely attractive. So, the old problem for almost every classical religion is
this: How is it possible that I could survive my bodily death if my mind is
realized in a body?
Now concerning the potential weaknesses in the Hesperus example as it
applies to the mind-body case, some philosophers admittedly do say that pain is
essentially the experience of pain and this differs from the Hesperus example
because Hesperus is not essentially the experience of Hesperus. Well, I can agree
with this up to a point. But we must keep in mind that the experience of pain is
the way pain comes to you. How is this relevant to the question we’re talking
about here—can we explain consciousness? Can we have a constructive theory
which can transcendentally explain how creatures like us can have mental states
that are consciously experienced? Can we embed this in a constructive
naturalistic solution?
First, we must take the phenomenon seriously. We must take the psychology
seriously. We must also study what consciousness does—when you’re awake, it
keeps you from falling into holes, it keeps you from bumping into cars, it keeps
you moving through space successfully, it leads you to water if you’re thirsty, etc.
Of course, I don’t think the motor features themselves are conscious, just that
consciousness has this ecological role.
Then the question is: What kinds of systems can realize experiential states like
336 Giulio Tononi, Owen Flanagan

these? The tradition that most of us think is promising is to proceed by focusing

on minimal sufficiency. Here’s an example. What conditions of the human body
are minimally sufficient for your not being alive anymore? I chop your head off.
That’s sufficient for your not being alive. That’s not necessary, however, since
there are a lot of other ways for your not being alive—your heart can stop. What
about for consciousness? The cerebellum doesn’t seem to do the job. 40 Hz
oscillation doesn’t seem to do the job.
We then try to build a theory in response to the question: How is it possible
that there are creatures that have experiences? Some people say—you better look
in the nervous system of people. Some people are even braver and say—well,
maybe it doesn’t take a nervous system but perhaps a Pythagorean or Euclidean
structure or set of logical structures. That, I believe, is what IIT tries to give us.
A final caution about responses to theories. The astronomers discovered that
the evening star is the morning star. In fact, they discovered that it’s not a star at
all, it’s the planet Venus. That requires a pretty big change in our conceptual
structure, but at least they are all heavenly bodies and we have a common
underlying framework for understanding motion. Some people might say,
however, no theory of consciousness can provide understanding of how
consciousness arises. The theory, after all, is just a model. It isn’t consciousness
itself. The theory doesn’t produce consciousness, it’s merely a set of equations or
generalizations about the kinds of systems that produce consciousness. A
generation of philosophers, in line with this sentiment, have gotten tenure by
scratching their heads and saying we can not solve the hard problem. That is
100% unhelpful. Yes, we do not have a solution to the hard problem. All we have
at the moment are some theories. But what you have to do is let the best
theoreticians start building neuroscientific explanations about how it’s possible.
Does quantum physics yield a perfectly satisfying explanation of how the world
works? Maybe not. I do not understand quantum physics so it’s not perfectly
satisfying for me. But then I think that’s just a problem with me, not the theory.

TONONI: First, going back to Chalmers and the hard problem of consciousness,
he acknowledged that it was a good public relations move on his part. This is
what he said in Tucson, Arizona (Chalmers 2014). It was the 300th anniversary of
Leibniz’s Monadology which, among other things, contains a beautiful thought
experiment: Leibniz’ Mill.
Leibniz invites each of us to enter this giant mill with wheels and ropes going
around. It’s a very intricate machine. Now think about your brain. Make
everything in your brain very big so that you can walk into it like Leibniz’ mill.
You will see the axons, the synapses, the dendrites and the cell bodies. You name
it, whatever you want. As you walk around your brain you realize it’s just a giant
machine! When you approach things that way it seems impossibly hard to
Philosophy and Science Dialogue: Consciousness 337

understand how your brain has anything to do with consciousness. If it’s just
mechanical or electro-chemical, why would there be experiences associated with
it? It seems impossible.
Second, in order to make progress, you must go in exactly the opposite
direction. You should not start with the postulated fact that the world is made of
atoms and neurons. Instead you must begin with yourself—you certainly exist
and you are certainly conscious. The model that we are working with is about the
world, not about my consciousness. My consciousness is foundational, it exists
period. But our understanding of physics is a model of the world. It can be a
good model, but it is, nevertheless, a model. This way of seeing things inverts the
problem and, I believe, this is the way to approach consciousness. If you don’t
start in this way you’ll never get anywhere.
Third, by using IIT as an example, let’s consider what it means to explain
something. In what sense is IIT a scientific theory? There are many things one
can ask of a scientific theory in terms of explanation. One is, it must have
theoretical validity and coherence. It needs to be a theory with a few primitives
or principles from which many other things follow. We strive to avoid
explanations that are as long and complex as the phenomenon itself. We know
this intuitively as scientists. We know that the best explanation is the shortest
explanation. The theory working from the fewest principles is good.
You also want this theory to account for a vast array of data based on these few
principles. The classic example in biology is natural selection. That has been
dubbed by Daniel Dennett as the best idea anybody ever had. Rightly so because
natural selection gives us a beautiful and general framework for explaining all of
the many different things on earth that all look impossibly different.
More generally, successful examples in physics involve the ability to take
many different phenomena and show that they boil down to the same thing. The
Hesperus and Phosphorus example is a simple-minded version of this idea. You
have two disparate things, two different stars, but then we learned that one planet
accounts for all the observable phenomena.
Finally, you want to have predictive power. If you have a good theory, not only
do you have relatively few principles to explain many disparate things, you also
make extrapolations. If things are like this, then others things should be like that.
And the more counterintuitive those predictions are the better the theory is. This
is what you want from a good theory.
IIT tries to do all of this. It’s based on a few principles. It explains many
phenomena. It explains why the cerebellum does not generate consciousness
even though it has so many neurons. It explains why you lose consciousness
during sleep even if the brain remains active. It makes predictions. For example,
if IIT is correct, then a brain that is not active yet functional should have
consciousness. Some of these predictions have been tested very roughly, others
338 Giulio Tononi, Owen Flanagan

need to be tested but the testing is, in principle, possible.

If consciousness can be explained, can it be explained in purely physical terms?

FLANAGAN: One idea I had when I wrote Consciousness Reconsidered (1992)

is this theory I called “Constructive Naturalism.” I took it for granted that we are
conscious. But it would be useful if we could have access to as much
non-theoretical phenomenology of persons as we can get. We can learn about this
even from non-human animals. Even at the time I was writing this book there
was good work being done by Nikos Logothetis and Jeffrey Schall on binocular
rivalry in Macaque monkeys (Logothetis and Schall 1990). So, I claim that you
can get phenomenological reports even from monkeys. It’s fairly simple. You
create a situation where you can’t see both the red and the blue lines at the same
time—this is a phenomenon called binocular rivalry. Even though the lines are
both there, you’ll flip-flop. It’s similar to a gestalt illusion like a Necker cube
where you see the back or the front. What they were able to do is to train these
monkeys to give reports of whether they saw the red or the blue lines (or the
vertical or the horizontal lines) by saccading. They rewarded the monkeys when
the monkeys gave reliable reports. By looking left they signaled that they saw
blue and by looking right they signaled that they saw red. Then the scientists
isolated the area of the brain that looked like it was usually active when they saw
blue (and a different area for red). So, the psychology relies on the behavioral
report connected to the phenomenology which cues studies into the neurological
structures of the brain. This kind of research would at least tell you something
about how the phenomenology lines up with the behavior and with what’s going
on in the brain.
Clearly, at least in the case of a human, something is going on in our nervous
system when we have experiences. This is under the assumption that we are
biological beings living in this particular world. If consciousness is supervening
on anything, it better supervene on that.
What about other non-biological creatures? I have always been interested in
how consciousness works in us, but here, with IIT, we have a theory that goes
deeper.

TONONI: Many people understand the term “physical” differently. In the end, it
has to do with things we can manipulate and observe. One of the important
manipulations is to cut. Common sense has gone a long way by cutting and
observing. This, after all, is what our naïve realist view of the world is based on.
Philosophy and Science Dialogue: Consciousness 339

The world is made of things, middle-sized typical objects, that we can

manipulate, observe, and cut. That’s what the world is made of.
The scientific view has gone much further than that. It has developed a model
of what can be manipulated, observed, and cut. What we try to do in science is to
go down to the smallest things we can manipulate, observe, and cut and claim
that these are the building blocks of reality. We want our theory to say that
everything is made of these little blocks. Depending on how they are configured
and interact with each other, we can explain everything. So, the physical world
boils down to whatever our best theories say that the world is made of. It’s made
of the smallest things we can manipulate, observe, and cut. That’s the physical
world.
Within this physical world there is my consciousness. Suddenly we have the
mind-body problem. The physical world is made of these little building blocks
that follow laws and have cause-effect power. But how does consciousness fit
into this picture?
IIT says that everything that exists has cause-effect power. This is what every
person intuitively accepts—everything that exists must make a causal difference.
Moreover, we can make a distinction among things that exist. Some things exist
intrinsically (in and of themselves). These things have cause-effect power
intrinsically. Other things exist extrinsically, things that exist for something else.
The physical world, then, would boil down to whatever are the smallest things
that exist extrinsically, since they are apprehended from our perspective.
Consciousness exists intrinsically, it exists in and of itself.
The beauty is that this can all be expressed in terms of cause-effect power. So,
we exist intrinsically and the physical world exists extrinsically, it is derivative of
our existence.

FLANAGAN: An interesting consequence of IIT is that there must be something

it is like, even minimally, to be, say, a diode—the light that flashes on and off in
response to darkness. The way IIT spells this out, of course, means that there are
different platforms that can have causal powers. IIT doesn’t say anything about
the constitution of the systems that have consciousness, it simply says that
conscious systems must have a high φ value. Otherwise it would be biologically
chauvinistic.
Is consciousness going to be explained neurobiologically? For us, it will be
explained neurobiologically. But, is the nature of consciousness limited to us?
It may be bad public relations for you, Giulio, to be doing three things
(actually three and a half) with your theory simultaneously. All we really want is
a scientific theory and you are providing us with one. In fact, you are providing
us with one of the most exciting scientific theories of consciousness presently
available.
340 Giulio Tononi, Owen Flanagan

Beyond this, there are also the general questions involving the metaphysical
and epistemological implications of a scientific theory. I think you say something
about both of these things that involve unfamiliar terminology for philosophers.
Insofar as they are rigorous, they don’t sound right even on your own view. So,
let me give you some examples.
First, you often say the only thing that is real, or the only thing that matters, is
intrinsic existence. Now intrinsic existence, philosophically, usually means
something like existence which has no relational properties whatsoever. But I
think, on your view, you emphasize the claim that the qualitative system has
inputs and outputs. That is, when I have a blue experience there are light waves
being received by my retina that form signals being sent up to my brain to be
computed. When I have an auditory experience, there are sound waves being
received by my ears that also form signals being sent up to my brain to be
computed. That means that the system that has the property of being conscious
has relations with the external world. I don’t want to say that the sound waves
coming in are conscious or that the light is conscious. While I’m happy to say
that the qualitative properties are contained in the system the way IIT suggests, I
think that we’re not really dealing with intrinsic existence in IIT since we’re
dealing with matters of function.
It also looks to me that your metaphysics can sound like transcendental
idealism. It’s almost as if you claim that the only things that exist are subjective
beings and their experiences. There’s no real reason to posit an external world. In
fact, the dream metaphor, which I like, can also function that way. All the work
that you and your team have done on the phenomenology of what we call
dreaming (non-REM and REM2) is well studied and it’s a different kind of
consciousness than wakeful consciousness. So, I think it’s important not to
tamper with these distinctions.
Moreover, we have a general theory about reality (cobbled together from all
the sciences). This then lends itself to an epistemology which tells us how
reliably we’re picking up and detecting what’s out there in the world. I don’t
think we need to tamper with this either—especially not everything at once.
On all our best scientific theories, the universe has been around for 10 billion
years. The cause and effect relations were operating for 10 billion years before
we arrived on the scene. And the world may well be here after we leave. Are you
questioning the physical view of the universe as a system of cause and effect
relationships that existed long before sentient beings ever showed up? We could
say that we’re the only kind of existence that matters. I’m almost certain that for
most of us, we’re the only kind of system that we care about, but that seems to be
a psychological fact about us, not a claim about what reality is.
                                                             
2
REM—rapid eye movement.
Philosophy and Science Dialogue: Consciousness 341

In your theory, however, all these different features are being addressed
simultaneously and there is potential for confusion and unnecessary conceptual
mistakes.

TONONI: I can extract two or three main criticisms from what you said. One,
leaving aside metaphysics and epistemology, is the natural criticism of
overreaching. I might respond by saying that if a theory of consciousness has any
pretense of being in the right direction then necessarily it will have very
far-reaching impact. It’s the nature of the beast. Schopenhauer called it the
world-knot because it’s at the center of everything. If consciousness turns out to
be, say, intrinsic existence as defined by IIT then it’s going to change our views
about what it means to know about the world. It will change our view about what
is worth treating as an entity that has value in and of itself versus an entity that
only has functional value. I think if it is a good theory of consciousness it will
necessarily have an impact on these things. Having said that, before trying to
make these connections one must rigorously and carefully develop the theory
along with its mathematics.
The second important point you raised is in essence that IIT may ground a
solipsistic nightmare. I will say with no hesitation that there is a real world “out
there.” I don’t have the slightest doubt about it. But the way to look at it is,
perhaps, different. One way to think about it is that we always have a model of
how the world is. There is the common-sense model of rocks and trees and there
is the scientific model of rocks and trees. We are constantly trying to integrate
these models of the world but the scientific model is becoming increasingly
difficult to understand. I look at you and I know that you have a brain that is
responsible for the way you are looking at me. 2000 years ago, I didn’t know that
and I would have understood you differently. All of this knowledge has been
incorporated into the scientific model.
Pure solipsism will say that it’s all just a dream. It is a position that is
impossible to prove wrong. But, at the same time, it doesn’t make sense.
Consider all the regularities that are in my experience. Does it happen according
to rules in my dream which have no further explanations? It will be predictively
useless. It will explain nothing. Instead, if there is a world outside of myself
which is made of stuff that follow certain rules then the scientific model is a good
model—a powerful one with excellent predictive resources.
But it’s still possible that cause-effect power is the foundation for everything
and, on this theory, we can tell ourselves a story that accounts for both the
regularities in my flow of experiences, my own “movie,” and also for the
regularities in the “movie itself.” That’s the divide between intrinsic and extrinsic
existence. It is a realist position. There is a world. The panpsychist version of IIT
would say, the world is constituted of things that exist intrinsically—lots of
342 Giulio Tononi, Owen Flanagan

various small entities. Everything that exists, exists intrinsically. If they enter
particular relationships, bigger entities come to exist intrinsically. They can get as
big as us. For these entities, all the other entities that exist intrinsically start to
exist extrinsically in the shape of trees, people, and so on. It is realist, but you
must keep in mind that it’s a model.

FLANANGAN: I agree with everything you said. I’m still going to resist the
use of “intrinsic” and the way you use it may not be as helpful as you think
since the entities you posit still seem to be constituted by relations. I agree that
the right scientific theory will change everything. It changes our metaphysics,
epistemology, and ethics. But I am concerned that because everything changes so
radically, it can easily become difficult to keep track of the relevant issues in
clear ways.

TONONI: Intrinsic in IIT means, in line with the original use of the term in
Latin, “from the inside, i.e. from the perspective of the entity itself.” It does not
mean independent of everything external, as in many philosophical writings.
I will add one thing. At the end of your second lecture, I showed a few slides
in which we have these animats3 and they interact with a simple environment.
Everything is simulated. Nevertheless, the animats increased their level of φ (or
their level of consciousness) by adapting to regularities in their environment in
order to survive. This is useful because it offers us the beginnings of a
perspective on why it might be that consciousness evolved. This suggests that
there is an explanation for why ceteris paribus there is a trend towards
consciousness. With limited resources, it pays to do things in an integrated way.
If consciousness, as IIT claims, is integrated information, then everything fits
nicely together.
The other thing we are working on related to epistemology is the notion of
matching. If you exist as a set of complicated forms that evolved over time, these
forms must be adapted to the environment such that you do the right things at the
right time. So, there must be some correspondence between the internal forms
and the external causal structure of the world. This correspondence can be
measured mathematically and is always relative to the conscious entity. There is
no one way for the world to be represented. It is essentially relative to how I
categorize it according to the kind of consciousness I have. It’s an interesting
story and it’s only beginning. While meaning is all internal, like the concepts
allied with the morning star and the evening star, it nevertheless refers to things
out there. This reference relation is captured by the notion of “matching” that we
are working out mathematically. It gives us another way of approaching
                                                             
3
To see these animats in action, see http://integratedinformationtheory.org/animats.html. 
Philosophy and Science Dialogue: Consciousness 343

perennial philosophical problems.

FLANAGAN: I was thrilled by the modeling of the animats—that was

interesting. I always wanted it to be the case because it always seemed plausible
that consciousness should be an adaptation in the classical sense of increased
fitness. The possibility that it was a spandrel or free-rider always bothered me. In
saying this, of course, I’m just reporting something about my own psychology. I
wanted it to work out this way.

Can we build conscious machines? What is it, if anything, that is special about
humans?

TONONI: You can build conscious machines. If IIT is correct then as long as a
machine satisfies the relevant principles it will be conscious. The problem is the
machines we are currently building are most likely not conscious. At most they
will amount to one bit of conscious dust. This is because of their nature as Turing
Machines. Everything goes through an ALU (Arithmetic Logic Unit) and that is
not, from what we can tell, a good architecture to have a system that has a high φ
value. Indeed, you can simulate something with very high φ without having any
φ. You can make a Turing Machine that does everything you can do. It simulates
everything perfectly. It simulates someone who has consciousness and high φ.
Yet, it’s implemented in such a way that it will have no φ.
In principle, we could build a conscious machine that is not made of neurons.
It’s unclear right now how close such a conscious machine would be in terms of
its similarity to a network of neurons, but there is no principled reason why a
silicon-based system (like a neuromorphic computer) could not be conscious.
The functionalist approach says that as long as you reproduce the function of
something, then there’s nothing else to be said. This is very common in
philosophy and in science. We study ourselves in terms of machines and every
neurophysiologist will tell you that once they understand how you go from input
to output with all the internal processing steps, there’s nothing more to
understand. That is absolutely false in IIT. If you did that for the brain, if you
knew exactly what happens in every neuron at every time step, you may be able
to predict what the brain does, but you would, in some sense, understand nothing.
You would miss what I call the elephant in the brain—consciousness! You would
have no idea that, while the whole brain is doing this, in some parts of the brain
there are experiences and in other parts things are going on, but there are no
experiences. You would miss the most important thing about the brain.
344 Giulio Tononi, Owen Flanagan

More generally, whenever we reproduce functions by brute force without all

the constraints that biology has to deal with, which we are doing more and more
successfully with computer programs, we might be able to carry out the functions
just as well, but there will be no conscious experiences. Those machines, even if
they were functionally indistinguishable from you, would be completely
unconscious. So, it depends on how the functions are carried out. What is done is
not important. If this is right, this will make a critical difference in the way we
work out our ethics, regulations, and laws. We want to know, for example,
whether a brain-damaged person is conscious or not. But soon we may have a
much bigger problem. We want to know whether or not these computers that will
behave better than humans (in a growing number of domains) are conscious. If
we don’t have a good answer to these questions, we’re going to be in trouble.

FLANAGAN: As far as human distinctiveness, my answer is connected to my

suspicions about language. I can illustrate this by thinking about what I call the
Q-self—the qualitative self. It’s the self we carry with us over time. How does
that Q-self work? I wrote a book called The Really Hard Problem (Flanagan
2007). I started that book by saying that the hard problem will be solved in
Wisconsin, but the really hard problem will emerge after we accept that we are
physical animals living in a physical world. Then we will be faced with the
problem of existential meaning or purpose. Why do I care so much about leading
a good human life? What does human flourishing consist in? These are questions
that seem to be uniquely human. What makes them possible, it seems, is that I
can look at my qualitative self over historical time. I know, right now, as a grown
man nearer to death than to birth that I have lived a certain life. There are some
things I can bring back very vividly. How do I bring them back? Part of how I do
this is connected to language. But it’s not completely linguistic. These are
interesting features of human consciousness that lead to our ability to hold each
other accountable and that hinge on our capacity to be accountable to ourselves
by holding our own biography in our heads. This has something to do with
memory and its being linguistically scaffolded. I think this is hugely important
and is what makes our consciousness and our sense of ourselves very different
from the rest of the world on earth.

Concerning the evolutionary role of consciousness, why do we have complicated

systems (like the cerebellum) that are not conscious? Wouldn’t everything be
driven to consciousness?
Philosophy and Science Dialogue: Consciousness 345

TONONI: Consider the outside world as a task domain—you must carry out
certain tasks in the world in order to survive. There are some tasks that are, by
nature, rather modular. Certain forms of physical movement, for example, can be
easily modularized. Other tasks can’t be broken down or analyzed in this way.
When you are faced with processing the causal structure of the world as an
organism, some aspects of that causal structure can be modularized. Other
aspects are not of this nature. Consciousness is what you need, given constraints
on the number of available processing units, to deal with the most complex and
integrated parts of the world. Without certain levels of integration, you will not
be able to effectively work with these parts of the world. These parts of the world
are integrated causally across time. Instead, things that can be dealt with in a
modular manner, like ensuring that a muscle doesn’t always stretch, can devolve
into an unconscious localized module that does its job very well. In fact, it does it
better if it’s not taking care of all kinds of other tasks at the same time.
So in some sense the prediction would be that for all of those tasks that can be
modularized, evolution will be able to construct mechanisms that successfully
perform the relevant tasks without developing consciousness. When you are
trying to take in regularities that work over large spatiotemporal scales, then
evolution will be forced to develop something like consciousness.

Does George Lakoff’s work have any influence on consciousness studies?

TONONI: One of Lakoff’s most famous books is Women, Fire, and Dangerous
Things (Lakoff 1990). What is consciousness most like of these three?
If you remember, IIT makes a distinction between low and high-level concepts,
where a concept is a cause-effect repertoire within a conceptual structure. This
may sound like a new language, but it’s interesting because it is an attempt to put
the perceptual and the conceptual together. A high-level concept is what is
usually called an invariant or an abstraction. A face can be a high-level concept
because it is a disjunction of many lower-level conjunctions. Much of our interest
in consciousness derives from high-level concepts that makes language possible.
Describing the details of a visual scene, pixel by pixel, is useless. But saying
there is a person there is very useful. The concept of a person, however, is a
high-level abstraction. One important thing about this that may seem trivial is
this: If we do have high level concepts in our brain (as we know we do) that
develop in a certain way through plasticity over the course of evolutionary
development, then anything more that you learn (including a beautiful poetic
metaphor) is going to have to be built on the basis of stuff you already have.
346 Giulio Tononi, Owen Flanagan

Consider, for example, the metaphor of the “container.” We know from

neuroscience that there are some neurons that quickly analyze an input to find
out whether it is “open” or “closed.” This is important evolutionarily because it
can tell us whether it is “escapable” or “inescapable”—can I get out or not? It’s
very important and very basic. The brain calculates this quickly and
automatically. We have neural structures that correspond to these concepts. When
you develop language and poetry, these basic concepts become the building
blocks by which you make metaphors. Every metaphor will refer back, in some
way, to these primitive blocks.
This is similar to evolutionary biology where you have a heart that pumps
blood. No matter how the heart develops, it will continue to be connected to the
issue of blood circulation. There is a lot of richness to be explored here. What
Lakoff was saying was beautiful and true. We are acquiring the means of
understanding this in progressively better ways.

FLANAGAN: A quick addition to that. Lakoff and Johnson talk about metaphors
we live by (Lakoff and Johnson 2003). The example of one’s beloved, the
brightness of her eyes, is a common cross-cultural metaphor. It’s ubiquitous. Of
course, the fact that we live by conceptual structures that operate at the
metaphorical level doesn’t mean that science is best done by using metaphors,
and that relates to my suggestion that the metaphor of consciousness of being
like a dream doesn’t quite work. The dream phenomenology that I have is
different from the phenomenology I have when I am awake. I don’t want to
absorb it all into one.

Are you a mysterian?

FLANAGAN: It’s certainly true that the problem of consciousness has been
around for a long time. Humans are very selfish about time intervals. We want
big scientific problems solved by next Thursday in our lab. So, I admit, the
problem of consciousness is difficult. The fact that a problem has a history that is
centuries (or even millennia) old, given that the universe has been around for 14
billion years, doesn’t seem to be very long.
There are two kinds of mysterians I defined back in the 1990s. One is an “old”
or traditional mysterian—they are dualists. They will hold that you’ll never be
able to solve the consciousness brain problem physically because the mind is not
a physical thing. You can’t absorb the mind into a physical theory because it’s
simply not a physical thing! It will mean that the nature of consciousness will
Philosophy and Science Dialogue: Consciousness 347

always remain outside the realm of science.

The “new” mysterians, like McGinn (2000) (and possibly Thomas Nagel and
David Chalmers), are people who think you can’t solve the mind-body problem
even though consciousness is realized in the physical world. It is completely
natural but the relevant conceptual connections are permanently missing. It
would be like Gödel’s theorem where you can’t prove the completeness of first
order arithmetic inside first order arithmetic (or Heisenberg’s uncertainty
principle). These are sometimes called impossibility results.
It’s an empirical question whether or not we’ll get to the point of having a
theory which satisfactorily explains consciousness as opposed to providing a
satisfying explanation for anyone who wants to “feel” like they understand. It’s
like quantum mechanics.
One significant obstacle is that, in the West, dualism is in the blood and bone.
We’re not born dualists, the way Paul Bloom argues, but there’s a long historical
tradition that makes us think dualistically that is partly connected to theological
issues. This makes it hard to resist dualism.
When I use the word “mystery,” I mean it more innocently than these two
versions of mysterianism.

At the end of inquiry, will the explanation of consciousness be a priori or a

posteriori?

FLANAGAN: Kant’s method involves what he called transcendental deduction.

We experience the world spatially and temporally. How do we do this? We must
have Euclidean geometry in our heads. What Kant calls transcendental deduction
is, upon closer inspection, really inference to the best explanation. He calls it
deduction but it’s not.
It is like Giulio’s view. A system is doing X. How is it possible for a system to
be doing X? We generate axioms and postulates and see how these can account
for X. This is not deduction. We’re not getting theorems out of the system the
way a teacher teaches the Pythagorean Theorem. If you understand the axioms
and postulates of Euclidean geometry and you follow the teacher’s proof each
step of the way, when you get to the end you don’t think: the teacher’s opinion is
interesting. Instead you will see that the Pythagorean Theorem is necessarily true.
This is what we can get out via the a priori. I don’t think we’ll get anything
remotely close to this out of a scientific theory of consciousness. It will be an
inference to the best explanation, not like Frank Jackson or David Lewis
348 Giulio Tononi, Owen Flanagan

(Australian rules philosophy). Rather, it will be a very complicated theory that

fits together with the phenomenology.

References
Chalmers, D. 1996. The Conscious Mind. Oxford: Oxford University Press.
Chalmers, D. 2014. The Hard Problem of Consciousness: 342 Years On. Toward a Science of
Consciousness. Tucson, Arizona: Center for Consciousness Studies, University of Arizona.
Flanagan, O. 1992. Consciousness Reconsidered. Cambridge, MA: MIT Press.
Flanagan, O. 2007. The Really Hard Problem: Meaning in a Material World. Cambridge, MA:
MIT Press.
Lakoff, G. 1990. Women, Fire, and Dangerous Things: What Categories Reveal About the
Mind. Chicago: University of Chicago Press.
Lakoff, G., & Johnson, M. 2003. Metaphors We Live By. Chicago: University of Chicago
Press.
Logothetis, N., & Schall, J. 1990. “Binocular Motion Rivalry in Macaque monkeys: Eye
Dominance and Tracking Eye Movements.” Vision Research 30.10: 1409–19.
McGinn, C. 2000. The Mysterious Flame: Conscious Minds in a Material World. New York:
Basic Books.
Strawson, G. 2015. “Consciousness Myth.” The Times Literary Supplement, February 25.
Tononi, G., Boly, M., Massimini, M., & Koch, C. 2016. “Integrated Information Theory: From
Consciousness to Its Physical Substrate.” Nature Reviews: Neuroscience 17: 450–61.