by Stanislas Dehaene (editor)
MIT Press, 2001
Review by David A. Patten, Ph.D. on Sep 6th 2002
This book achieves its aim of
demonstrating how recent advances in cognitive neuroscience have contributed to
our understanding of conscious experience.
The book is a reprint of a special issue of Cognition: International Journal of Cognitive Science. Consequently, its audience is not the layperson,
but philosophers, psychologists, or anyone else who has given serious scholarly
thought to the nature of consciousness.
The volume brings together eight articles from leading proponents of the
cognitive neuroscience approach to understanding consciousness and the mind in
general. The articles can be divided
into three basic types. Some of the
authors utilize a cognitive neuroscientific approach to explaining empirical
data, others address methodological concerns, and still others wrestle with
philosophical questions about what it is that is being explained. There appears to be a wide range of
agreement amongst the authors concerning the basic tenets of the program:
subjectively experienced mental phenomena can be studied objectively, most
mental processing takes place unconsciously, most mental processes are modular,
mental states are token identical to brain states, and mental states are type
identical to functional states. There
also seems to be general convergence upon what is called a global neuronal
workspace model of consciousness. For
the most part, the collection comes across as a team effort with each member
performing his or her own well defined function. There are a small number of intramural disagreements, some
significant (is phenomenal experience real or illusory?), but even here there
is agreement on how to approach the problems.
introductory article, Dehaene and Naccache introduce us to the global neuronal
workspace hypothesis. The thesis brings together the idea that the mind is
modular with the idea that all mental states are brain states. The proposal is that we can isolate and
identify a broad range of mental acts, then locate the regions of the brain
that are active when we perform them.
There are distinct, though not necessarily anatomically isolated, neural
systems that are dedicated to processing different types of information. Modules perform their function
automatically, and the information in one module is not ordinarily available to
other modules. The great number of
autonomous, domain specific, and informationally encapsulated modules makes it
possible to explain a great deal of the complexity of human behavior. We do not need to be conscious of much of
what we do in order to do it.
Nevertheless, we sometimes do need to know what our minds are doing and
it certainly appears that we often are aware of what we are thinking. So the question is: what is going on in the
brain that accounts for my consciousness of a particular mental state of which
I am conscious? According to the global neuronal workspace model, the
subjective experience of being in a particular conscious state is explained in
terms of the interconnection among certain high-order mental modules,
specifically the perceptual, motor, evaluational, attentional, and long term
memory systems. That is, the contents
of a given mental module are conscious when they are accessed by these five
systems. This is what global
availability is meant to refer to. The
global workspace, is not a space at all, despite its name. The authors maintain that this account
explains why some sorts of mental processes are impervious to consciousness
those processes are not connected to the workspace. The authors also make the bold claim that the so-called hard
problems free will, qualia, sense of self, evolution of consciousness will disappear after a satisfactory
framework for understanding consciousness has been constructed (p. 29).
The second article, by Driver and
Vuilleumier, is a good representative of the empirical work being done by
cognitive neuroscientists. It relies
heavily on empirical studies done on monkeys and brain damaged humans. The general pattern of these studies is to
specify the cognitive functions an organism can and cannot perform as the
result of a lesion in a particular part of the brain. Of specific interest are those lesions that allow the organism to
continue to represent and manipulate some aspect of his environment, but
without the conscious awareness of such cognitive activity that, in humans at
least, is characteristic of such cognitive activity. What these studies are designed to show is that due to the
modularity of the mind, the organism is sometimes capable of performing certain
complex cognitive tasks in the absence of conscious awareness. In these rare cases the cerebral substrates
that enable the mind to perform these tasks are unaffected by the lesion, but
the modules connection to higher order modules (such as long-term memory,
intentional action) are damaged, thus accounting for the organisms lack of
consciousness of what he is doing.
article we are presented with the phenomena of neglect and extinction. People with a certain sort of brain damage
act as if half of their physical environment were not there. For example, they will read only one side of
the newspaper, or, if asked to sketch an object that is in front of them, they
will draw only one half of it.
(Shockingly, these individuals are typically unaware of their
deficit.) There is nothing wrong with
these individuals sensory faculties.
They are neither blind, nor incapable of processing visual
information. There remains, in fact,
much residual information processing.
Hence the neglect patient perceives the neglected information, but is
not aware that she does so; her perception is unconscious. What is illuminating is just how much a
person is capable of doing with information she doesnt know she has.
third article Kanwisher takes up the problem of finding the necessary
conditions for a mental states being conscious. She argues against the
activation strength hypothesis, i.e. that a mental state is conscious if the
neural process underlying it is powerful enough. Her focus is to show that a mental states connectivity to
workspace neurons is an additionally necessary condition for conscious
awareness of that state. (I)t seems
reasonable to hypothesize that awareness of a particular element of perceptual
information must entail not just a strong enough neural representation of that
information, but also access to that information by most of the rest of the
mind/brain (105). Kanwisher steers a
careful course in the article, avoiding the scylla of having to posit a
homunculus witness to the neural event, and the Charybdis of implying that
isolated brain processes magically become conscious in virtue of their mere occurrence.
In the fourth article, Merikle,
Smilek, and Eastwood take up the important
methodological problem of measuring perception without awareness, i.e.
unconscious perception. Unfortunately,
much of the article is concerned with demonstrating that perception without
awareness is possible a claim that is hardly controversial and that is
addressed by every other article anyhow.
The crucial methodological issue in question is whether awareness ought
to be measured objectively or subjectively.
An objective measure of awareness would count any ability to make a
discrimination as evidence for awareness of what is being discriminated. Subjective measures of awareness, by
contrast, take seriously subjects claims not to be aware of the items which,
it turns out, they can distinguish at rates better than chance. What is at stake is the definition of
consciousness (or awareness). Can
consciousness be defined in a way that ignores the subjects first-person
perspective? If we use objective
measures of consciousness, it can.
While the authors do make a solid argument in favor of prioritizing
subjective measures, they often seem to be pulling their punches against the
use of objective measures. Hence, the
authors do not seem to adequately emphasize that the philosophy underlying the
reliance on objective measures (behaviorism) is antithetical to the project of
cognitive neuroscience as conceived of by most of the other authors of this
book. For example, Dennett makes clear
that he thinks every study reported in every article in this volume has been
conducted according to the tenets of heterophenomenology (p. 231). Heterophenomenology is a method for
understanding the subjective, first-person, aspects of consciousness. Heterophenomenology, Dennett maintains,
takes subjectivity seriously by taking the reports of subjects seriously as
reports of their subjective experience (p. 230). Dehaene and Naccache also emphasize quite clearly that the aim of
cognitive neuroscience is to account for the subjective experience of
consciousness. They write: In various
daily life and psychophysical testing situations, (people) use phrases such as
I was not conscious of X, I suddenly realized that Y, or I knew that Z, therefore
I decided to do X. In other words,
they use a vocabulary of psychological attitudes such as believing, pretending,
knowing, etc., that all involve to various extents the concept of being
conscious. . . The task of cognitive
neuroscience is to identify which mental representations and, ultimately, which
brain states are associated with such reports. (p. 3). It seems as if this chapters task was to
dispatch the behaviorists notion of consciousness, but the authors seem to
stop short, contenting themselves with dismissing it as too conservative (p.
While it includes some intriguing
proposals, the fifth article, by Parvizi and Damasio, is almost exclusively of
interest to other neuroscientists doing technical research on the human
brain. A typical sentence looks like:
Both the superficial dorsal horn and the caudal spinal trigeminal subnucleus
receive primary afferents through unmyelinated C-fibers and lightly myelinated
Ad fibers which convey signals related
to pain and temperature. Unless the reader is capable of making sense of
sentences like that, he would be well advised to look elsewhere to find out
what Parvisi and Damasio are trying to do.
The sixth article, by Jack and
Shallice, is the most ambitious in the volume.
The authors set out to determine the function of consciousness. They believe that just as Watson and Crick
had to know the function of the gene before they could discover the double
helix structure of DNA, so too, must we know the function of consciousness
before we can find its physical structure in the brain. The authors begin with folk psychological
criteria for determining the function of consciousness: we are sometimes
conscious of something and we sometimes consciously perform some action (p.
165). They propose that the challenge ahead is to enumerate the sorts of
processes that require consciousness, calling such processes type-C. Type-C processes are defined as processes
that can only operate effectively on information when normal subjects report
awareness of that information (p. 170).
This definition seems arbitrary and counter-intuitive since most people
think that non-linguistic animals are conscious. What the authors are really looking for, however, are processes
that require self-reflexive awareness, or introspection. And the issue seems to
be that when we introspect, we report having experiences that have non-physical
properties; and since there are no
non-physical properties, we cannot be having the experiences we report
having. The solution, they propose, is
that the reporting just is the awareness; we do not enter into a conscious
state, witness it, then report it.
Rather the ability to make the report, somehow, is the experiencing. The authors call this the endowing view because the type-C process
is supposed to endow awareness, not just enable
an allegedly already conscious state to become reportable (172). The authors then go on to identify numerous
processes that they think are strong candidates to be type-C processes. This article has the advantage of being the
most overt in its statement of the goal of cognitive neuroscience: The closest
that science can come to accounting for subjectivity is through elucidating the
mechanisms that allow us to understand ourselves from our own point of view
(p. 190). So, the goal is not to
explain the subjective properties of consciousness, but to explain the
mechanisms of human understanding that lead us to believe that consciousness
has irreducibly subjective properties.
two articles, one by Block, the other by Dennett, take on the form of a debate
over whether and how the cognitive neuroscientific framework must leave some
aspect of conscious experience unexplained.
Block takes the position that there are three distinct types of
consciousness and that the authors of this volume are concerned, or should be
concerned, exclusively with reflexive consciousness. Cognitive neuroscience does not address phenomenality, he
argues. Global access to an otherwise
isolated mental state may explain the persons consciousness of that particular
mental state, but even without such reflexive consciousness, Block suggests, mental states have phenomenal properties:
there is something it is like to be in them.
In fact, it is the phenomenal properties that reflexive consciousness is
consciousness of. Reflexivity involves phenomenality plus more reflection on
the phenomenality (p. 213). Dennett
counters that there is no such will-of-the-wisp property of
consciousness. A conscious state is
conscious in virtue of the effects it has, if phenomenal properties have no
effect, they are a philosophers illusion.
It does seem that Dennett gets the best of the argument since Block must
resort to appealing to unconscious phenomenal properties, i.e. states that it
is like something to be in, even though we have no awareness of being in them.
On the whole this is an exceptional
book and well worth reading if you are interested in learning what cognitive
neuroscience can contribute to our understanding of mind. It is a breath of fresh air compared to the
numerous hackneyed debates over whether such a contribution is, in principle,
2002 David A. Patten
David A. Patten, Ph.D., Department
of Philosophy, SUNY Stony Brook, NY.