Learning to Work Together:
Review of Consciousness In Philosophy And Cognitive Neuroscience Antti Revonsuo & Matti Kamppinen (Eds.)
W.F.G. Haselager
Theoretical Psychology
Vrije Universiteit
Amsterdam
The Netherlands
also at
Cognitive Science Department
Katholieke Universiteit Nijmegen
The Netherlands
wfg.haselager@psy.vu.nl
Copyright (c) W.F.G. Haselager 1998
PSYCHE, 4(9), June 1998
http://psyche.cs.monash.edu.au/v4/psyche-4-09-haselager.html
KEYWORDS: consciousness, philosophy, neuropsychology, neurobiology.
REVIEW OF: Antti Revonsuo & Matti Kamppinen (Eds.). (1994). Consciousness in philosophy and cognitive neuroscience.
Hillsdale: Lawrence Erlbaum, xii + 298pp. ISBN: 0-8058-1509-0.
Price: $US59.95 hbk.
This book is based on papers presented at the International
Symposium of Consciousness held at the University of Turku, Finland,
in 1992. As the editors claim, the explicit aim of the conference
was to relate philosophical views with empirical findings by
encouraging the dialogue between philosophers and cognitive
scientists. The editors and participants (Bernard Baars, Patricia
Smith Churchland, Daniel Dennett, John Haugeland, Vilayanur
Ramachandran, Antti Revonsuo, John Searle, and Andrew Young, among
others) have done an admirable job in presenting more than a loose
collection of essays. In contrast to many edited volumes based on
conference proceedings, the authors actually do address positions
defended by other contributors, thus contributing to the coherence
of the book. Books of this kind may not, in general, be read first
page to last, but this one contains valuable material to choose
from.
The book is composed of four thematic parts, dealing with the
ontological status of consciousness, its problematic aspects, its
models, and its future. Each part is preceded by a short
introduction by Revonsuo, outlining how the papers relate to the
specific theme. The general introduction by Revonsuo and Sajama
provides a brief review of the history of the problem of
consciousness in philosophy and they argue that the issues of
subjectivity, qualia and intentionality have become the main
focusing points of the current debate. Neuroscientific research
(e.g., direct excitation of the cortex, split-brain studies,
blindsight and the timing of conscious experience) is taken to
present empirical findings that are of direct relevance to
philosophical theses. Yet, the authors are not overly optimistic.
The results themselves are often in need of further explanation, and
although there are converging lines of research, there is as of yet
no agreement on what the study of consciousness actually amounts to.
As they say, 'nobody is going to tell you the orthodox view, because
there is no such thing' (p. 21).
Niiniluoto argues in the first chapter that scientific realism is a
valuable position to consider because it has played an important
role in rehabilitating the empirical study of human consciousness by
holding, in contrast with positivism and instrumentalism, that
science can disclose the true nature of unobservable mental
events. Echoing a point made in the general introduction, Niiniluoto
claims that given the current multitude of rival theories, science
still cannot decide metaphysical controversies surrounding
consciousness, for lack of a serious unified theory. To me,
Niiniluoto's dismissal of instrumentalism is a first indication of
an underlying theme that seems to pervade much of the book:
Dennett's position on consciousness is a recurrent topic of attack.
Interestingly, in the second chapter, Dennett tries to defend
himself against the accusation that his theory is a verificationist
or eliminativist denial of consciousness. He claims not to be
denying the reality of consciousness at all, but just to be pointing
out that consciousness just isn't like what many people
(philosophers, neuroscientists) think it is. The reality of
consciousness consists in the 'cerebral celebrity' of distributed
micro-processes in response to whatever is given as input. There is
nothing over and above these micro-processes other than their
leaving traces for further processing. Having effect, winning the
cerebral competition, is all there is to becoming part of
consciousness. No additional semi-mysterious properties (to be
discovered by neuroscience) are required, as far as Dennett is
concerned.
Patricia Smith Churchland and Vilayanur Ramachandran attack a
central element of Dennett's position, namely his suggestion that in
the case of 'filling in' phenomena, the brain just represents that
there is more of the same. They provide many instances of filling in
phenomena, related to blind spots and cortical and artificial
scotomata, to show that instead of merely ignoring the gap, as
Dennett would have it, the brain actually provides something.
Dennett's claim is shown to be incorrect by cases where the filling
in is quite contrary to the overall pattern of the presented image.
In conclusion, Churchland and Ramachandran argue that Dennett too
readily accepts a thesis that is contrary to neuroscientific data
because of his behaviorist background.
In chapter 4, another opponent of Dennett, Searle, provides his own
view on the nature of consciousness, its relation to the brain and
its main features needing explanation, which are, as he says he
would like to think, 'just a form of common sense' (p. 101).
Well-known by now is his view that the Turing test disposes one to
make the two most common mistakes at the same time, namely to assume
that consciousness can be analyzed behavioristically and
computationally. Behaviorism is wrong because there is no necessary
connection between inner states and behavior, so even if the
behavior of a system is convincing, it need not be conscious.
Computationalism cannot properly deal with consciousness because
computation is an observer-relative notion. It depends on an
observer who imposes a computational interpretation on some
phenomenon. Since natural sciences deal with intrinsic features of
the world (one of which is consciousness), computational properties
are just not the right kind of concept with which to approach
consciousness. Both 'mistakes' seem to be made by Dennett, but
unfortunately Searle here refrains from discussing Dennett's views
at all. At this point, given the aim of the book, one would like to
see at least an indication of how empirical data could help to
decide between two such completely opposing philosophical views. To
his credit, Dennett does relate his views to models and data,
thereby at least creating the possibility of being empirically
refuted. Searle is content by calling his own point 'obvious' (p.
103).
The second part of the book examines some problematic properties of
consciousness. Haugeland opens with a significant article on
intentionality. He agrees with Searle that intentionality is
intrinsic and subjective rather than observer-relative and objective
a la Dennett. Yet, he tries to find some common ground by noting
that both Dennett and Searle recognize the normativity of
intentionality (by appealing to rationality (Dennett) or
satisfaction conditions (Searle)). The question then is how a
naturally evolved physical brain can be intrinsically normative.
Haugeland argues that it is the commitment of a subject to the
standards that operate within a domain that constitutes the
intentionality of his or her states. This suggestion is, Haugeland
claims, Dennettian in the sense that it involves taking a stance,
though not from the outside, but by the subject itself. By taking a
committed stance towards domain-regulating principles intentionality
is constituted. Haugeland's suggestion is at the same time Searlian
in that intentionality is seen to be intrinsic to the subject
itself. Yet, Haugeland argues that animals do not have the intrinsic
intentionality that Searle attributes to them since they do not
submit themselves to any norms or constitutive principles on their
own responsibility. In that respect, animals are on par with
computers in having, as Haugeland calls it, ersatz intentionality.
There is, to be sure, a lot in this paper that asks for more
explication. An important task, I think, would be to explain the
ability to commit oneself without invoking normative notions. Yet,
the interesting promise contained in Haugeland's article is that
there is a way of characterizing intrinsic intentionality in terms
of something else that is more susceptible to analysis.
Dennett, in chapter 6, concentrates on another problematic aspect of
consciousness: qualia. He argues that the perception of color is a
discriminative internal state of the brain that underlies various
behavioral dispositions and that that is all. These internal
discriminative states do not also have some special intrinsic
properties. In short, though there sure seem to be qualia, there are
not any genuinely qualitative properties. Dennett illustrates his
suggestion by describing a color-detecting robot that can compare
colors internally by relating representations of color, instead of
using a colored representation. The way the robot would perform its
task is not in any significant way different from the way we do it,
Dennett claims.
Here we see Dennett making a point that was strongly attacked by
Churchland and Ramachandran in chapter 3. Obviously, they granted,
there are no colored representations in the brain. However, in the
case of blindsight, the brain actually provides representations of a
nature specific to the perception of visual phenomena, instead of
merely representing that there is more of the same. Speaking about
the special visual nature of certain representations does not entail
that the brain produces actual images or colors to be perceived by
an internal observer. But it does suggest that there is more going
on than merely representing a color by a label such as 'green' or
'11011'. What this 'more' actually amounts to, and how it is related
to the emergence of a subjective experience, is quite unclear, but
Dennett has not succeeded in showing that the attempt to find an
empirical answer to such a question is essentially misguided.
Baars, opening the third part of the book dealing with models of
consciousness, describes his own cognitive model in chapter 7. By
comparing characteristics of conscious and non-conscious processes,
Baars concludes that the basic question is how a serial, integrated
and limited stream of consciousness can emerge from a nervous system
that is mostly unconscious, distributed, parallel and of enormous
capacity (p. 154). His answer is a global workspace theory.
Specialist processes broadcast their results to the global
workspace, and compete to access information displayed in it. He
suggests that his model can explain many of the functional
characteristics of consciousness (e.g., its relation to voluntary
action, problem solving, etc.). Baars (p. 162) reports that Searle,
in discussion, asked whether his model was truly about
consciousness. To Baars the answer is yes, although he also claims
that his aim is not to answer the Nagelian question of 'what it is
like', but to understand the role of consciousness from a
third-person perspective. A global workspace is necessary, though
not sufficient, for consciousness. Searle's response is not
reported, though it would be of considerable interest. If he found
the answer convincing, then at least implicitly Searle would be
acknowledging that computationally inspired models such as the
global workspace theory might help to understand how consciousness
arises in the brain. If he rejected the answer, then one might point
out that Baars (in this chapter as well as in chapter 9) does
attempt to specify neurobiological mechanisms that Searle himself
supposes (p. 96) must be there in order for consciousness to arise.
Young, in chapter 8, goes into the neuropsychology of awareness in
great detail. He makes clear at the outset that awareness comes in
different kinds. After presenting data in the areas of visual field
defects, achromatopsia, prosopagnosia, amnesia and anosagnosia, his
general conclusion is that these deficits show that different
aspects of awareness are lost in each specific case, and that there
is no general perturbation of consciousness. Because of this, Young
echoes a theme from Dennett in wondering "whether the subjective
unity of conscious phenomena is not largely illusory" (p. 195).
Indeed, Young warns against theories that treat consciousness as a
single homogeneous entity. As Revonsuo notes in the introduction to
this part of the book (p. 146), this puts Young in opposition to
Baars' global workspace theory, but although one may speculate about
discussions concerning this issue during the conference, there are
no reports of it here.
The final part of the book addresses the future of consciousness
research. Baars and Newman offer a neurobiological interpretation of
the global workspace theory outlined in chapter 7. Especially, the
suggestions concerning competition for access to the global
workspace and the widespread dissemination of conscious information
(broadcasting) can employ the expanding neurophysiological evidence.
Neural structures associated with conscious wakefulness (the
brainstem reticular formation and the nucleus reticularis) have an
insufficient bandwidth to carry the information for a conscious full
visual scene. Thus competition between potentially conscious
perceptual contents for access to these neural structures seems
likely. With respect to the broadcasting aspect of consciousness,
Baars and Newman point, among other things, at the fact that the
performance of novel tasks and the orienting response both require
conscious attention and both involve widespread activity in the
brain.
Tuomela addresses something of a side-issue with respect to the main
theme of the book. He reviews the fate of folk psychology on the
basis of a consideration of consciousness and argues for
'correctivism' instead of eliminativism. He discusses aspects of the
traditional mind-body problem as related to self-awareness, qualia,
intelligence and personality, and claims that one can be
ontologically monist about these phenomena while continuing to use
irreducible psychological predicates. He systematically examines
twelve arguments for and against the elimination of folk psychology,
which in itself is a welcome clarifying contribution to the
eliminativist debate, and argues that none of these arguments is
fully decisive alone, but together make the case for either
eliminativism or full-blown anti-eliminativism rather weak. The
option of correctivism remains alive and he attempts to strengthen
it further by his 'thesis of correction': as long as human beings
are viewed as thinking, feeling and acting persons, elimination is
only possible to the extent that the topic is not changed. Though
there may be a substantial change in the basic conceptual
categories, the main topic of empirical investigation as given above
need, by conceptual necessity, remain constant (p. 243).
Revonsuo deserves credit for his attempt, in the final chapter, to
systematize and interrelate many of the issues and views regarding
consciousness as discussed in the book. His starting point is the
claim, made earlier in the introduction, that current theorizing
about consciousness lacks a common core and he proposes a
distinction between a philosophical and an empirical aspect of the
problem of consciousness.
Philosophically, the problem is whether consciousness ultimately is
a property of a neurobiological, functional, computational or
completely different kind. Empirically, an important issue is how
the brain solves the binding problem. By focusing on the binding
problem, Revonsuo is uniting with Young in asking how the phenomenal
unity of consciousness is possible, given the divergence at the
neurophysiological level.
Armed with this distinction, Revonsuo reviews five models of
consciousness. Dennett's multiple drafts model gets criticized for
being 'eccentric' in its answers to the ontological and binding
problem. Consciousness is turned into a mere observer-relative
label, whereas the binding problem is pushed away as originating in
a mere seeming. In this way, Revonsuo concludes, the multiple drafts
model distorts the meaning of consciousness and can therefore not
provide the basis for a proper science of it. Dennett, of course,
could simply repeat his defence, given in chapter 2, that he is
being a realist about consciousness and is just stating that
consciousness is not like what many people think it is. Yet, it
seems to me, by treating the subjective experienced quality of
qualia as mere seeming, Dennett is denying a central
characteristic of the phenomenon under investigation. As long as
empirical scientists see ways of investigating consciousness as
traditionally conceived, Dennett's position is likely to be
dismissed in the way of Revonsuo.
Jackendoff's computational model of consciousness leads, according
to Revonsuo, to no satisfactory results. The underlying problem of
all computational approaches, Revonsuo argues, is that computations
play no causal or explanatory role in the system but is just a
formal description of the system's operation. Agreeing with Searle
that the notion of computation denotes nothing in nature itself but
implies reference to an external interpreter, Revonsuo rejects the
notion of computation as just not suited as a basis for theories of
consciousness.
The remaining three models of consciousness Revonsuo deems to be
more promising. The neuropsychological model is based on
brain-injured patients having unconscious but not explicit knowledge
of stimuli (among others, as discussed by Young in chapter 8). This
model is taken by Revonsuo to be in agreement with the cognitive
model, outlined by Baars in two earlier chapters, in assuming one
common system that underlies all conscious phenomena. Finally, the
neurobiological model (a label covering a variety of theories of,
among others, Damasio, Crick, and Edelman) is of great interest
because it pays close attention to the borderline between macro and
micro levels of organization. Basically the binding problem is
solved by means of the suggestion of a 'spotlight mechanism' based
on time-locked neural activation or frequency-locked oscillations. A
combination of these three models with a realist ontology is,
according to Revonsuo, the way to go in the future study of
consciousness.
This conclusion to the book may still be a long way from
establishing the orthodoxy that, in the general introduction, was
judged to be missing. Yet, all in all, one can safely say that the
book presents a clear illustration of the current state of art in
the interdisciplinary debate about consciousness. It reveals the
stark contrast between certain positions, the intricacies of the
discussion, the many loose ends and unsolved puzzles, but it also
provides vivid instances of the growing inter-relation between
philosophical views and empirical science.
