Trusting the Subject? Volume 1
(edited by Anthony Jack and Andreas Roepstorff) is an interesting and timely
anthology.  Indeed, given that recent decades have seen such a robust
re-emergence of interest in the study of consciousness, few topics are more pressing
than the appropriate use of introspection in consciousness research.  Utilizing
a dozen articles from clinicians and philosophers, this book makes an important
contribution to the endeavor. 

The first article, by K. Anders
Ericsson, takes, as its starting point, the fierce controversies over
introspection from the early part of the 20^th century; only by
understanding these debates, it is argued, will methods be improved such that
they meet the standards of reproducible scientific data.  Given the lessons
drawn from these debates Ericsson presents an account of verbal reports
(developed in conjunction with Herb Simon) that, it is argued, provides useful
and necessary scientific data.  In the second article, Daniel C. Dennett
presents a defense of "heterophenomenology," his explicitly
third-person approach to human consciousness.  Dennett defends heterophenomenology
against a number of objections; we return to one of those objections below. 
The third article, by Antoine Lutz and Evan Thompson, presents a research
program for the neuroscience of consciousness; Lutz and Thomson (following
Varela) refer to this program as "neurophenomenology."  Utilizing
first-person data as a heuristic in the description and quantification of the neurodynamics
of consciousness, Thompson and Lutz address a number of methodological
challenges and pursue an embodied and dynamical approach to the
study of the neurophysiology of consciousness.  In the fourth article, Dan Zahavi
and Josef Parnas take up the issue of introspective methodologies and
phenomenology in the context of infantile autism.  What is the capacity of autists
to understand their own minds?  While the answer to this question raises
potential problems for the "theory-theory" account of autism, Zahavi
and Parnas argue that it also shows the potential value of introspection, and
phenomenology more generally, in the cognitive sciences.  The fifth article, by
Patrick Haggard and Helen Johnson, addresses the issue of the phenomenology of
voluntary action.  Reviewing some historical attempts to deal with the
phenomenology of action, Haggard and Johnson contend that the neurophysiological
processes of voluntary action do posses distinctive experiential features.  In
an attempt to characterize the phenomenology of action, the authors conclude
that the study of action may be a more promising avenue of investigation than
perception in the study of what they call "primary" and "secondary"
consciousness.  In the sixth article, Shaun Gallagher addresses the question of
how phenomenology can contribute to the experimental cognitive neurosciences. 
Gallagher distinguishes three approaches: neurophenomenology, retrospective and
indirect phenomenology (e.g., heterophenomenology), and "front-loaded"
phenomenology  (which is the approach he favors).  Gallagher concludes that
where introspective reports are required, the approach of front-loaded
phenomenology follows neurophenomenological rather than heterophenomological
procedures.  In the seventh article, Bernard Baars considers a number of areas
in which conscious experience converges with observations of neurophysiological
activity.  Consider, for example, the positive correlations between brain
activation and metabolism levels and consciousness, or the various brain
correlates of conscious mental states such as imagery, inner speech, and
volition.  To understand this great convergence of phenomena, experiential
(introspective) reports are indispensable.  In the eighth article, David
A. Leopold, Alexander Maier, and Nikos K. Logothetis utilize animal studies in
an attempt to understand how brain activity leads to subjective visual
percepts.  To do this it is necessary to ensure that the responses of the
animal subjects accurately and reliably reflect what is perceived.  With this
in mind, the authors discuss a number of approaches meant to optimize the
reliability of the animal’s behavioral responses.  In the ninth article,
Timothy D. Wilson contends that the most fruitful use of introspection requires
distinguishing between those mental phenomena which are accessible via
introspection (conscious sensations, emotions, and thoughts) and those which
are not accessible but which are vital to human survival, what Wilson refers to
as the "adaptive unconscious."  Researchers, therefore, must get
clear on precisely which mental phenomena are accessible to the subject, and
which are not.  The tenth article, by Gualtiero Piccinini, proffers a
commonsensical account of introspection.  While we should not blindly accept a
person’s introspective reports, neither should we simply dismiss them as
entirely uninformative.  Instead, we should endeavor to understand our commonsense
treatment of introspective reports.  By answering the question, ‘how do we
decide what to infer from the introspective reports of everyday life?’ we are
in a position to develop a more methodologically precise account of
introspection.  Richard E. Cytowic provides the penultimate chapter of this
book.  Cytowic contends that while first-person reports can often be
problematic, the investigators own theory-laden assumptions about how to
interpret the results of introspective reports can also bias conclusions. 
Hence, not only must great care be taken in training subjects, but
experimenters must also be cognizant of their own biases.   Cytowic concludes
that while we should trust the subject, introspective reports should not be
taken literally, but instead should be carefully interpreted and revised in the
light of the assumed biases of both the subject and the investigator.  The final
contribution to this book, by Anthony J. Marcel, addresses the question of just
how transparent and accurate our access to our own conscious experiences is. 
According to Marcel, one should be skeptical not only of the veracity of the
introspective reports of others, but even of one’s knowledge of one’s own
conscious experiences.   Nevertheless, Marcel closes by offering four "coping
strategies" for dealing with the problems of trying to accurately capture
the contents of conscious experience.

Before closing, I wish to say a
little more about Dennett’s third-person approach to consciousness, heterophenomenology.
Here, as elsewhere, Dennett contends that all the ground of
consciousness can be capture by a straightforward and conservative extension of
objective science.  According to Dennett, the only thing, besides unproblematic
scientific phenomena (neurons, electrons, etc.) that the heterophenomenological
methodology commits us to is the existence of beliefs (which, from the
perspective of the Dennett’s intentional stance are abstracta, "theorists’
fictions" (p. 20, his emphasis)).  Of course, such a position leads to
the immediate objection (Levine, Joseph, ‘Out of the closet: A qualophile
confronts qualophobia’, Philosophical Topics, 22, pp. 107-26, 1994) that
Dennett has eliminated the experiences themselves, the very
phenomena to be studied via introspection.  Rather than treating conscious
experiences as had by a subject as the primary data in question, the heterophenomenologist
takes the verbal utterances of the subject as the primary data; these data can
carry us, through interpretation, to the subject’s beliefs about the
experiences in question.  Dennett demurs going all the way to the experiences
themselves, instead stopping at the beliefs about the experiences.  To go all
the way to the conscious experiences themselves "is not a good idea"
(p. 21).  Why must we stop with beliefs?  According to Dennett, "if you
have conscious experiences you don’t believe you have . . . those extra
conscious experiences are just as inaccessible to you as to the external
observer.  So Levine’s proposed alternative garners you no more useable data
than heterophenomenology does" (p. 21, his emphasis).  But many
contemporary theorists might agree that for a state to be conscious one must
have a belief about it.  However, it is far from clear that a higher-order
thought (HOT) theory of consciousness prevents us from going "all the way
to the conscious experiences themselves."  In other words, the eliminativist
tendencies of Dennett’s heterophenomenology do not seem to follow from the
dependence of conscious experience on beliefs.  And in any case, to say that
one could have an occurrent conscious experience that one does not have any
beliefs about is not to say, as Dennett seems to imply, that said
experience is (cognitively or epistemically) inaccessible, but only that
it is not, presently, being accessed by the subject.  We do have a wide variety
of conscious experiences and we are able to form various beliefs about them. 
And while such beliefs are far from indubitable, the great promise of
introspective methodologies is that they will provide us with a form of access
to the experiences themselves, not just to our experiential beliefs as
expressed in speech acts.

In closing, Trusting the
Subject? Volume 1 is a very useful and timely collection of articles from a
wide array of sources on what all would agree is a very important topic.  For
anyone interested in the scientific study of consciousness, Trusting the
Subject? is worthwhile, even essential, reading.

© 2004 Liam Dempsey

Liam Dempsey, Assistant
Professor, Department of Philosophy, Dalhousie University, Halifax, Nova Scotia
CA

Categories: Philosophical, Psychology