The title of Edelman’s book comes
from a poem by Emily Dickenson that celebrates the brain ("The brain €“ is
wider than the Sky €“ "), from which the reader who is new to Edelman’s
work will correctly infer that he is not a Cartesian dualist. The Preface
relates that his books and articles on consciousness over the past twenty-five
years prompt him "to present an account of consciousness to the general
reader" (xi). He adopts James’s distinction between science and
metaphysics and undertakes "to avoid extensive discussion of metaphysical
matters, [hoping] to disenthrall those who believe the subject is exclusively
metaphysical or necessarily mysterious" (xiii). In Chapter 1, "The
Mind of Man: Completing Darwin’s Program," he undertakes to establish the
causal efficacy of consciousness and "to show how a neural mechanism
entails a subjective conscious state, or quale, as it is called" (3). In
Chapter 2, "Consciousness: The Remembered Present," he salutes what
he calls the Jamesian properties of consciousness, including James’s point in
"Does Consciousness Exist? " that consciousness is a process rather
than a thing, but also the points James makes in the famous chapter "The
Stream of Thought" in The Principles of Psychology, about privacy,
continuity, "aboutness", selectivity, having a "fringe,"
and so forth. He introduces some technical terms. Primary consciousness
"is the state of being mentally aware of things in the world, of having
mental images in the present," whereas higher-order consciousness
"involves the ability to be conscious of being conscious, and it allows
the recognition by a thinking subject of his or her own acts and affectiuons"
(9). He introduces some dramatic tension by announcing that qualia are to be
understood as "high-order discriminations that constitute
consciousness". Does Edelman mean that qualia are constitutive of
consciousness in the high-order sense but not the primary sense?  Chapter 3,
"Elements of the Brain" is an engaging Brain Science for Dummies
exercise that includes such colorful thoughts as

€¢    "If the cerebral cortex were unfolded
(making the gyri, its protrusions, and the sulci, its clefts, disappear, it
would have the size and thickness of a large table napkin." (15)

€¢    "[The
thalamus] is only somewhat larger than the last bone in your own thumb."
(19)

€¢    "The
hippocampus [is] lined up like a pair of curled sausages [and] each sausage
looks like a sea horse, hence the name ‘hippocampus’." (21)

€¢    Inability
to convert "short-term memory of events into a permanent narrative record
[is] a condition that was depicted dramatically in the movie Memento."
(23)

The chapter ends with the argument that it "would seem
unlikely" that the brain is a computer because the development of the
brain quickly becomes epigenetic or not-hardwired, resulting in "highly
individual networks in each animal" (29). He concludes, "This is no
way to build a computer, which must execute input algorithms or effective
procedures according to a precise prearranged program and with no error in
wiring" (29). He does not reveal in this chapter what he understands a
computer to be, though his remark about building one suggests he has in mind
something like a desktop PC. But proponents of the view that the brain is a
computer €“ cognitivists in John Searle’s terminology, or language-of-thought
theorists in Jerry Fodor’s €“ have in mind a much more abstract conception of a
computer, one that draws on Alan Turing’s work in 1930s: what we call a
universal Turing machine. The prospect that the brain is a computer in that
sense is not undercut by Brain Science for Dummies. For consider:
Edelman would presumably agree that for every brain event there is a cause, and
that these causal relationships instantiate a causal law. But such law-governed
processes will be algorithmic in the sense that’s relevant to describing the
brain as a Turing machine. He will be entitled to his point that each of these
machines will be importantly different because of epigenesis, and that any such
machine would be a bad blueprint for building a desktop PC. But brains might
still be them.

Chapter 4, "Neural Darwinism:
A global brain theory," begins by stipulating that "neural Darwinism
or the theory of neuronal group selection" is incompatible with a model of
the brain as a computer or Turing machine (33). (If the argument above is
correct, that is not true.) Neural Darwinism accepts that neural variability as
fundamental and that the population of variants in each brain exhibits a
pattern because of some constraint of value or fitness (35; and see 25 for
technical notion of value, a value system being a pattern of activity related
to rewards and responses necessary for survival). He also continues his attack
on the computer model, dismissing as unlikely "an a priori program"
(37). But the program hypothesized by AI theorists is not a priori (whatever
Edelman means by that), but is brought into being and molded by evolutionary
forces. This suggests that neural Darwinism and the computer model should not
be regarded as incompatible, but rather as, respectively, more and less
specific descriptions of the brain. He also refers skeptically to the computer
model’s "fixed template" or "predetermined algorithms"
(38), but the algorithms need be predetermined only in the sense that there are
causal laws governing the brain, and if a fixed template is something like a
PC’s silicon chip, nothing like that is presupposed in the Turing-machine
conception of a computer. He objects that a computer uses formal rules governed
by explicit, unambiguous instructions, but that is no bar to the computer’s
generating the sort of neural selection that Edelman’s Darwinism predicates.
Connectionist systems have comparable complexity, but those systems are virtual
machines running on computers with formal rules. Edelman also introduces in
this chapter the key idea of reentry, "the ongoing recursive interchange
of parallel signals among brain areas, which serves to coordinate the
activities of different brain areas in space and time" (41). He is at
pains to show that reentry is not feedback, "not a sequential transmission
of an error signal in a simple loop," but reentry looks very much like a
complex (nonsimple) system of the sort that systems theory describes, hence a
kind of computer, one with a PDP (parallel distributed processing)
architecture. Chapter 5, "The Mechanism of Consciousness," declares
that the sufficient condition for consciousness, is the following
evolutionary event: the emergence of massively reentrant, reciprocal
connectivity in the thalomocortical system (54), an event which dynamically
linked value-category memory to perceptual categorization. This raises the
mind-body analog of Hume’s famous question about getting from IS to OUGHT: how
do you get from BRAIN PROCESSES to QUALIA?  Chapter 6, "Wider than the
Sky: Qualia, Unity, and Complexity" makes the attempt, but when he
asserts, "Qualia are these discriminations" (high-order
discriminations enabled by activity in the thalamocortical system), this seems
like saying, on the basis of a particulary thorough police investigation,
"The butler ought not to have done that." The police
investigation begs Hume’s question, and arguably neural Darwinism begs the qualia
freak’s question. Chapter 7, "Consciousness and Causation: the phenomenal
transform," addresses the other question Edelman set himself, about the
efficacy of consciousness. He asserts that consciousness C is
"entailed" by C’, his neural Darwinist account of brain processes,
but evidently he is not clear about the difference between logical entailment,
as in "It’s raining" entails "It’s not not raining", and
the looser relationship in which accomplishing some goal, for example, would
entail a lot of work. The latter gestures at the physical necessity of working
in order to achieve the goal, and is compatible with the logical possibility
(magic, say) of achieving the goal without working; there is no logical
entailment. By the same token, zombies who have C’ without C are logically
possible, even if neural Darwinists are right in thinking that the invariant
association of C’ and C is physically necessary; there is no logical
entailment. I don’t see that Edelman has established the efficacy of
consciousness. In fact he seems to admit that he hasn’t, in passages like this
one. Talking about people like us, he writes

But such individuals can exchange
information even on the basis of the mistaken belief that their C states are
causal. The belief is safe, even if scientifically false, for evolution has set
up reentrant circuits to yield C states as properties of C’ states. (137)

I don’t see how this amounts to anything over than a form of
epiphenomenalism: belief in the causal efficacy of consciousness is safe but
false. Edelman may be leaning very hard on his statement that qualia
"are" high-order states of the thalomocortical system, turning it
into assertion of identity, but he disavows that approach (125) in favor of
understanding qualia as properties of that system.

In short, Edelman’s little book
fails to do what it sets out to do, and it is of little use to the general
reader, unless that reader wants a sketchy tour de horizon of Edelman’s oeuvre.

 

© 2004 Wesley Cooper