Customer Reviews

Bright Air, Brilliant Fire: On The Matter Of The Mind

5 star:		(10)
4 star:		(3)
3 star:		(2)
2 star:		(1)
1 star:		(0)

 

 

Although Edelman tried to make "Bright Air, Brilliant Fire : On the Matter of the Mind" a self-contained story, it really is based on his trilogy of books "Topobiology : An Introduction to Molecular Embryology", "Neural Darwinism; The Theory of Neuronal Group Selection", and "The Remembered Present : A Biological Theory of Consciousness". I am not sure that any mortal can read only "Bright Air" and really understand what Edelman is talking about.

The claim that Edelman's Theory of Neuronal Group Selection (TNGS) "does not appear to have the potential to really crack the problem" of how a brain makes a mind is a claim that is often made without any suggestion of exactly what Edelman might have missed. These claims are like people in the 1940's saying, "a rocket does not have what it takes to get to the moon." Certainly a 1940's rocket could not reach the moon, and certainly Edelman's TNGS is not a complete theory of mind, but Edelman, like von Braun, was visionary in being able to see that with future improvements, the path to the desired future was in sight. The claim that no correct materialistic theory of mind will ever be found is now nearly as impossible to defend as the claim that "men will never walk on the moon" would have been in 1965.

Speculation about why Edelman's books so annoy and infuriate his critics: 1) Edelman has constructed an new language which he uses to describe his theory mind. He provides no glossary with definitions of his terms. This alone is a horrible tactical error that can only alienate his readers. 2) Edelman builds his theory from a foundation that is unfamiliar to most of his critics.People like Crick, Dennett, and Johnson have never read the literature of "topobiology" and they are also not able to conceptualize how synapse regulation rules must be integrated into the proper types of neural networks in order to allow for learning and memory. 3) Philosophers of Mind, in particular, the many who are "Functionalists" as well as the huge swarm of Parallel Distributed Processing connectionists are shown by Edelman to be taking an inferior approach to mind. Having your professional career side-swiped by an interloper from Biolgy is enough to enrage most philosophers and AI researchers.

New species arise from subtle recombinations of mutations and their birth is a fragile process. The fundamentally correct components of Neural Edelmanism will survive the memetic selection process within the Science of Mind. In the next century Edelman will be viewed in much the same way biologists of this century now view Gregor Mendel and Charles Darwin: men who published their ideas well before science as a whole was ready.

In Bright Air, Brilliant Fire, Gerald M. Edelman accomplishes what seems to be an almost impossible task: He helps the non-scientist to understand the connections between what is known about the mind with what is beginning to be known about the brain. For Edelman, this subject "is the most important one imaginable" because it is charged "with the excitement of being on the threshold of knowing how we know." At the outset, he poses "some commonsense notions":

1. Things do not have minds.

2. Normal humans have minds; some animals act as if they do.

3. Beings with minds can refer to other beings or things; things without minds do not refer to beings or things.

The book is divided into four main parts (Problems, Origins, Proposals, and Harmonies), concluding with "Mind Without Biology: A Critical Postscript" in which Edelman dispels the notion that the mind can be understood in the absence of biology. Stated another way (in Chapter 2), "There must be ways to put the mind back into nature that are concordant with how it got there in the first place."

Obviously, this is not a book for browsers, for grasshoppers, or for dilettantes. It makes great demands on the mind (and patience) of its reader. But consider Edelman's original objective: to explore the connections between what is known about the mind with what is beginning to be known about the brain. For him, this subject is (to reiterate) "the most important one imaginable" because it is charged "with the excitement of being on the threshold of knowing how we know."

Is there any other knowledge of greater importance?

Dr. Edelman has many critics, who all sound the same in their attacks on his work. First, a sly remark about his personality, his egomania, his obsession with grandeur. Then they claim that he cannot write clearly, that he obscures with his highly technical language. And then you get the usual complaint about the lack of empirical evidence, etc., etc. finally, they claim that they, even being the experts that they are, cannot understand Edelman at all. All these critisms seem convincing enough until one reads Edelman's recent book. Yes, he is highly ambitious, attempting to construct a complex theory of consciousness. But he is a clear and direct writer, who exposes the problems at every step instead of hiding them; he is modest, generously acknowledging his debts to earlier work, providing a helpful bibliography for the interested reader. As to his theory, I find it more convincing than all the theories offered by his critics. That is not to say that it's flawless. Edelman himself, in fact, explicitly says that many aspects of the theory are in need of further revision based on empirical evidence. But his work, clearly the product of a powerful and erudite mind, seems to me the best there is in this immature theoretical field. It needs criticism, but not stupid cricism, as offered by his current critics (e.g. Crick, Dennett, Johnson, etc.), who are obviously off the mark. It would be interesting to speculate why Edelman's books so annoy and infuriate them; after all, it is just another theory, why all the sound and fury when most people in the neuroscience community haven't even read the new book, or any of the previous ones. Is it dangerous--to certain people, for certain unknown reason? Now I hope that Dr. Edelman will continue his line of work, writing more enlightening books that will gradually engage the specific problems he mentioned in his previous work.

I am a huge fan of Edelman, but I regretted having bought this book; I would say this is a kind of half-successful attempt at vulgarizing what he explained so well elsewhere: there is nothing to be found here that wasn`t already explained in more detail in "Neural Darwinism" and "The Remembered Present". So stay away from this one if you read the others. If you never read anything by him, go for his "Neural Darwinism". Reading his books was definitely a great intellectual experience of my life. Go for it!

PS: Da man is a genius.

The worst kind of reviewers are those who come when the play is half over. There is no "royal road" to the truth, either in mathematics, as the old tale has it, or in philosophy and science. Ethically, you can struggle through the basics and talk about your conclusions; or you should shut up. If you have no inkling about the Theory of Neuronal Group Selection, you have no inkling about consciousness. Go back and do your homework. So Edelman's books are tough going? Did you expect to understand calculus without taking algebra? I read "Neural Darwinism" right after it was published, and gagged. My biology background is thin; one college course. I looked up "natural selection" and found Dawkins; I read all of his books and knew something about genetics. I read "ND" again, understanding a bit more, especially the part about "ground-breaking" ... When "Topobiology" appeared, I read it. I had to reread "ND" to get through it, but I began to understand the vocabulary, the ideas, the logic, the structure. Same thing with "Remembered Present". At that point I was working in a university psychology department (strictly as a hardware specialist, never having had a single psychology course). I mentioned this exciting work to a few professors in developmental psych who might have been expected to look at it; none of them did. Not till 1999, when a visiting professor offered a course in "Consciousness" based on "The Remembered Present", was there any hint that Edelman had some relevance for those researching the development of the mind. Draw your own conclusions. If you're not afraid to know the facts, if you really want to know the answers to the hard questions (What is life? Why is the universe only 4-dimensional? Where does thought come from?), dig for them; they do exist. Start with Edelman.

Having read this book roughly twice, I'm still not sure I'm up to offering a review but that may, itself, be an important comment because clarity often reflects theoretical adequacy. The book's main thesis seems to boil down to this: Brains are a product of what Edleman calls selectional events, reflecting an operational principle specific to living systems. In this he draws a distinction at the very outset between the hit or miss outcomes of evolutionary development by random selection (which produce species differentiation by changing the individual members of species over time) and the instructional development people cause to happen via algorithms (step by step procedures for getting things done, either by describing the steps needed to get things done to one another, by following such steps, or by producing those steps in computational code which can then be translated into the digitized electronic signals we think of as 0's and 1's that govern computer operations).

Selection, of course, is serendipitous, whatever blueprint establishes each individual being the result of eons of trial and error in the selectional process that preceded it. Edelman takes this idea and applies it to how individual brains operate on a daily, and instant by instant, basis. Extending the idea of selection, as he has previously identified it in human immune systems (where antibodies in our bloodstreams selectively bond with invading agents, based on their "fit", and those which bond more closely out-produce those which don't, thus increasing the "right" type of antibodies in the bloodstream), Edleman suggests that the waxing and waning of the brain linkages (which, he claims, embody our mental processes) occurs in roughly the same way.

This is somewhat controversial because of 1) its implications (it challenges more accepted functionalist accounts which liken brain operations to algorithms such as those a computer can perform); and because of 2) its failure to fully account for how the selecting actually alters the strengths of the linkages or how the linkages themselves give us the features we associate with consciousness (things like awareness, intentionality, understanding, etc.). And yet, given that he has so thoroughly laid out his case for this kind of biological mechanism driving brains, and that he seems to have covered the obvious operational issues as they apply to what we presently know about brains, his thesis has a certain plausiblity. But it remains highly complex and even controversial because it boils down, in the end, to a number of somewhat abstruse and not intuitively apparent claims.

It's really all about memory in the end, Edelman says, though he defines this function so broadly that it's no longer recognizable as what we call memory in ourselves. Memory at its most basic level is the capacity of a self-contained, self-replicating system to retain changes and, via self-replication, to pass them on to descendent systems. Of course, "memory" also means the ability of a computational system to recover stored coded data and, in us, the ability to summon up past images, thoughts, associations, etc. Are these all the same thing? Edelman suggests they are and claims that it is the increasing systemic complexity and resultant sophistication that makes the primitive memory we find in relatively simple systems and their components (including the capacity to reproduce progeny with like "fits" as we find in immune systems) into the kinds of things we call memory in ourselves. It is this memory principle, a product of the selectionist dynamic (or perhaps the reason that selection manages to occur?) that, on his view, is the foundation of the kinds of features we recognize as consciousness in ourselves.

Alas, Edelman's thesis is anything but simple (which I'm guessing, is already apparent from my efforts here). In the end it seems to amount to the claim that brains, being organic and the result of the previously described selectional processes, are uniquely complex in structure and operation in the universe, each uniquely differentiated from all other brains (because even twins' brains, constructed on a shared blueprint, will have changed in their lifetimes in ways unique to each individual). It is this complexity of structure, this "morphology", he tells us, that is responsible for the things brains do. Consciousness, he argues, is built on levels of operation in the brain and primary or animal level consciousness (which itself has levels) forms the necessary foundation for higher level consciousness in which self-awareness and self-reflexiveness become possible. These levels are only possible because of the intrinsic complexity of brain morphology.

Edelman lays out a picture of how the different parts of the brain are responsible for different features and notes how many of the features we treat as unitary in ourselves (e.g., memory, perception, conception, etc.) may be best understood, upon reflection, to really be the result of the elaborate collaboration, through cross brain linking of neuronal groups in different brain centers responsible for different things. These combine to produce the various constituent operations that, when further joined through what he calls "reentrant loops", become the features we recognize in ourselves like awareness, understanding, remembering, intentionality, conceptualizing, believing, etc. He calls his model of how brains work, built on all of this, the Theory of Neuronal Group Selection (TNGS), because, he explains, different neuronal groups (not individual neurons) must operate together through the cross-brain linkages already mentioned to produce each instance of consciousness in ourselves. Such instances are thus complex occurrences and not the simple unified phenomena they appear to be on introspection.

Human memory, for instance, he describes as a complex phenomenon that differs from the relatively simple memory function which computers rely on. What we recognize as memory in ourselves is never exact, he points out, because it consists of a bunch of different linked constituent elements running in parallel and impinging on one another, albeit never exactly in the same way, each time a remembered item is summoned up. Unlike computational memory, which involves the exact replication of a coded instruction in each instance (a failure of this level of precision being a failure of the computational memory function), human memory is seen to be rough, approximate and ever-changing (think Roshomon or the related phenomenon of "false memory" as described by modern psychology). This feature of our memory happens because remembering in us is a function of a complex reconstruction process that depends on the relative strengths and linkages of the connecting loops between the implicated neuronal groups. As such this makes creativity and metaphor possible but it also means that each instance of recall introduces changes through new associations.

It's because of his claim that brains are uniquely complex (more like a jungle than a powerplant as he puts it) that he insists that computationalism, the thesis that consciousness can be replicated on computers, is unsound. We need brains he says, because consciousness is a biological product requiring the complex morphology and historical development, reflecting selection, that only organic systems can achieve. On this last point, though, I think he may well be mistaken.

Edelman often uses "consciousness" in a relatively unsophisticated way and seems to be under the impression that computations, as performed by computers, are purely abstract without, as John Searle (whom he cites) puts it, causal efficacy in the world. In fact, computers are as physically instantiated as brains and computational processes on computers as physical as they must be in human brains. But his answer as to whether a computer can be conscious does not stop with this claim about the abstract nature of computational processes. He makes the further point that computers, running on logic, are instructionally based whereas our brains, running on "selection", are not. In this he seems to disregard the notion that selection in evolution itself provides its own set of instructions (our genomes), even if they're not conceived in advance by any mind. Similarly he seems to overlook the obvious fact that the genetic mechanism, itself, appears to be a form of information processing just as computations in computers are (though genetic information processing may, in fact, be more complex).

Moreover, if what constitutes consciousness in us is a range of features which occur together (awareness, intentionality, understanding, perception, conception, etc.), then the real question is not whether brains and computers can accomplish the production of these features in the same way but whether they can both achieve the same functional output at all. If computers can produce the features we associate with consciousness, how they do it is likely to be less important than that they can and his argument that computers can never be expected to do what brains do would be wrong.

Nevertheless, he has offered a very compelling picture of the incredible complexity that confronts us (especially if you're a computationalist) in endeavoring to achieve the goal of creating synthetic minds. For Edelman, everything hinges on his argument for uniqueness and complexity, which he derives from the principle of selection that, he tells us, is unique to self-contained, self-perpetuating systems like living things. Computers, on the other hand, are pre-set operational functions ("syntactical" systems, again echoing another of Searle's famous criticisms) which are guided by an outside mind (the programmer) and therefore unable to react to an open-ended world of unpredictable and ever changing inputs. But in claiming this, he presumes that computers lack the ability to be open-ended because they are programmed though this is by no means clear -- or likely to be any truer than his related claim that consciously programmed computational machines are qualitatively different from organic machines that have evolved over eons as a result of a selectional mechanism. After all, does it really matter where the programming comes from, as long as there's programming there?

Edelman's underlying position is that the genetic programming of evolutionary selection leads to an open-ended tangle of unpredicted, and unpredictable, synaptic connections -- and their more macro neuronal group links which strengthen and weaken according to their ongoing interaction with external stimuli. It is this waxing and waning of the neuronal group pathways between the different brain centers and their sub-parts that gives rise to the facets of consciousness on his view -- an activity that is nothing like the computational processing of machines as found on computers.

Yet, in this he seems to be disregarding the fact that computational programs can also involve multiple sub-systems doing different, interrelated things, and that such systems can be designed to implement open-ended reactions to an unpredicted world of stimuli once basic parameters are set. Such parameters need be no more obtrusive than our own evolution-driven genetic codes, after all.

Edelman's thesis is worth thinking about and may well offer some interesting directions for understanding how brains work and for developing new thinking machines that may depart from the standard computational model we rely on today (the universal Turing machine). Unfortunately Edelman's writing is not up to the level of the ideas presented (on my view). It may be the case that the complexity of Edelman's thesis would not admit of any more succinct or precisely written text than he has provided but, in spite of my own difficulties here in explicating his ideas, I rather think he could have (and should have) been clearer. Certainly his failure to be clear is at least a prima facie reason to question the potency of the underlying ideas.

SWM

More than a quarter of a century ago, Edelman was distinguished with a Nobel Price for his research on the immunity system. Later he switched to neurology. He sees a parallel between the way the immunity system is programmed and the way the brain is programmed. These procedures have much in common with Darwins process of evolution by way of natural selection. Edelman speaks of "neural Darwinism".

When you just think about it, after reading the arguments Edelman brings forward, you will see that evolution could hardly have produced something as complex as the human brain (or even more simple animal brains) by any other means. A lot has to be investigated yet, of course, but I think Edelman has shown the way to a deeper understanding of our brain.

It is a pity Edelman and Daniel Dennet get along so badly. Edelman never mentions Dennet, and Dennet is extremely critical in the few remarks he makes about Edelmans work. I think their approaches are complementary, not contradictory.

Yes, the human brain is a computer of sorts. Edelman has the clearest ideas about the structure of this computer, but he denies that the metaphor of the computer is valid. I think that, even taking Edelmans ideas about the deeper structure in account, the metaphor remains valid, up to a point at least. Really, Edelman has much more in common with Dennet (whom he seems to despise) than with the "mysterian" Searle, whom he praises.

Edelman's detailed reasoning may be beyond the abilities of a layperson to follow, but his concepts are clear and logically sequenced. Single cells communicate internally by diffusion, multi-celled organisms need specialist linking cells from one remote part to another because diffusion through successsive cell walls is too slow. Hence the beginning of a neural net. Evolutionary pressure gives a benefit from storing records of responses to stimuli in a specialist area of this net, access in due course to these records give Primary Consciousness. As a consequence of homo sapiens' upright gait, a unique construction over the windpipe to prevent food falling into the lungs coincidentally allows a far greater range of noises to be made in response to stimuli - enough different noises for a very fine degree of specialisation. Not merely "woof woof" = danger but "tree falling" = specific danger. A data store of the words associated with the responses permits them to be called up separately from the occurrences that crteated them - hence an ability to imagine and plan with Higher Consciousness. An interesting idea associated with this developmental scheme is that many animals (up to, say, lobsters) may have no consciousness at all and are merely machines. Better minds than mine may say that Edelman's details do not stand up to rigorous examination, but the concepts form the most persuasive explanation of consciousness that this reader has come across.

Wow. Obviously we've emerged from the dark ages into the light and....wait, Plato already made this analogy. Anyway, its always edifying to see people, anyone, confirm in print what we all already know but are afraid, for political reasons I guess, to utter. It's a material world. Edelman takes off on this theme and proceeds to deliver the underlying neural blueprints for ego, soul, cognition, and a host of other mystical magical metaphysics, minus the meta. Edelman makes some gratuitous forays into philosophy in this book, but by and large sticks to the point and discusses his work on axonal feedback/forward pathways and TNGS; the theory of neuronal group selection, an evolutionary take on neurobiology. It's always nice, in my view, when researchers deign to wtite these pop-sci books to share with us plebes what we could never encounter in our own pedestrian lives. I can't really evaluate if this is a "good" piece of literature, but anyone remotely interested in who and what they and the people they know are simply must devour evey morsel of this neurophysiology primer.

I didn't know Edelman and his books till last year.I had to write a thesis for my degree and i only was looking for ideas related to IA and cognitive processes. I first read "The Remembered Present" and i felt entusiastic!!! Then i studied carefully his other books (Neuronal darwinism and this one); seldom it happens to find such a deep thought and clarity. Its a must for all people searching new kind of theories , both in philisophics both in neuroscience field.