Daniel Dennett: "From Bacteria to Bach and Back: The Evolution of Minds"

This talk introduces a recent book by Dennett, rehashing and updating his theories on the mind. I have the hard copy of the book on order. These comments refer to his summary of it delivered as part of the Google Talks series.

It is much to Dennett's credit that, as he goes along, his ideas become more precise and articulate. His examples are more apt, his analogies closer to the "essence". He never seems to say, "I was totally wrong about this", but he does make progress - partly by not taking himself too seriously. It is much to Dennett's credit that he is rarely caught in an "argument" -- he sees an element of truth in almost everything, always putting his own ideas on the table without allowing opposition to get personal. This is good because he gets a lot of opposition.

"Consciousness Explained" (1991) was a serious over-reach. Rather a mess that didn't come close to delivering on the promise in its title.

Yet, in the intervening 25 years or so, Dennett has wrestled with the ideas in "Consciousness" and made them almost work. There are many signs of progress ...

COMPUTERS

His facile understanding (1971) of the way computers work has been replaced by a deeper insight -- especially contrasting the machine with the living cell. Yet he manages to navigate the tricky issue of what is and is not alive ("Animus"). It's about "internationality" - behavour that makes little sense if you don't take into account a future "purpose". He's very good at showing that this behavour does not necessarily involve consciousness. I note that this particular talk was delivered to the folks at Google, who arguably know more about computers than anyone else on the planet. There were not patronizing giggles from the audience. Questions from the audience were respectful - it was a dialogue between peers.

COMPETENCE

He makes it clear that some entities are more "competent" than others even if they don't actually know what they are doing. Increasing competence is happening in evolution even though it is a mindless process. It is, in the end, governed by a close partnership between the entity and its environment. Not all things are possible but "R&D" is required to find out what is actually possible. In a sense, "thinking" takes place in partnership with reality! How very Zen!

It's worth noting that he sees his pet 3D vision of the evolution of mind from blind chance to intelligent design as not a picture of what is real but an arbitrary choice of 3 axes for the purpose of explanation. In other words, he's very aware of the role of analogy (as brilliantly explained in "Surfaces and Essences"). There are good analogies and bad ones. Dennett is aware of these things as "memes" - (thinking tools) things that evolve to become more useful or are discarded as misleading. This applies to the meme of the meme itself, which may become much more useful than it was when it was born in the pages of "Selfish Gene".

REDUCTIONISM

In his early years, Dennett was literally the poster child for "reductionism", the idea that behavour of systems was "merely" the result of the behavour of their constituent parts. Dennett has become very good not only at accepting the idea of emergence (the whole is more than the parts) but explaining how this becomes true. A good theory of emergence (the classic being the theory of gases based on basic Newtonean physics) don't just explain emergent behavour but explicity tie it down to what is going on at a simpler level. This is never easy.

MEMES

Dennett is still very respectful of the ideas of his friend Richard Dawkins and recommends the book that made Dawkins famous (Selfish Gene). It is interesting to note that Dawkins has since downplayed the idea of memes presented in that book. Dennett, on the other hand, has picked up the idea and run with it. Dennett's "memes" are useful and, in fact, it's possible for Dennett to talk about memes without using the word. An echo of the talking points of "four horsemen" of Atheism is still to be found here, as Dennett speaks of the virus-infected ape - the viruses being memes. This is gratifyingly close to my term: "The Programmable Ape". In both cases, we are referring to the fact that our minds consist largely of culturally determined "memes". To understand this idea, it is not necessary to see the genetic evolution theory of memes as literally true (it is hopelessly imprecise) but it's a great analogy and points precisely to what Zen would claim the "mind" is an illusion - there are only "ideas" running around in our heads. No "observer".

It is culture and especially language that is evolving and getting smarter at an exponential rate over the last 100 years or so.

Dennett's little talk is perhaps the clearest explanation I've seen of what is going on and why. My old "Dragon Theory" (now Programmable Ape) blog explores a special case of this insight - especially the darker implications.

LIFE ITSELF AND "ANIMUS"

Dennett does us a big favor by breaking down the concept of "life" (not merely consciousness). We still don't know how life arose but Dennett has provided a way of talking about the emergence of life without the need for mystical language, a mysterious miraculous event or a whole set of new words to describe what is going on.

M1: A Theory of Everything

I'm beginning to feel the need to organize my ideas into some kind of structure. Let's call that structure "M1". The name comes from "M for Mind and 1 for the primacy of the first person experience".

This should probably be document in a Web Site, rather than a blog, but such a site needs a few initial ideas and some sketch of what it's about. In any case, documentation of M1 needs to be on an interactive platform (not a paper book) to allow dynamic growth and interaction with interested individuals.

Mind Maps

One key tool on the site should be the "Mind Map", which allows us to draw informal diagrams of the relationship between ideas. This is considerably less formal than UML (Universal Modelling Language) that I was first tempted to use. It takes considerable training to understand a UML diagram and there are subtle objections to the ultimate trap of equating an "idea" with a UML "class". The Mind Map allows us to informally model the "space" of ideas and say intelligible things about what ideas are "close to each other" in that space. This includes the ability to visualize the analogy, which, according to Hofstadter, is the underlying organizing principle of the mind.

Core Ideas

1. M1 builds on the first person experience. Everything we know comes from "unpacking" our experience, forming new associations into categories. This is articulated in Hofstadter's "I Am a Strange Loop". The first person experience is also primary in Zen, although Zen and Hofstadter arrive at quite different conclusions about what's "really going on".

1. M1 is ultimately about mind, mainly but not only human minds.
2. M1 builds on first person experience to describe (if not "explain") the emergence of personal memes such as "you" and "them".
3. M1 builds on personal memes to describe social memes and the emergence of "group minds" with "thoughts", "concepts" and "perceptions" that cannot be productively described in terms of some kind of function of individuals in the group (Dragon Theory)
4. M1 is about all possible ideas. Every idea that has ever appeared in all possible mind plus any idea that could appear in the mind, including "right" and "wrong" ideas. M1 is therefore a "Theory of Everything". Annoyingly to some, it includes all other possible "Theories of Everything". This "universe" of memes is sometimes called the "noosphere" but that word is also used in an mystical way to refer to the "mind of the universe". 
5. M1 includes the idea that M1 cannot embrace all of reality. This amounts to acceptance that there is a real world "out there" and our ideas map to that world only incompletely and approximately. Of course, since it includes all ideas, M1 also includes the opposite idea, which is that there is no actual real world out there: it's all in our heads (Strong Idealism). However, M1 concerns mainly with actual, productive and powerful human ideas. Acceptance of the existence of the external world allows M1 to embrace all of "Science".
6. M1 is not "epistemology" which is the philosophical discipline attempting to describe what "we" know and how we know it. How we "know" things is of specific interest in M1, including how and why we "know" things that are manifestly false or inconsistent with the other things we "know".
7. M1 collects imagination pumps (Dennett)
8. The focus of M1 is on "core" memes. These include the "Self", "Mind", "Reality", "Person", "Category" and of course "M1" itself and the meme of a "core meme". 
9. M1 is not reductionist. It makes no attempt to "explain" the universe in terms of a privileged "fundamental" collection of memes (such as those from neurology or quantum mechanics). If anything, the "self" is fundamental, yet, when examined, turns out to depend on other things going on that could be regarded as "more fundamental". 
10. M1 attempts to "unpack" the "self" meme to show the connection between the way cells specialize in brain tissues to produce a "brain" that is, at least in principle, capable of supporting a mind.
11. M1 is shamelessly self-referential and recursive. This is to be expected, since M1 is a product of the mind which has the remarkable ability to know itself and to know it is knowing or even doubt that it is knowing ...
12. M1 is to the universe of ideas as cartography is to the world. It's about mapping and exploring - specifically about tools we can develop to map the meme world - how we get from one meme to another, how we generalize and form analogies.

Reductionism

Descartes is the poster-boy for the method of "reductionism", which I think of as "Math Envy". This is the idea that you can build up all knowledge from simple principles. These days, we pretend to start with Quantum  Mechanics, yet we are awarding Nobel Prizes for people who make "baby steps" from Quantum Mechanics to the chemistry of Hydrogen. The fact is that reductionism is dead. There is no path from "simple" to "complex". We need to start somewhere else. That's good news, since it turns out that we all know a Hell of a lot more than we thought we did.

It turns out that we can start with what we all share and what we know: the fantastic and virtually limitless capabilities of the human mind. We don't need to "boil it down" to something simpler, or "explain" it in terms of chattering neurons. We learn very little about our minds by comparing them to the "minds" of chimpanzees, cats and frogs. If we assume that our minds are "nothing but" two pounds of meat on our shoulders, we succeed in explaining almost nothing while ignoring almost everything we know about our minds from first-hand experience.

The most cogent attack on reductionism I've seen comes from Nobel Prize winner Robert Laughlin: "A Different Universe: Reinventing Physics From the Bottom Down". Laughlin is not commenting outside his field here. He has first hand experience of how we actually figure out what is going on in the invisible world of Quantum Mechanics. His point is that we infer properties of the "small" world from observations in the everyday world and not the other way around. He provides example after example of the fact that we always fail when we attempt to "explain" the whole in terms of the parts. This applies as we climb from simple to complex going from Quantum Mechanics => Physics=> Chemistry => Biology. At each new level, a whole new universe of "emergent" phenomena emerge. Laughlin makes the case that there is a natural barrier between these levels. It's not just the way humans partition the world into University departments. Each new level of complexity is characterized by chaos, non-linearity and a whole new class of organizing principles.

If Laughlin is right, this turns all of Science on its head. Since Descartes and before, the project of Science has been to break down phenomena into the simplest pieces, study the pieces and bring it all together to understand the whole. It just doesn't work [2].

I hope I'm not being unfair to Dennett in guessing that, in spite of his attempt to present himself as a knowledgeable observer of Science, Laughlin's book would be well beyond his ability to comprehend. You need to "weigh a few electrons" in the lab to get what Laughlin is talking about. There is a good reason why it takes a thousand physicists and the biggest, most expensive machine in the world to "discover" the Higgs Boson. It turns out to be exponentially difficult for Quantum Mechanics to "explain" itself in detail, let alone explain simple chemical phenomena, such as the eleven different ways water crystallizes into ice.

[1] It is often forgotten that the ability of Europeans to think in anything like a logical manner was relatively new in the 17th century.  The Moors from Spain  in the 11 the century, resulting in the rediscovery of the Greeks philosophers. Before this, Europeans were not in possession of anything we would call Mathematics or logic. It took centuries to shake off the dead hand of theology and absorb the logic of the Greeks and centuries more to advance it. Progress beyond the Greeks started with Galileo and blossomed with Newton, a contemporary of Descartes. Seen from this perspective, both Newton and Descartes were still in the thrall of "Math Envy", following from the fact that the only thing they "knew for sure" came from Greek Geometry, its axioms and proofs. At least in the popular imagination, Mathematics is the ultimate reductionist discipline.

[2] Stanford's Robert Sapolsky gives an entertaining lecture on reductionism in psychology.

[3] Hofstader provides a devastating attack on reductionism in "I Am a Strange Loop", ironically taking issue with his close friend, Daniel Denett, the poster boy for reductionism. A close read of "Strange Loop" will reveal a further irony: Hofstader takes for granted that, in reality itself, the wriggling of atoms ultimately (in principle) accounts for what we experience as "mind", except the account is irrelevant and useless. He goes to great lengths to convince us that a "higher level" way of seeing things "makes sense".

Deep down, a human brain is a chaotic seething soup of particles, on a higher level it is a jungle of neurons, and on a yet higher level it is a network of abstractions that we call "symbols."

Thus, he still clings to the idea that "out there" in the universe, physics is more "fundamental" than psychology. To let go of reductionism, one must finally accept that "emergent" phenomena are genuinely new and cannot be explained in simpler terms, even "in principle".

John Horgan's Skepticism

John Horgan is a journalist, not a Scientist. A very successful journalist.

You can get the thrust of his approach in

The End of Science

or a short video

Soft vs. Hard Targets
Interview

Science is a vast endeavour and Horgan is bound to miss as often as he hits. His main technique is to go around and interview famous Scientists, then write articles for the layman exposing the flaws in the reasoning of the experts. His main thesis, which he has tirelessly defended for decades is that there nothing fundamentally discovered in Science since, say, the 1950's. This is ridiculous on its face and he's buried with opposition. He's just plain annoying.

His response to the critics is

• Lots of "advances" are simply not true. He especially likes to cite fads in medicine, but these are "straw men". Does proving that some results are false establish that no advances are being made? Anyone who follows the subject is stunned at how often we find answers to questions we never thought of asking. For example, knowing how a cell "knows" how to grow into a horse instead of a tulip sounds like fundamental knowledge to me.
• The promise of many discoveries is over-hyped. The promise implied in decoding DNA is in this category. But he's confusing public confusion and Science Fiction with actual claims and assuming that major breakthroughs are not possible since they have no already happened. For example, we don't know how DNA is decoded into proteins. Paradoxically, there may be a good reason why we will never know, but this is, itself, a major discovery.
• All discoveries are simply working out of existing theory. This is "reductionism", the idea that you can understand the whole by understanding the parts. Reductionism is on its death bead. The explosive growth of the field of complex systems (including chaos, self-organizing systems) is challenging all branches of Science to re-think fundamental assumptions. It's a revolution on a "Copernican" scale and Horgan doesn't see it. I suspect that he doesn't understand how Science works or how Scientist think. That's why he's likely to miss major change to the culture.
• He loves to bash the modern theologies of Multiple Universes, String Theory and Super Symmetry because they are (so far) immune to disproof. But just because many Scientists have followed unproductive lines of research doesn't prove that no productive lines of research remain.
• He is particularly abstruse when it comes to Biology. This field has grown explosively since Darwin provided its organizing principal - Evolution. "Natural Selection" must now take it place among dozens of other theoretical frameworks. 
• He is resolutely ignorant of the fundamental differences between the various branches of science. Why is it so difficult to use Quantum Mechanics to predict useful things in Chemistry? Why do economists so famously fail to predict anything about the economy? There are very good answers to these questions and these answers qualify as major breakthroughs. If, as he claims, we have hit some kind of wall that prevents further progress, the discovery of that wall is itself progress. We are, in fact, learning a lot about these walls. Horgan suffers from a common misconception about Science. Many major findings tell us what is not possible.

As a journalist, Horgan lacks a solid grounding in advanced mathematics, or possibly any mathematics. He was an English major in University and took a few undergrad courses in Physics and Math that left him with a seriously warped idea of what science and math is all about. Undergrads don't "do" science or mathematics. They just learn the answer to exam questions.

His grasp of philosophy also seems wobbly at times. He often mentions Relativity and Quantum Mechanics as fundamental breakthroughs, but these fields cannot be understood without grasping the philosophical and mathematical underpinnings. He underestimates the breakthroughs achieved by Galileo and Newton, chiefly because he's a bit weak when it comes to the history of Science, which is, in the end, his subject. You can't plausibly claim that Science is "over" if you have no idea what Science has been doing in the past. What, exactly is "over"?

I wound up reading Horgan, the journalist, just after I read two books by Laughlin, the Nobel Prize winning scientist. Horgan thinks science (not really his field) is over. Laughlin thinks science is finally purging itself of its fundamental misconceptions and it's just getting started. If I hadn't read Laughlin, I think I might have been seduced by Horgan's writing style without instantly spotting his pathetic ignorance of his subject.

In spite of the fact that he constantly passes judgement on scientists, Horgan doesn't claim to know anything substantial about science itself. But is he a good journalist? He seeks out outrageous comments by "leading" scientists, then quotes them out of context to support his thesis. "The End of Science" is packed with this kind of stuff: completely wrong assertions presented without a shred of evidence. In the end, it's just bad journalism.

Horgan can be safely ignored.