Daniel Dennett at ASU

Last night, Daniel Dennett gave the 2009 Beyond Center lecture with a talk appropriate for the bicentennial of Charles Darwin's birthday, titled "Darwin's 'Strange Inversion of Reasoning.'" While not quite drawing the crowd that last year's lecture by Richard Dawkins did (3000 people at Gammage Auditorium), Dennett filled the 485-seat Galvin Playhouse and an overflow room was set up with a video link. The Phoenix Atheists Meetup group alone had about 57 members who attended.

The talk was videotaped by the Beyond Center, and what may be an unauthorized video has been made available on YouTube.

Skyhooks and Cranes
The content of Dennett's talk was largely drawn from his book, Darwin's Dangerous Idea, and centered on the idea that Darwin brought about a change from thinking of the world as the product of top-down design to a recognition of apparent design as the result of bottom-up processes. Dennett referred to the former as the "trickle-down theory of creation" and the latter as the "bubble-up theory of creation," and used his "intuition pump" of skyhooks vs. cranes to make the point.

"Skyhooks" are explanations of design in terms of miraculous intervention by an entity which itself has no explanation, a deus ex machina. Dennett illustrated that with the drawing above, a Guy Billout illustration titled "Deus ex Machina," from the May 1999 issue of The Atlantic Monthly. By contrast, "cranes" are built up from the ground to provide scaffolding for constructing new things. The dome of the Florence Cathedral (Santa Maria del Fiore), depicted in Billout's illustration, was a marvel of engineering by Filippo Brunelleschi, which used some innovative construction techniques to build something that many thought was not possible.

Darwin's "Strange Inversion of Reasoning"
The title of Dennett's talk came from a critique of Darwin's theory of natural selection by Robert Beverley MacKenzie in 1868, who wrote (as quoted by Dennett in DDA, p. 65):

In the theory with which we have to deal, Absolute Ignorance is the artificer; so that we may enunciate as the fundamental principle of the whole system that, IN ORDER TO MAKE A PERFECT AND BEAUTIFUL MACHINE, IT IS NOT REQUISITE TO KNOW HOW TO MAKE IT. This proposition will be found, on careful examination, to express, in condensed form, the essential purport of the Theory, and to express in a few words all Mr. Darwin's meaning; who, by a strange inversion of reasoning, seems to think Absolute Ignorance fully qualified to take the place of Absolute Wisdom in all the achievements of creative skill.

To which Dennett's response was: "Exactly!" He illustrated the point with an example that is now somewhat commonplace, the computer. Dennett observed that prior to Alan Turing, "computers" referred to people who were hired to perform tasks that today are performed by mechanical devices with the same name. In order to perform these functions, people had to understand arithmetic. Dennett cited Turing's 1936 paper, "On computable numbers, with an application to the Entscheidungsproblem" (PDF), a demonstration that arithmetic computation is a specific case where, in fact, understanding is not required to perform the action--another example of the same kind of "strange inversion of reasoning." Dennett quotes Turing: "The behaviour of the computer [meaning a person] at any moment is determined by the symbols which he is observing and his 'state of mind' at that moment," noting that "state of mind" is in quotes because Turing's showing a method by which no mental activity or understanding is actually required. Substituting into MacKenzie's argument, we get "IN ORDER TO BE A PERFECT AND BEAUTIFUL COMPUTING MACHINE, IT IS NOT REQUISITE TO KNOW WHAT ARITHMETIC IS."

Creationists and Mind-Creationists
Dennett observed that many people cannot abide Darwin, and we call them creationists. There are also people who can't abide Turing, and he suggests we call them mind-creationists. (Steve Novella's presentation at last year's The Amazing Meeting, on "Dualism and Creationism," drew this same analogy.) Dennett said that there are some people who can't abide either--including both Jerry Fodor and Thomas Nagel, referring to his paper "Public Education and Intelligent Design" in Philosophy and Public Affairs vol. 36, no. 2. I think Dennett mischaracterizes Nagel's position here--Nagel is an atheist who thinks that we don't have the full account of evolutionary theory, and who also thinks that if a god exists, there's no reason to think science couldn't study such a being and its effects. I agree with Nagel about that--methodological naturalism could potentially find its own limits and suggest the existence of entities that operate independently of the laws of physics we've discovered. I think we'd end up just modifying our understanding of those laws and continuing to call the result "natural." Jake Young, at the Pure Pedantry ScienceBlog, argues otherwise, defending Stephen J. Gould's "Nonoverlapping Magisteria" (NOMA), the view that science and religion are completely distinct subjects with no intersection, a view I find implausible unless religion is restricted to matters that are completely unobservable and have no causal consequences in the empirical world--which is not the case for any actual religion that I'm aware of.

A few of the "mind-creationists" Dennett pointed out were Jerry Fodor and John Searle. Another is Victor Reppert, author of C.S. Lewis's Dangerous Idea: In Defense of the Argument from Reason, the main argument of which I criticized in a short paper ("Historical But Indistinguishable Differences: Some Notes on Victor Reppert's Paper," Philo vol. 2, no. 1, 1999, pp. 45-47). Reppert's position is that Turing machines don't actually do arithmetic, because they have no semantics, only syntax, and that you only get meaning through original intentionality of the sort that John Searle argues is an irreducible feature of the world. Computers only have semantics when we impute it to them. My argument was that if you have two possible worlds that are exactly alike, except that one was created by a top-down designer and one evolved, there's no reason to say that one has semantics and the other one doesn't--how they got to the point at which they have creatures with internal representations that stand in the right causal relationships to the external world doesn't make a difference to whether or not those representations actually refer and have meaning. [UPDATE (March 3, 2009): Victor Reppert says I've misdescribed his position and elaborates a bit at his own blog.]

Hunting for Skyhooks
Dennett observed that people's issues with bubble-up theories of creation and design center around the fact that some designs seem to be too remarkable to have evolved. Michael Behe's notion of "irreducible complexity" is the idea that some structures require all of their parts in place to function at all, and cannot evolve step-by-step from a previous structure that doesn't also have all of those parts. (The mistake there is that the previous structure may have some other function.) So those arguing for intelligent design have gone "hunting for skyhooks," to try to find examples of design in nature that require a top-down designer's intervening hand to bring into existence. Dennett observed that all of the hunting for skyhooks has failed to come up with any actual examples, but instead has resulted in multiple new discoveries of cranes. This is certainly true for the main examples of "irreducible complexity," blood clotting systems and bacterial flagella. This has led to the quip, "evolution is cleverer than you are," which Dennett discussed in the Q&A as "Orgel's Second Rule."

Another example Dennett gave was the discovery of motor proteins, which he showed using a clip from the film "The Inner Life of the Cell," produced by XVIVO for Harvard University. Dennett didn't mention that this film was the subject of a controversy regarding the film "Expelled," pre-release versions of which used XVIVO footage without permission. Earlier still, intelligent design advocate William Dembski used an overdubbed version of their film in his lectures.

The Bubble-up Path
"We are made of trillions of mindless little robots," Dennett said, "but not a one of them knows who you are or cares." But we do know, and we do care. How is that possible? The bubble-up view has to provide an explanation. Dennett provided some examples of how certain evolutionary changes in the past have created entirely new ways for evolution to proceed. His first example was one that was championed for years by Lynn Margulis to much resistance, but which has now become mainstream, which is the idea of a symbiotic origin for eukaryotes.

For the first 2.5 billion years of life, everything was prokaryotic--single-celled organisms without a nucleus. But then, one form of single-celled organisms invaded another without destroying each other, and came to evolve together, forming eukaryotic life. Each of our cells has not only its own genome in the cell nucleus, but a separate genome in its mitochondria, which is inherited only from our mothers. This development allowed cells to become more complex and versatile, as well as allowed a division of labor that made multicellular life possible.

The Need-to-Know Principle
Dennett showed a video clip about the cuckoo (the link is to a different but similar one). The mother bird lays her egg in the nest of another bird, and removes one of the other bird's eggs. The other bird is then surprised to find that one of its eggs--the cuckoo's egg--hatches first, and the hatchling pushes the other eggs out of the nest. It seems evil, Dennett said, but "don't worry, the cuckoo chick doesn't know what it does. It doesn't need to know."

A principle something like the CIA's need-to-know principle applies in evolution as a matter of thrift, but matters are often confused because biologists tend to attribute more understanding when explaining a feature of living things than actually exists. This, Dennett says, is partly a linguistic matter, because we don't have a word for a "semi-understood quasi-representation" or a "hemi-semi-demi-understood quasi-representation." But Turing does give us models of competence without comprehension.

He then showed a video of a New Caledonian crow trying to use a bit of metal wire to get a worm out of a glass beaker. The crow bends the wire around the glass to make it into a hook, then uses it to fish the worm out of the beaker. This was an example of a creature that goes a step beyond the cuckoo chick. Dennett cited the work of Ruth Millikan, noting that the crow is an example of an animal that represents its goals in the same system in which it represents its facts--but not its reasons for those goals, which are produced by evolution and not represented within the organism.

The MacCready Explosion and Memes
Dennett observed that there has been about 3.5 billion years since the start of the whole tree of life, and only about 6 million years since the divergence of humans from chimps and bonobos, our closest hominid relatives. But a mere 10,000 years ago, as Paul MacCready pointed out, the total human population plus livestock and pets composed about a tenth of one percent of the terrestrial vertebrate biomass. Today, however, we consume 98% of it (most of which is cattle).

The Cambrian "explosion" in which multicellular life became dramatically more diversified took place over millions of years, while the "MacCready explosion" took place over a mere 500 generations, and the explanation is science and technology, communicated from parents to children not by biological evolution but through culture.

Here Dennett gave an introduction to memes by analogy--the cultural highway of transmission of ideas, once it exists, can be invaded by "rogue cultural variants," or "memes," as Richard Dawkins originally called them They are vehicles of information, like viruses, that invade our brains.

He then paused for a "skeptical interlude" to address the question of what's the evidence that memes even exist. He asked, "do you believe that words exist?" If so, then those are examples of a subset of memes, those that can be pronounced. (I'm not sure of the practical benefit of talk of memes as opposed to ideas, concepts, and language, but I'll save commentary on that until I read the meme chapters in DDI.)

So, said Dennett, we are apes with "infected" brains, or, on analogy to prokaryotes/eukaryotes, we are "euprimates." We carry with us virtual machines that give us new powers and versatility to bring organization of the world up another level.

Mind Tools
Dennett quoted one of his own students, Bo Dahlborn, who wrote, "Just as you cannot do very much carpentry with your bare hands, there is not much thinking you can do with your bare brain." We have conceptual tools and methods. At the very simplest level, there are words as tools, such as passwords or labels. Douglas Hofstadter's I Am A Strange Loop identifies a bunch of phrases that are frequently used as tools for analogies, such as "wild goose chases," "tackiness," "loose cannons," "feet of clay," "feedback," "slamdunks," "lip service," and "elbow grease." Dennett compared these to Java applets for the mind--collections of information transmitted from one person to another that allow them to do something more.

Long division is a more complex example. With a sufficiently well developed English (or other language) "virtual machine," you can "download" the procedure in the form of mathematical instruction or from a book, to be able to perform the process. Cost-benefit analysis is a bigger, more complex set of tools learned in the same way.

While some such tools have distinct authors, others have evolved. Language itself, money, and tonal music are examples of such mental tools that were not created at once by individual authors, but have evolved over time.

What this implies for who we are is that we are not Cartesian egos with original intentionality, but "an alliance of hemi-semi-demi-understood virtual machines."

Darwin's Trio
Darwin proposed three types of selection. First, two types of selection where the selective force is human beings. 1. Methodical selection, or intentional artificial selection, where humans intentionally breed creatures for particular characteristics. 2. Unconscious selection, where humans simply preferred certain organisms to others, and helped those to reproduce--such as in farming, and raising domestic animals. To those, Darwin added 3. Natural selection.

Now we've also added 4. Genetic engineering.

And the same categories can be applied to memes. There are original, synanthropic memes, those which live with us but are not domesticated, such as superstitions; these are analogous to memes created by natural selection. There are memes replicated by unconscious selection, such as differential replication of tunes based on how catchy they are. Dennett noted that the Germans call tunes that get stuck in your head "earworms." And then there is methodical selection of domesticated memes, which would include science, literature, and calculus. Dennett compared calculus to laying hens, for which broodiness has been selected out--you have to work hard to get it to reproduce.

And to these categories we can add memetic engineering--spin-doctoring, marketing, propaganda, etc.

Bootstrapping
Dennett asked, how do you draw a straight line? We use a straight edge. And how do we make straight edges? By drawing a line along a piece of metal with a straight edge, and cutting it. How do we get the first straight edge? He pointed to a book on the history of straight edges, and observed that over time we have gradually improved our technology for making straight edges, and can now measure far more precisely how we fall short in reaching the unattainable goal of a perfectly straight line. We can represent our goal, our reasons for achieving the goal, and the imperfections and errors in reaching that goal.

He suggested that the Platonic "form of the true" has a similar history, and that in science "memes have been selected for veridicality."

At this point, we really do have the capacity for genuine top-down design.

Dennett concluded his talk (apart from the next section, which seemed more like an afterword) by stating that "What makes us human is not our genetic children, but our brainchildren. We've finally reached genuine intelligent design."

Darwin Fish
Dennett concluded his prepared lecture by pointing out that he was wearing a Darwin fish lapel pin. The physicist Murray Gell-Mann observed to Dennett that this was patterned after the Jesus fish, a fish symbol which contains the Greek word for fish, which was apparently the first acronym. The Greek letters ΙΧΘΥΣ stand for the Greek words for Jesus Christ, God's Son, the Savior, said Gell-Mann. But what does "DARWIN" stand for?

Dennett took that as a challenge, and came up with a Latin expansion for "DARUUIN" (since there is no letter "W" in Latin):

Delere
Auctorem
Rerum
Ut Universum
Infinitum
Noscas

This translates into English as

Destroy
the author
of things
in order to understand
the infinite
universe

I'm not too fond of this--it confirms anti-evolutionists' worst fears of evolution, and refers to an "author of things" to be destroyed, as though there is one that exists, rather than a myth not to be believed. It's clever, though.

UPDATE (February 20, 2009):

Dennett then answered a few brief questions, and then signed a bunch of books. The first question (and the only one I'll note) was what it was like to work with W.V. Quine, his mentor. Dennett said that he transferred to Harvard University as an undergraduate specifically to work with Quine, and that two of the most significant influences from Quine were the view that science and philosophy are significantly overlapping and parts of the same larger project, and that the quality of Quine's writing (in contrast to his lecture style) was something to aspire to.

He's well-spoken, entertaining, and thought-provoking, and I encourage you to hear him speak if you have the opportunity. I think that his views, like those of Richard Dawkins, argue that science and evolution in particular either imply or at least cohere better with or provide evidence for atheism. I don't think there is a logical implication, and I'm not sure Dennett and Dawkins do, either--that's something that anti-evolutionist lawyer Phillip Johnson has argued, which I've critiqued at the talkorigins.org website, and which the views of Christian evolutionists like Kenneth Miller, Glenn Morton, and Mike Beidler contradict by their very existence. On the other hand, I'm not sure Miller's position is coherent (I really should get around to writing a summary of last year's Skeptics Society conference), and I reject the NOMA view that there is no overlap between the domains of religion and science and agree with Dennett's and Quine's views that there is significant overlap between science and philosophy (and history, for that matter).

The National Center for Science Education and many scientists argue for a sharp divide between science and philosophy, and between science and religion, and find cases like those made by Dawkins and Dennett (and P.Z. Myers) to be problematic, especially when it comes to the legal arena and the goal of keeping intelligent design and creationism out of the public schools (though public universities have more freedom). I think that this is ultimately due to a tension between the principles of separation of church and state, public education, and academic freedom, given that there is no sharp divide between the domains of science and religion (or science and philosophy). In my view, in any case where a religion makes an empirical claim, if there's scientific evidence against that claim, it should be legitimate to discuss that scientific evidence in a public school classroom even if that has the primary effect of inhibiting (or promoting) religion (violating the second prong of the Lemon Test for measuring whether a government action is a violation of the Constitution's establishment clause). I consider it a flaw in the Lemon Test that people can always create new religions which attempt to turn secular ideas into religious content with the specific intent of turning government actions into church-state violations (e.g., creating a doctrine that paying taxes is a sin), as well as the fact that it provides an unwarranted immunity to criticism in the classroom for religious claims, even if they are empirically falsified or conceptually incoherent. (See the comments of this Ed Brayton post at Dispatches from the Culture Wars on the Summum monument case for some legal puzzles. BTW, Justice O'Connor argued for a either a different test in Lynch v. Donnelly, the "endorsement test," which asks whether a reasonable person would conclude government is endorsing or disapproving religion from the action. This has sometimes been interpreted as a complement to the Lemon Test, and sometimes as a substitute for it. Judge Jones in the Dover case applied both the endorsement test and the Lemon Test, and argued that the Dover school district violated both, including all three prongs of Lemon.)

Another resolution is to finesse the issue by getting government out of the business of being a direct provider of education, and instead meet the goal of free public education by providing government funding and standards that include mandatory curriculum requirements that any school can exceed with content that expresses particular religious viewpoints. By providing a fixed amount of per-pupil funding and a mandatory minimum curriculum that doesn't include religious content, those two items are tied together and anything beyond it would be considered provided at the school's own expense, and thus not a church-state violation. In my view, more discussion and debate of religious claims at a younger age will yield better-educated adults (and probably more atheists). Ironically, it is western democracies without a strong history of separation of church and state where religion is weakest and acceptance of evolution is strongest.

Without finessing the problem like that or modifying the Lemon Test, views like those of Dennett and Dawkins must be excluded from public school classrooms along with creationism for the same reasons (to the extent that they express a religious viewpoint), and I think that ultimately the "exploring evolution" or "academic freedom" strategies of the creationists for getting critiques of evolution into the public school classrooms will succeed in passing constitutional muster. Ultimately, the reason their arguments should be excluded from science classrooms is not that they are religious, but that they are bad arguments, and there's no constitutional provision prohibiting the establishment of bad arguments.