Todd Berkebile’s Blog

I watched The Brothers Grimm the other night which, surprisingly enough, I had not seen before.  I consider myself to be a fan of Terry Gilliam movies even though they can be a mixed bag. 12 Monkeys and The Adventures of Baron Munchausen are two of my all time favorite movies.  Anyhow, I really enjoyed the atmosphere generated in this film.  The costumes and the sets did an excellent job of creating a surreal fairy tale setting.  In a lot of ways this movie is to the Grimm fairy tales what Munchausen is to its Greek and Arabian Nights style fairy tales, complete with occasionally cheesey monsters and special effects.  If you enjoyed the Baron I’d certainly recommend this film.

Last week I finished reading On Intelligence by Jeff Hawkins. In this book he presents the Memory-Prediction Framework for human intelligence. The wikipedia article I linked above can explain this theory far better than I could hope to, I highly recommend reading at least that article if you have any interest in the human mind. Of course if you’re truly interested I’d highly recommend this book. This is one of those theories that, once you read it, makes you think, “Gee, that all makes sense, why didn’t I think of that, it seems so obvious.” In effect his theory attributes most of the abilities of human intelligence to the special way in which our memory works. The most important ability of our brain is the ability to store hierarchal auto-associative invariant representations of the patterns we experience. In layman’s terms, this means our brain naturally forms abstract representations of the things we experience, can associate multiple patterns (sight and sound, for example) related to a thing with one another, and can recall that memory using a similar or partial pattern. This is then coupled with extensive feedback mechanisms so that the abstract representations can affect what we perceive next. Hawkins postulates that a set of neural columns effectively create a “name” for a pattern. This name then feeds back down towards the input layer and can actually be used to fill-in missing information. In other words, a memory is being used to predict the next thing we will perceive. In fact, when we see the world around us as a steady picture a lot of that visual data is coming not from our eyes but from our memory-prediction. Because of the associative nature we might see an object and predict the sound it will make. This association allows us to connect the sight of a bell with the sound of a bell ringing and with the English word “bell” and with the muscle movements needed to speak the word “bell”, for example.

I found the book fascinating, I actually read chapter 6 about three times. That is the most technical chapter in which he discusses how the layout and connectivity of neurons in the brain might explain how our brains can form these hierarchal associative invariant memories. The rest of the book is not very technical but still contains some fascinating insights. Even though this book has had a significant impact on my thinking about intelligence I’m still left with many questions. The obvious neurological questions involve things like the details of how neurons accomplish this task. It almost seems to me that each neuron would require a good bit of computational power to accomplish some of the tasks Hawkins proposed. Can it really be as simple as synapse strength and timing or are logical computations required? Perhaps the quantum processes proposed by Penrose and Hammeroff have a role to play within Hawkins’ framework even if it’s not the role they proposed? My search for answers continues.

When I was flying home from GDC a few weeks back I saw a new root beer while passing though Dulles. I picked up a few bottles but as I was traveling my review of this root beer got delayed in the ether. Hank’s root beer is a fairly traditional root beer. It is flavorful but not too spicy with sarsaparilla being the dominant flavor. It was smooth but not creamy, which is how I like it. The bottle I had was a bit flat, but it did come from an airport convenience cart and survived a flight before being consumed. If you want a decent traditional root beer this one fits the bill, but it you’re looking for something unique this isn’t the drink for you.

A friend of mine somehow managed to con a beautiful and talented woman into marrying him.  The irony is that this is a man who would have sworn off a religious ceremony such as marriage before meeting his perfect compliment.  Anyway, you should all buy her music!  Check out the soulful tunes of Alyse Black.

The discussion on Front-Loaded Evolution was certainly the part of the Design Matrix which I had the biggest trouble with.  This is an Intelligent Design theory that effectively states that even though we know evolution occurs life was originally designed with that knowledge in mind and the designer, being a really clever sort, actually accounted for this.  In fact, the designer placed specific features into the original life for the purpose of getting it to evolve into some other later life form (for example, mammals, or multi-celled organisms, or human beings).  What sort of evidence could possibly support such a theory?

Read the rest of this entry »

I’ve been dipping my toes into the ID debate and I’ve noticed a trend.  So far it seems all of the ID arguments from Mike Gene to William A Dembski to Micheal Behe all center around the idea of complexity.  The basic idea is that some aspect of life, most commonly aspects of cellular biology, is too complex to have arisen by any process other than the deliberate actions of an intelligent agent.  But just what is complexity anyway?

Read the rest of this entry »

I’ve been reading the Telic Thoughts blog and slogged my way through the extensive comments for the Evolving Scientism article.  Surprisingly enough at lunch today before I had read this debate we happened to be discussing whether language was “designed.”  In the case of language we know that it has been developed through the actions of intelligent agents, however it does not seem to have been purposefully developed with any specific goal in mind.  Rather language seems to have evolved over the years through a number of methods that, although very different from genetic evolution, seem to have many analogies to classical biological evolution models.  The debate on the Evolving Scientism article mainly seems to focus around whether we can call something “designed” simply because intelligent agents were involved in this gradual incremental process, or must something be planned for a specific purpose in order to qualify as designed.  Does design require purpose?  If not, does design require intelligence?  If design doesn’t require intelligence then evidence of design cannot support an ID theory.  So just what is intelligence anyway?  Could a gradual process such as the adaptation of language be called “intelligent”, or does that property only apply to the individual agents?  Could a natural process like evolution be called “intelligent” in and of itself?  What is so special about our minds that we tend to think something like them is required for intelligence? 

Read the rest of this entry »

I have finally finished reading the Design Matrix. I enjoyed reading it but mainly because I love picking apart theories I disagree with . I took copious notes while reading the book, you can read those if you want the play-by-play. Here I plan only to give a summary.

Summarized Evidence from the Design Matrix
This is my attempt to briefly summarize the main points that Mike makes in the book. In effect this summarizes chapters three through ten. I think my previous post adequately summarized the earlier chapters. It is my hope that this summary, although brief and greatly simplified, will capture Mike’s intentions without coloring them with my own biases.

• Cellular Biology uses the same terminology that Mechanical Engineers use to talk about the things they design. This common terminology shows an analogy between the cell and things humans design.
• DNA is an information encoding scheme. Human designed things also use encoding schemes.
• DNA is an arbitrary encoding scheme. There are no natural laws that dictate that codons always be three nucleic acids long or that any given codon should code for any particular amino acid. Thus DNA is very analogous to things like human codes and languages.
• DNA is highly optimized as you would expect something designed to be.
• Life uses proofreading and error correction. Humans also use proofreading and error correction when designing things.
• DNA might contain a parity code as one of its error correction methods. Such an algorithm is something a designer would be likely to use but is unlikely to arise otherwise.
• Life really is like carbon based nanotechnology. The analogy between cells and machines is stronger than just terminology but rather indicates a true similarity.
• To jump-start ID research its all right to show a little conformational bias because a lot of research is needed to counter the large existing body of research supporting the old theories. If the old hypothesis is the only theory on the table then it will simply be ad hoc applied to all the evidence otherwise. If you’re not willing to “chase the rabbit” just to see where it goes then you’ll never be able to build up enough data to make a reasonable conclusion as to whether an ID based theory might also explain things.
• The most obvious difference between something that is designed and something created by classical evolution alone is that a design has intelligence and thus foresight whereas natural selection cannot plan ahead. Therefore we can look for foresight in biotic systems to detect design.
• The very design of DNA could have been used to “front-load” later evolutionary changes. In effect, a designer could design things such that a desired outcome was more likely to occur.
• Although evolution can account for Irreducibly Complex (IC) systems through gradual cooption it wouldn’t account for a “system dependent” IC system. As such, if we find many IC systems for which we don’t find the precursor states predicted by gradual cooption then this is a likely candidate for an SD-IC system and it strengths the inference of design.
• Good design implies intelligent design. If something like a cell can be rationally explained via methods like functional decomposition and still holds up as an efficient and practical design then it is more likely to have been intelligently designed.

---

Many of Mike’s claims are supported with interesting examples, some of Mike’s claims aren’t supported with any evidence at all, and a few are supported with highly suspect evidence. For a point-by-point breakdown you’d have to read through my complete notes. Since the notes are a reactionary stream of conciseness they don’t paint a very coherent picture though, so I will try to summarize the main high level problems I see with the theories Mike is proposing:

Read the rest of this entry »

I thought after I moved out of downtown San Francisco I’d see less police actions but it seems every cop in Durham is in my yard right now. There are five or six squad cars parked along the street with their lights flashing. I’ve seen several officers searching through the woods along the jogging trail with flashlights and now a K9 unit just did a sweep around my house. What’s that noise? A helicopter just joined the search. My bedroom is like a disco-tech from all the flashing squad car lights and the helicopter is shaking the walls. Not the sort of excitement I expected at 2 am in a quite suburb. Well the helicopter has moved on so I guess I’ll try to get some sleep.

I have a friend at work who is into Intelligent Design theories. He’s a big fan of Mike Gene from the TeleoLogic blog. I like to consider myself a scientifically minded person and from my experience all ID theories tend to be so-called God of the Gaps theories, which is simply a specialization of the Argument from Ignorance logical fallacy. Both Mike and my friend claim this book is different and that it has a unique approach to the issue so I’ve started reading this book to find out for myself.

So far I’ve only just started, I have read part 1, so I can’t reach any conclusions yet. I did want to share some initial insights though. Perhaps the biggest difference is that Mike is trying to walk the line between the religious ID supporters and the secular Darwinian supporters. He fully accepts the theory of evolution, the forces of natural selection, and, it seems, even the idea of differentiation of species from common origin. This is directly counter to the ID advocates who seem to make the most noise in the press; they argue that evolution is a lie and that ID is an alternative theory. Mike is very clear that he supports a compromise that considers the possibility that both ID and Evolution might be correct theories. The other significant difference is that Mike is not suggesting that scientific methodology be used to test for or prove that life was intelligently designed. Instead he proposes using inductive reasoning to search for clues the way a police detective might search for clues to solve a crime. My friend at work put it this way, the claim of ID is that at some point in the past an event occurred in which life on Earth was designed so why wouldn’t we apply the same methods historians use to study past events to studying ID? Mike calls this process Inductive Gradualism with the idea being the more clues you find that something might be possible the more reasonable it is to believe in that something (or to move through the Explanatory Continuum from “possible” through “plausible” and into “probable” to use Mike’s terminology). He is not claiming this approach could ever prove ID to be true, but rather that it simply gives us justification to seriously consider the possibility that ID might be true.

Of course this is not a scientific approach. In fact even a near infinite number of clues that something might be possible actually does nothing to imply that something is probable. There could be infinite weak clues suggesting a possibility and just one strong clue against it that might outweigh them all. You could just as easily collect clues that the Invisible Pink Unicorn is possible, in fact there is nothing in all of human understanding that denies the possibility that the IPU exists. When a scientist thinks something might be possible she develops a testable hypothesis about what it would mean if that something was true. She doesn’t then say, “We have hundreds of hypothesis that all suggest something is possible, therefore it is probably true.” Instead she makes predictions based on those hypothesis and tests the predictions. Well, actually the scientist does make an estimation about the probability of finding something interesting before committing valuable limited research funds to the experiment. The point is that where the inductive reasoning stops is where science begins.

Mike is trying to walk a very thin line between two rather hostile groups so I can respect his efforts to create a common ground in the middle. Even though I have doubts about Mike’s methods from the get-go I’m still interesting in reading what clues Mike has to offer.