Pennock's Pragmatic Test
by bipodJohn Timmer is to be congratulated for providing some excellent coverage of the "Teaching Evolution and the Nature of Science" conference. Very good reporting.
What I'd like to focus on in this blog is Pennock's pragmatic test for understanding evolution as science. Take note TelicThought readers. Pennock places a huge amount of value in *what evolution can do for us today* - in other words, how we can use evolution. Directed evolution, by engineers, for people.
The first session was on the nature of science and biology, presented in part by Robert T. Pennock of Michigan State, who testified at the Dover trial. He suggested that teachers should present evolution as part of a discussion of the nature of science, as the development of the theory is an example of science done right. He even suggested that science itself can be viewed as a selective process that discriminates among competing ideas. In contrast, he presented ID as a negative argument against evolution with no explanatory power. Ultimately, however, he suggested that the key feature of evolution is that it passes the pragamatic test: evolutionary processes work in both engineering and computer programming, producing efficient products that would not have been proposed by intentional design, including an antenna used by NASA. I asked him later about the prominent roles played by engineers and chemists (who have careers centered around goal oriented design) in providing creationist arguments with academic credentials, and he suggested that ultimately, the success of evolved designs will win over these fields.
[emphasis added]
April 25th, 2006 at 5:07 pm
Hardly. Evolution by natural selection was a creationist conception articulated by creationist Blyth. For evolution to work it must be designed and created. Further evolvability has limits. Engineers and chemists know this.
Pennock only displays his insulated exposure. Last Fall, I asked some creationist computer science types, "you write software. Do you write Genetic Algorithms?"
"Yes," they replied.
"You think mindless forcess can create software like that," I ask.
"No way," they replied.
Pennock makes a misrepresentation of ID. ID can operate through surrogates like an evolutionary algorithms, but evolutionary algorithms have limits. That was the point of Dembski's No Free Lunch.
Comment by Salvador T. Cordova — April 25, 2006 @ 5:07 pm
April 25th, 2006 at 6:08 pm
As a software developer, I have watched AI techniques, evolutionary programming techniques, and genetic programming techniques for a long time. I have found them to produce tidbits of value here and there, but very little. I feel much more threatened by inexpensive software developers in India than I do by evolutionary programs providing software development expertise.
Comment by bFast — April 25, 2006 @ 6:08 pm
April 25th, 2006 at 9:51 pm
I also have to agree - as a software developer for two decades, AI and its offspring such as GA are very limited. Many of the genetic algorithms that I am familiar with are probably good demonstrations of front-loading, since the selection criteria is very specific to the goal in mind.
Scientists that point to GA's as examples of evolutionary mechanisms should have their tenure revoked!
Comment by David — April 25, 2006 @ 9:51 pm
April 25th, 2006 at 10:20 pm
Of course. That's because AI and GA do not have at their disposal all of the tools that factor in evolution.
Comment by Art — April 25, 2006 @ 10:20 pm
April 26th, 2006 at 1:28 am
Art - Could you elaborate?
Comment by David — April 26, 2006 @ 1:28 am
April 26th, 2006 at 10:25 am
Such as intelligence. . .
Comment by MatthewCromer — April 26, 2006 @ 10:25 am
April 26th, 2006 at 11:28 am
Hmmm, it seems that this forum is dominated by software developers. Maybe we should start a "software deveopers discuss evolution" forum or something. I bet that the software development community would be much more inclined towards ID than the biological community is. I suspect that most of us naturally see DNA as program code.
Comment by bFast — April 26, 2006 @ 11:28 am
April 26th, 2006 at 1:04 pm
"Of course. That's because AI and GA do not have at their disposal all of the tools that factor in evolution."
Art, you crack me up. What tools might those be? Surely you don't mean RM+NS?
Comment by Eric Anderson — April 26, 2006 @ 1:04 pm