Wednesday, December 21, 2005
Defining Science
As of yesterday morning, you can't teach Intelligent Design in some schools. Here's the ruling. (PDF) Hugh Hewitt points out that it really is the only viable decision under current law.
One of the more interesting reads on the topic was Evangelical Outpost. Joe looks at the limits of "methodological naturalism." He uses a definition for this that is posited in a Scientific American piece in July of 2002. Here's the money quote Joe uses:
Back when I was in undergrad school (Dinosaurs still wandered the earth and I had to walk to school, uphill, both ways, 20 miles, but that is a story for another time) I was awarded membership into the Sigma Xi honor fraternity for science. There was a huge debate over the inclusion of psychology majors in the awards -- there was a question as to whether psychology was a science. That may take some of you aback, but lets think about it.
"Science" is really a spectrum. We used to talk all the time about "hard" (physics and to some extent chemistry) and "soft" (everything else) science. But I think it useful to show the spectrum a little better.
Physics stands at the peak. It is a science of the absolutely measurable and the absolutely mathematically modelable. In fact, much of mathematics was invented precisely to do physics. Next comes chemistry, as time has progressed much of chemistry has become very hard, but its inception was purely descriptive - and a good deal of it remains that way.
What do I mean by "purely descriptive?" That means the science observes, records, catalogs, and predicts, but a lot of the time we just know that X will happen when you do Y because it always does - not because our model or theory tells us precisely what is going on -- which is the case with Newtonian physics.
After chemistry comes any number of descriptive or taxonomic sciences. These are sciences of pure observation, description and grouping - biology, zoology, geology.... Now, each of these, some more so than others are becoming harder. For example -- there are now a number of hard theories about fault mechanics in geology (they don't work very well, but they are out there) - so geology is not just about categorizing rocks anymore. But in the end, you develop "ideas" about what might happen next, but that doesn't advance to the level of model in the sense that it does in physics. Sometimes you do very advanced statistics which is very useful for predicting the next event, but that is still very different from a true mathematical model like comes from physics, it's just a very sophisticated calculation of the odds.
Finally, there are the behavioral sciences - sociology, psychology, etc. These are fields of study which may utilize the scientific method and do a lot of statistical study, but you will only every be able to make statistical predictions - human behavior is never predictable on an individual basis. Back in my school days, we hard scientists looked down our noses at these types as pretenders - hence the Sigma Xi debate.
These distinctions I have just described are largely disappearing. Some of this shift stems from the fact that quantum mechanics are highly statistical in nature, so the seeming pinnacle of physics starts to look an awful lot like behavioral science. (Of course, I can't observe the behavior of a single quantum particle like a can a single individual, but that discussion will take us too far afield) But more the shift stems from our desire to grant the less hard sciences the same level of certainty that we associate with physics, a certainty that simply isn't there.
Which takes me to the money quote from SA above. Note that the author tries to equate the certainty of a physics concept with the squishiness of a zoological model. Oops.
Consider what evolution really is -- it's an attempt to explain the taxonomy that we see in the life on the planet, and the fact that that taxonomy is not quite as discrete as we would like -- that it is in fact more spectrum than grouping. It is the application of the scientific method to that taxonomy in an effort to build a model - in other words it's pretty soft science.
One of the reasons physics can be a hard science is that "the system" is well defined with discrete boundaries for any particular problem. Thus methodological naturalism works really well, because we always define the system in the fully observable and natural world. The problems physics investigates are not subject to random deviations or any sort of individualistic behavior.
However, once we move into the softer sciences it becomes extremely difficult to draw those boundaries and define the system. In zoology how many variables does one include to determine which direction evolution will point? In geology, do you consider fault stresses only within 10 miles of the breakpoint, or 100? In the behavioral sciences how do you even begin to account for individually erratic beahvior? And then , returning to evolution, what about behavorial variables? Who mates with whom and why? The boundaries of the system are fuzzy at best in soft science. One must in fact impose upon the problem, particularly when there are behavioral variables, the assumption of purely observable and natural as opposed to having those boundaries arise naturally from the problem itself. Thus, evolution is "science" because an a priori decision was made about the supernatural interference not being involved as opposed to that conclusion arising naturally from the study itself.
Which leads me to what I really hate about all this discussion of Intelligent Design and evolution and everything else. Where you come down on the problem will always be a result of the pre-notions you carry with you into the discussion - always.
Which means Hugh is right not just from a legal viewpoint, but from a scientific one. You want to talk about God in public schools, don't try and do it by developing a new, soft "theory." Do it by addressing the notions that currently prevent it really, the poor court decisions regarding church/state, the prejudice against all things religious, and on the list goes.
BTW, if all this seems a little fuzzy to you, tune in tomorrow when I will post in more detail about mathematical, statistical and decriptive models.
given the judge's explanation of his view of the "endorsement" test and the facts in the record, it is a reasonable result, just as the Ninth Circuit's first "pledge" case was a reasonable result given the "tests" under which the judges involved said they were working.In an interestingly timed article, the NYTimes looked at "science fraud" in light of the revelations of cheating by the Korean scientist and stem cell work. Not related you say? - Right, neither was the "domestic spying" piece related to the Patriot Act vote.
One of the more interesting reads on the topic was Evangelical Outpost. Joe looks at the limits of "methodological naturalism." He uses a definition for this that is posited in a Scientific American piece in July of 2002. Here's the money quote Joe uses:
"Creation science" is a contradiction in terms. A central tenet of modern science is methodological naturalism--it seeks to explain the universe purely in terms of observed or testable natural mechanisms. Thus, physics describes the atomic nucleus with specific concepts governing matter and energy, and it tests those descriptions experimentally. Physicists introduce new particles, such as quarks, to flesh out their theories only when data show that the previous descriptions cannot adequately explain observed phenomena. The new particles do not have arbitrary properties, moreover--their definitions are tightly constrained, because the new particles must fit within the existing framework of physics.Joe responds to this with an interesting thought experiment, but I have another problem with the quote.
In contrast, intelligent-design theorists invoke shadowy entities that conveniently have whatever unconstrained abilities are needed to solve the mystery at hand. Rather than expanding scientific inquiry, such answers shut it down. (How does one disprove the existence of omnipotent intelligences?)
Back when I was in undergrad school (Dinosaurs still wandered the earth and I had to walk to school, uphill, both ways, 20 miles, but that is a story for another time) I was awarded membership into the Sigma Xi honor fraternity for science. There was a huge debate over the inclusion of psychology majors in the awards -- there was a question as to whether psychology was a science. That may take some of you aback, but lets think about it.
"Science" is really a spectrum. We used to talk all the time about "hard" (physics and to some extent chemistry) and "soft" (everything else) science. But I think it useful to show the spectrum a little better.
Physics stands at the peak. It is a science of the absolutely measurable and the absolutely mathematically modelable. In fact, much of mathematics was invented precisely to do physics. Next comes chemistry, as time has progressed much of chemistry has become very hard, but its inception was purely descriptive - and a good deal of it remains that way.
What do I mean by "purely descriptive?" That means the science observes, records, catalogs, and predicts, but a lot of the time we just know that X will happen when you do Y because it always does - not because our model or theory tells us precisely what is going on -- which is the case with Newtonian physics.
After chemistry comes any number of descriptive or taxonomic sciences. These are sciences of pure observation, description and grouping - biology, zoology, geology.... Now, each of these, some more so than others are becoming harder. For example -- there are now a number of hard theories about fault mechanics in geology (they don't work very well, but they are out there) - so geology is not just about categorizing rocks anymore. But in the end, you develop "ideas" about what might happen next, but that doesn't advance to the level of model in the sense that it does in physics. Sometimes you do very advanced statistics which is very useful for predicting the next event, but that is still very different from a true mathematical model like comes from physics, it's just a very sophisticated calculation of the odds.
Finally, there are the behavioral sciences - sociology, psychology, etc. These are fields of study which may utilize the scientific method and do a lot of statistical study, but you will only every be able to make statistical predictions - human behavior is never predictable on an individual basis. Back in my school days, we hard scientists looked down our noses at these types as pretenders - hence the Sigma Xi debate.
These distinctions I have just described are largely disappearing. Some of this shift stems from the fact that quantum mechanics are highly statistical in nature, so the seeming pinnacle of physics starts to look an awful lot like behavioral science. (Of course, I can't observe the behavior of a single quantum particle like a can a single individual, but that discussion will take us too far afield) But more the shift stems from our desire to grant the less hard sciences the same level of certainty that we associate with physics, a certainty that simply isn't there.
Which takes me to the money quote from SA above. Note that the author tries to equate the certainty of a physics concept with the squishiness of a zoological model. Oops.
Consider what evolution really is -- it's an attempt to explain the taxonomy that we see in the life on the planet, and the fact that that taxonomy is not quite as discrete as we would like -- that it is in fact more spectrum than grouping. It is the application of the scientific method to that taxonomy in an effort to build a model - in other words it's pretty soft science.
One of the reasons physics can be a hard science is that "the system" is well defined with discrete boundaries for any particular problem. Thus methodological naturalism works really well, because we always define the system in the fully observable and natural world. The problems physics investigates are not subject to random deviations or any sort of individualistic behavior.
However, once we move into the softer sciences it becomes extremely difficult to draw those boundaries and define the system. In zoology how many variables does one include to determine which direction evolution will point? In geology, do you consider fault stresses only within 10 miles of the breakpoint, or 100? In the behavioral sciences how do you even begin to account for individually erratic beahvior? And then , returning to evolution, what about behavorial variables? Who mates with whom and why? The boundaries of the system are fuzzy at best in soft science. One must in fact impose upon the problem, particularly when there are behavioral variables, the assumption of purely observable and natural as opposed to having those boundaries arise naturally from the problem itself. Thus, evolution is "science" because an a priori decision was made about the supernatural interference not being involved as opposed to that conclusion arising naturally from the study itself.
Which leads me to what I really hate about all this discussion of Intelligent Design and evolution and everything else. Where you come down on the problem will always be a result of the pre-notions you carry with you into the discussion - always.
Which means Hugh is right not just from a legal viewpoint, but from a scientific one. You want to talk about God in public schools, don't try and do it by developing a new, soft "theory." Do it by addressing the notions that currently prevent it really, the poor court decisions regarding church/state, the prejudice against all things religious, and on the list goes.
BTW, if all this seems a little fuzzy to you, tune in tomorrow when I will post in more detail about mathematical, statistical and decriptive models.
