Of the many arguments made against design, three are central; many of the others rest upon them. They are that (a) design does not explain by using natural law alone; (b) design assumes an unobservable Intelligent Designer; and (c) design is not testable. In practice, (b) and (c) are linked, and Meyer analyzes them together. Here is his analysis.

                             "Scientific theories must explain using natural laws." Meyer points out two problems with this criterion. First, "many [natural] laws are descriptive and not explanatory." Indeed, this fits the most fundamental laws of physics, which simply state a mathematical relationship without explaining the phenomena involved [think, for example, of Einstein's E = mc2]. Second, many scientific explanations do not use natural laws. These explanations are of the cause-and-effect type; Meyer calls them "historical" explanations, because they rest on "evidence of a unique or distinctive set of past conditions." Among the physical sciences, geology offers many examples of explanations "invoking past events as causes."

                             Evolutionary biology, of course, fits this latter pattern. Evolution as common descent "is asserted to be a well-established scientific theory because it alone explains a diverse class of present data (fossil progression, homology, biogeographical distribution, etc.)" even though biologists offer competing views of how this change might have occurred. The study of origins, then, offers "historical" (relying on past causes) rather than "nomological" (relying on natural laws) explanations.

                             But are some causes - to be exact, supernatural ones - ruled out? To demand that a scientific theory must explain naturalistically is to set up a circular argument. The action of an agent does not necessarily violate laws of nature; it may simply change the conditions. [Meyer does not offer an example, but one might be helpful. A forest fire and an internal combustion engine both use involve the same chemical reaction and result in an energy output consistent with the amount of fuel, but the engine is designed by a human agent to optimize useful energy.]

                             "Design is unscientific because it involves an unobservable Designer and therefore cannot be tested." This criterion links observability and testability. To test something you must be able to observe it; if you can observe it, you can test it.

                             However, Meyer points out that "many entities and events cannot be directly observed or studied - in practice or in principle. …Forces, fields, atoms, quarks, past events, mental states, subsurface geological features, molecular biological structures - all are unobservables inferred from observable phenomena." However, these unobservables are often tested scientifically; they are "evaluated indirectly by comparing their explanatory power against competing hypotheses." Meyer offers the famous example of the choice of the double helix structure for DNA over the triple helix (also consistent with the x-ray crystallography data) because the double helix explained more of the properties of DNA.

                             Meyer's next point is particularly interesting. In considering evolutionary biology, Darwinian transitional, neo-Darwinian mutational events, punctuationalism's 'rapid branching' events, the past action of a designing agent - none of these are directly observable. With respect to direct observability, each of these theoretical entities is equivalent.

                             These hypotheses can be tested only indirectly, by comparing their explanatory power against relevant data. Specifically, "the past action of an unobservable agent could have empirical consequences in the present just as an unobservable genealogical connection between organisms does."

                             Those who make the argument against design by linking observability with testability often find themselves is a self-contradictory bind. If design is inherently untestable, it cannot also be rejected by calling on empirical observations - which implies that it is testable. Both design and descent use similar methods to make their case.

                             Other proposed demarcation criteria. Design and descent show a similar level of "predictive" ability (both are concerned with "predicting" the world we, in fact, see) and both are weak in offering an unambiguous mechanism. However, incomplete scientific theories are neither uncommon nor useless. Indeed, asserting that "all scientifically acceptable causes are mechanistic causes" simply leads to a circular argument for rejecting mental or intelligent agency as a cause. "Theoretically there are at least two possible types of causes: mechanistic and intelligent. The demarcationist has yet to offer a noncircular reason for excluding the latter type."

                             If some sciences are "historical" (rely on past causes) rather than "nomological" (rely on natural laws), it is appropriate to study the methodological approach characteristic of historical science to see if both design and decent fit this methodology, and conform to practices in other scientific fields such as historical geology and archaeology.

                             First and foremost, both design and descent are obviously concerned with historical questions. They, like the other historical sciences, ask "What happened?" rather than "How does nature usually work in such a case?"

                             In addition, the historical sciences in general use a particular type of logic: they "infer a past event from a present fact or clue." In addition, they compare reconstructions of possible past events. [Forensic science - the scientific side of detective work --constantly uses this methodology.]

                             Third, the historical sciences rely on causes rather than natural laws in crafting explanations. They do their testing by assessing relative explanatory power. This last method, assessing "the relative explanatory power of competing theories," is widely used throughout science, including those disciplines characterized as nomological.

                             Many examples of each of these methodological strategies can be found in Darwin's work, which set the pattern for evolutionary biology. Meyer then turns to the question whether design also uses the methods of historical science.

                             Certainly, design, like descent, is concerned with historical questions about how various natural features came about. Design also infers a past cause and uses "event scenarios" to explain present data, although these causes are mental in nature. For example, it postulates intelligent agency rather than random assembly to account for the information content of DNA. Finally, design is testable in the same way as descent: by evaluating its explanatory power against a wide range of current findings.

                             There is, however, a remaining question about the methods and the type of explanations offered by design: is it ultimately a trivialization of science? In other words, "Would a theory of design leave scientists with nothing to do, since presumably the phrase 'God did it' could be invoked as the answer to every scientific question?" Here Meyer turns again to the distinction between historical and nomological sciences. Design answers are not appropriate - indeed, are even somewhat nonsensical - for nomological questions, the questions about how nature usually works. The actions of intelligent agents are not laws, and do not explain general relationships in nature. However, they may be very reasonable answers to questions involving past causes; indeed, forensic science and archaeology routinely consider past agency as causes (although the agents considered are human).

                             Ultimately, however, what really restrains design from becoming a "theory of everything" is the competition among ideas involved in testing explanatory power. Design is not a good explanation for every present observation; where it can be tested against descent is in those cases where descent cannot explain an observation. Thus, modern scientific advocates of intelligent design… postulate antecedent intelligent activity, not because of what we do not know, but because of what we do know about what is, and is not, capable of producing, for example, "information"…. Conversely, there are many effects that do not, based upon our present background knowledge of causal powers, suggest design as either a necessary or best explanation.

                             Here Meyer illustrates with an example that constitutes the most accessible part of his paper. He considers the distribution of pine cones on the ground in his neighborhood and on his campus. Intelligent agency is by far the best explanation of piles of pine cones spelling out "Stewart Hall" appearing overnight in front the dorm of that name. Intelligent agency is a reasonable explanation of piles of pine cones gathered on a tarp by the curb. Intelligent agency is a highly unlikely explanation for pine cones randomly distributed on a lawn or clustered under a pine tree. Intelligent design "can function as the best explanation or best theory of only some things."

                             Intelligent design is not a "theory of everything" for another reason as well. All scientific theories use background assumptions; a theistic approach to design [one which conceptualizes the Designer in a way consistent with beliefs about God found in the monotheistic religions] will be limited by assumptions about divine action. These assumptions include the idea that God acts usually through natural laws and processes and only rarely through exceptions to these processes [in other words, miracles are rare events]. For example, Isaac Newton proposed that the orbits of the Jupiter and Saturn were kept stable by periodic exercise of special divine intervention. Scientists of the time objected, not because Newton invoked divine power, but because such regular interventions seemed inconsistent with the regular order associated with the usual exercise of divine power. As this illustrates, "allowing the design hypothesis… will not cause science to come grinding to a halt."

                             The result so far seems ambiguous: there are problems with every strategy purporting to distinguish science from non-science, and all origins research is subject to special difficulties. However, distinguishing among kinds of sciences - considering what "proof" means in the historical versus the nomological sciences - allows one to decide whether or not design fits the research methods used by other historical sciences. But is all this really beside the point? "What we really want to know is not whether a theory is scientific but whether a theory is true or false, well confirmed or not, worthy of our belief or not." This does not mean that methods don't matter. Explanations involving causes need to exhibit logical consistency and fit present data, whether they deal with naturalistic processes or intelligent agency.

                             Here is the catch. If there is a bias in favor of purely naturalistic causes - as Meyer contends that there is - rejection of design on those grounds gives an unwarranted boost to the apparent solidity of "Darwinian dogma." "If competing hypotheses are eliminated before they are evaluated, remaining theories may acquire an undeserved dominance."

                             Is this harmful?

                             Perhaps scientists should just accept the definition of science that has come down to them. After all, the search for natural causes has served science well. What harm can come from continuing with the status quo? What compelling reasons can be offered for overturning the prohibition against non-naturalistic explanation in science?

                             Meyer offers several reasons why making naturalism "the only game in town" is a flawed idea. Is it really possible to rule out that a personal agent existed before human life on earth? This is not verifiable. Furthermore, it cuts off a rational direction of inquiry. And inasmuch as scientific evaluation of theories is a comparative enterprise, this sets us "an artificially limited set of options." The result of this is that one of a limited number of possible programs in origins research is eliminated at the outset.

                             Meyer concludes, An openness to empirical arguments for design is therefore a necessary condition of a fully rational historical biology…. To insist otherwise is to insist that materialism holds a metaphysically privileged position. Since there seems to be no reason to concede that assumption, I see no reason concede that origins theories must be strictly naturalistic.

                             It's too bad that scientists so rarely read philosophy of science. I avoided it myself: it seemed so much less interesting than actually doing science. But now I'm more than willing to admit that I was, at least sometimes, missing out. Meyer's article made sense: I understood why physical scientists think that a lot of contemporary biology (not to mention sociology or psychology!) sounds pretty sketchy. And I found it fascinating to get a different perspective on some of the formative stories of science, such as brave Galileo bucking the closed-minded Church. It becomes a rather different story when you learn that a large part of the Vatican's problem with Galileo was that his conclusions relied on inductive reasoning and therefore did not meet the mathematical standard that defined "science" in his day.

                             If there is a problem with Meyer's approach from a scientific point of view, it is that he may underestimate the extent to which scientists define science to mean whatever scientists do. In other words, the reason that most scientists don't see a problem with demarcation arguments is less philosophical than social or cultural: it is easy to tell who is a scientist and who is not. Here's where the demarcation lies: a scientist has a Ph.D. in a scientific field from an accredited university. This is how scientists recognize other scientists. It's as simple as that.

                             And that's what makes it complicated. The whole academic and peer review system not only reinforces this definition but also enculturates the scientist into the kinds of research projects and avenues that are fruitful within a given discipline. This is neither bad nor stupid: indeed, it provides a structure for collaboration, affords accountability and checks fraud, for example.

                             But it may also, as Meyer argues, work to close off certain directions of inquiry. A scientist who is not fully committed to philosophical naturalism is very far from ready to commit the professional suicide of being thought a "creationist." (Those who argue that ID is a way to sneak creationism - meaning the young-earth, six-day version - into schools are not only ignorant of what ID is and is not but do incalculable harm to their fellow Christians in science.)

                             And yet, there is one crucial factor in the cultural make-up of scientists that gives me a glimmer of hope that ID might get a fair hearing one of these days. What is this factor? It's simple: scientists are ornery. We like to stick with the herd on some things, but a good, novel research idea - especially an idea that "takes down" one of the powers in one's field - trumps all. Perhaps even today some assistant professor with the chutzpah of the very young and talented will see ID as the ideal win-win project, offering the chance either to kill an unpopular idea…or to be first on the barricades of a scientific revolution.