Design and Science (long version)

Science and Intelligent Design
(the long version)

some introductory comments:

This is the original version (May 2001) of Sections 7A-7G in my
Overview of Origins Questions for Science and Theology.

Our questions about Science and Intelligent Design are important, so
initially I gave them the attention (and the writing space) they deserve.
Then, in August 2001, I condensed the main ideas from this version
to make the medium-short and medium-long versions that are
14% and 29% the size of Sections 7A-7G in this page.

TABLE OF CONTENTS

Added Value

7A. What is design?
7B. Can we prove design?
7C. Can design be scientific?
7D. The Freedom of Open Science
7E. The Problems of Closed Science
7F. Cultural-Personal Factors in Science
7G. Can evolution be scientific?

Added Value
Generally, in this page the topics are treated in more detail than in Sections 7A-7G of the Main Overview. The "tips" below are for special places that I think offer "added value" in some way: with new ideas or details, a different explanation, an interesting example,...

In Section 7A (What is Design?),
In this page, 7A is similar to what's in the Main Overview, but here it's more detailed.
design of the universe (connecting Sections 3 and 7A)

In Section 7B (Can design be proved?),
Details (and examples) to supplement design by natural process (which in the Overview was moved into 7E), distinguishing between a design theory and design, and the mental act of design and physical act of design.
WHY a scientific theory cannot be proved or disproved and the concept of theory status.
Theory evaluation by eliminative testing and comparative testing.
An extreme example (with super-intelligent space aliens!) of design denial.
Current Probability and Future Variability and two possible effects of future knowledge with "origin of life" examples.
"The main goal of science is to find truth, but [sometimes] claims that we have found the truth should be made with cautious humility."

In Section 7C (Can design be scientific?),
Of the 7 sections (7A-7G), I think this section is the most worthy of a complete reading, in terms of overall added value.
Why is design controversial? (it's metaphysics, not methodology)
A "two worlds" example of unavoidable error. (from Paul Nelson)
Is science a search for truth? Although it's not the only goal, for most scientists "constructing accurate theories about nature" is an important goal.
Historical judgments about design can be reversed.
Is science a game with rules? A "strong man" example from ESPN.
Positivism, Part 1: regarding observability, with historical analysis by Laudan.
The logic of historical science: deduction, prediction, postdiction, retroduction, and hypothetico-deduction (using agreement and predictive contrast), repeatability and controls.
Is supernatural agency possible?
Evidence and Testing: counterflow (from Del Ratzsch), and an unusual relationship while testing.
Mechanisms (necessary?) and Matching (of claims with methods).
Information (concepts, examples, details) about Supplemented Theories + Design and Creation.
A summary of ideas from Section 7C.
A reminder: reading this whole section may be worthwhile.

In Section 7D (The Freedom of Open Science),
A Goal and Strategy, Problem and Solution.
A theory of design can improve our confidence in a theory of non-design.
Responses to design before history and during history are different. And severe criticism is acceptable unless design is proposed.
More about Behe's correspondence with journals.
Intrinsic Status and Relative Status plus Criticism, Competition, and Confidence (in non-design).
A creative tension between perseverance and flexibility: If the keys are not in the kitchen, can you find them in the kitchen?
What scientists can search for in the second stage of design research.
Positivism, Part 2: Yes, "Scientists cherish their freedom of thought." (So why do they accept MN?)
Should we ask the question? Before we are justified in asking, do we need proof? Mike Behe responds.
A Process of Change: thoughts about some reasons (intellectual and practical) for opposing design.

In Section 7E (The Problems of Open Science),
Science as an expert witness but not judge and jury. But does MN make science a potentially hostile witness?
Shifting from "science is not a search for truth" to "science is our best method for finding truth."
Converting methodological naturalism into metaphysical materialism. Scientism and "explaining God and religion" with MN-Science.
A more in-depth look at hidden arguments and open discussions.

In Section 7F (Cultural-Personal Factors in Science),
The ideas here are similar to those in the Main Outline, exept in more detail, especially in Recognize and Minimize.

In Section 7G (Can evolution be scientific?),
As in Section 7F, the ideas here are similar to those in the Main Outline, exept in more detail.
And there is a subsection with details about historical science: evolutionary retroductions (including scenarios, goals, and levels), observations, and predictions.
And a variety of topics to be written about later.

7A. What is a theory of design?
( the title has been changed to "What is design?" )
Imagine that you receive a radio signal -- 2, 3, 5, 7, 11, 13, 17, 19, 23, 29,... -- and you think it is extremely unlikely that this long string of prime numbers was produced by a natural cause that wasn't intelligently directed. Your conclusion that "the cause of this feature (the signal) involved intelligent design" was reached by a scientific process, by a logical analysis of empirical data. Your scientific thinking began with data, with observations of the signal. Next, you creatively constructed a variety of tentative theories (by imagining various explanations involving either design or non-design) and critically evaluated each theory, then you reached a conclusion that design was the most plausible explanation.
Your theory about the signal is analogous to other theories of intelligent design proposing that a particular feature (which is defined broadly to include any type of phenomenon, any object, attribute, system, or event) was not produced by undirected natural process, but was the result of intelligent design and directed action. In the area of origins, design theories are being proposed to explain the origin of life and complex life and (before the history of nature began) the universe.
Can a theory of design be scientific? As described above, the process of proposing and evaluating a design theory involves a scientific analysis of observations. And logic demands that, during any intellectually rigorous attempt to explain the origin of an observed feature, scientists should consider all possibilities. The feature could have been produced by: 1) undirected natural process; 2A) design and construction with intelligent design by a natural agent (a human,...) and construction using directed natural action (such as shaping a mountainside with dynamite and chisels to make a sculptured face); 2B) design and creation with intelligent design by a supernatural agent and creation using directed supernatural action. A basic theory of design proposes only that a feature was intelligently designed by an agent and was produced by directed action; this "design only" theory, instead of trying to distinguish between 2A and 2B, proposes "either 2A or 2B" and makes no claims about the identity of the agent or the characteristics of the action.

Another type of theory, design of the universe, proposes design to explain why the natural properties of our universe -- its initial conditions (amount of matter/energy, rate of expansion following the Big Bang,...) and characteristics (interactive forces, wave-particle duality, values of constants,...) -- are "just right" for features (nuclei and sunshine, water and DNA,...) that allow life. A theory of design and natural process proposes that the universe was designed so natural process would not just allow the operation of complex physical and biological structures, such as stars and life, but would also produce some of them (or all of them) by undirected natural process, after the design-directed action that produced the universe.

7B. Evaluating Design with Confidence and Humility
( the title has been changed to "Can design be proved?" )
Proponents of intelligent design (ID) claim that, based on a logical analysis of empirical data, we can distinguish between two types of causes: undirected natural process, and intelligent design followed by design-directed action. William Dembski (1999) describes two methods for inferring design, by elimination and competition: A) an eliminative explanatory filter can distinguish between necessity, chance, and design, in order to identify features that have a low probability of being produced by undirected natural process; B) a competitive inference to the best explanation compares the relative explanatory power of ID theories and non-ID theories.

When scientists evaluate a theory of non-design, which proposes that a particular feature was produced by undirected natural process, their conclusions about the theory can range from "very plausible" to "definitely not plausible" and anything in between. Let's examine the two extremes:

IF NON-DESIGN SEEMS PLAUSIBLE

Two Questions, Three Actions, and Mutual Exclusion
If scientists are certain that a feature could be produced by undirected natural process, what can they conclude about design? It depends on the question being asked.

Two Questions
A design question: Was design-directed action involved in producing the feature?
A design theory question: Is it scientifically justifiable to claim the feature shows detectable "signs of design" indicating that design-directed action was involved in producing it? / Unless otherwise specified, a design theory claims there is scientific evidence (i.e., empirically detectable signs of design) supporting a theory that a particular feature was produced by design-directed action during the history of nature. { This definition is used throughout my overview, except in rare cases where the context indicates that another meaning is intended.

Three Actions
To show why answers can be different for the two types of questions, re: design and design theory, consider four possible scenarios for design-directed action (or a lack of it) in producing a feature:
2AB. The feature shows detectable signs of design, and was produced by design-action during history. { e.g., the faces on Mt Rushmore }
1A. The feature does not show detectable signs of design, and was produced by undirected natural process, but this natural process was possible due to design-action at the beginning of history. { This would occur if the universe was intentionally designed so it would naturally produce the feature. }
1B. The feature does not show detectable signs of design, and was produced by design-action during history. { An appearance of undesign is the goal in movies when "special effects" are used to make an artifact, such as a cinematic representation of a tornado or tidal wave, look like the real thing. Or a feature might have been produced with no concerns about whether or not any signs of design would be detectable. If design-action is undetectable and theistic, it is normal-appearing guiding theistic action.
1C. There was no design. The feature was produced by undirected natural process, and the universe was not designed to naturally produce the feature, so there was no design.

The first three scenarios (2AB, 1A, 1B) involve design, but the two requirements of a design theory -- detectability and action during history -- are met by only 2. In 1B the characteristics of design are not detectable, and in 1A neither requirement is satisfied. {although evidence for "design of the universe"
The table below shows the logical outcomes for two realities and two theories. Two possible realities are that "detectable design-action in history" did occur (as in 2AB) or did not occur (in 1A, 1B, or 1C). A design theory claims "it did occur" while a non-design theory says "it did not occur" and denies the claim for design.

	DESIGN theory: a claim for detectable design-action in history.	NON-DESIGN theory: a denial of detectable design-action in history.
With 2AB, there is detectable design-action in history.	The claim (made by a design theory) is TRUE.	The denial (made by a non-design theory) is FALSE.
With 1A, 1B, or 1C, there is no detectable design-action in history.	The claim is FALSE, even if (as in 1A or 1B) design-action did occur.	The denial is TRUE, even if (as in 1A or 1B) design-action did occur.

The table body points out that a theory is TRUE if it matches reality, and is FALSE if it doesn't. The purple text shows that the correct answers can differ for questions about a design theory and design. Even if a design theory (claiming detectable design-directed action during history) is false, it is possible that design did occur, with undetectable design-action either at the beginning of history (1A) or during history (1B).
Now we can answer the question that began this subsection, re: what scientists can conclude. If scientists were certain that a feature could be produced by undirected natural process, a theory of design -- which claims that the evidence supports a conclusion of design -- would be scientifically falsified. The scientists could justifiably claim that "there is no detectable design" and "maybe it was not designed," but they could not logically conclude that "it was undesigned" because the feature might have been designed in a way that would make it appear to be undesigned.
In other words, a theory of design (claiming evidence for design-action) could be scientifically falsified, but a more general theory (claiming an occurrence of design-action, either detectable or undetectable) could not be falsified, and a theory of "total non-design" that denies any occurrence of design (either detectable or undetectable) could not be proved true.
And, of course, we could reject a theory of detectable design for one feature, but accept detectable design as a plausible explanation for another feature.

some thoughts about timing: A mental act of "design" usually precedes the physical act of design-directed action.* The timing between mental and physical acts can be small or large. For example, a creationist who claims that design was involved in the origin of stars (with production by natural process) and the first life (with production by non-natural process) might propose that the mental act of design occurred before the beginning of history, so the properties of our universe (nuclear forces, gravity,... for stars; and for life, the chemical characteristics of carbon, oxygen, sodium, potassium,..., and of DNA, enzymes, water, nerve fibers,...) would, at a much later time, allow a natural production-and-operation of sunshine and a naturally continuing operation (following the miraculous initial production) of life.
* a question to ponder: In situations like musical improvisation, does a mental act always precede (even if by a tiny split-second) the physical act that produces the music? { Theologians disagree about whether God "improvises" during the history of nature, in formative history or human history. }

Mutual Exclusion
In Sections 7B-7D, design and non-design will refer to theories that affirm or deny a claim for empirically detectable design-directed action during the history of nature. When this definition is used, design (in scenario 2AB above) and non-design (in 1A, 1B, or 1C) are mutually exclusive. Either design was involved in producing a feature, or it wasn't.
In principle, the logic of mutual exclusion is simple. But in practice, we cannot be certain when we try to estimate the extent to which signs of design are empirically detectable, as discussed in the rest of this section.

The following subsection is a summary (from the Introductory Overview) of the "1A-1B-1C, 2A-2B" numbering system, plus a table (from the Main Overview) summarizing the characteristics of these five theories:

Design by Natural Process
There are three types of design.
Usually, a design theory is a claim for design-directed action that is empirically detectable and occurs during history. This is the meaning in Sections 7B-7D, and in the design theories (2, 2A and 2B) of Section 7A.
But production of a feature by "non-design" (by what appears to be undirected natural process) could be due to: 1A) design-directed action that occurred at the beginning of history (in a design of nature), 1B) design-directed action that is empirically undetectable and occurs during history, or 1C) a process that is undesigned and undirected.
Evidence against detectable design (2A or 2B) is not evidence against design by natural process (1A or 1B). .....
Creationists propose three types of design (2B, 1A, 1B), while theistic evolutionists propose only design by natural process (either 1A or 1A-and-1B).

This table shows that: a theory is a non-design theory if either condition (empirically detectable, occurs in history) is a no; the origin of a feature can involve design even if a "design theory claim" is not justified.

		empirically detectable?	occurs in history?	involves design?
non- design theory	1A	no	no	YES
	1B	no	YES	YES
	1C	no	YES	-
design theory	2A	YES	YES	YES
design theory	2B	YES	YES	YES

IF NON-DESIGN DOES NOT SEEM PLAUSIBLE

Certainty, Confidence, and Falsification
If all theories of non-design seem highly implausible, have we proved that a design theory is true? No. According to formal logic, in science it is impossible to prove any theory is true or false. Why? Because inductive logic cannot lead to certainty about the truth of a theory claiming that "____ never happens." And deductive logic (if..., then...) cannot prove truth. In principle, deductive logic can prove a theory is false, when deductive theory-based predictions differ from observations, but rigorous proof is possible only in a simple situation. In a complex real-life scientific situation, we cannot be certain whether a disagreement between prediction and observation is caused by a theory or by one of the many other factors that are involved in making and comparing the predictions and observations.
But a radical skeptic is wrong to imply that if science cannot claim certainty, it can claim nothing. Even though proof is impossible, and modern science has given up a quest for certainty, scientists can often develop a rationally justified confidence that a particular theory is approximately true or is false. Instead of demanding a formal proof, scientists are willing to settle for a high level of confidence in a conclusion, and this is the standard we should use when evaluating theories of design.
As a reminder that the outcome of theory evaluation is an estimate of confidence rather than a claim for certainty, we can think in terms of a theory status that describes our level of confidence in a theory, which can vary along a continuum from very low to very high. This concept of a "continuously variable status" is useful because it encourages flexible thinking instead of prematurely forcing ideas into the narrow channels of yes-or-no dichotomies (like proof or falsification) that reinforce mental rigidity and limit the options we are capable of imagining and thoughtfully considering. But binary judgments are still available: if the status of a theory rises above a certain level we can accept this theory as a confident conclusion, and if it falls too low we can reject the theory and consider it falsified. A rationally justified confidence in the falsity of a low-status theory is scientific falsification.

Evaluation by Elimination and Comparison
To test and evaluate a theory of design, we can use eliminative logic, comparative logic, or both.
Eliminative Testing: One approach to inferring design is eliminative, based on the fact that ID and non-ID are mutually exclusive by definition, as discussed above. A feature either was produced by a non-design process (an undirected natural process) or it wasn't. This either-or relationship leads to a bold ontological deduction (that if an ID theory is true, every non-ID theory is false) and a modest epistemological corollary (that perhaps each non-ID theory will seem implausible). If all non-ID theories seem highly implausible, can we conclude that an ID theory is probably true? Yes, we probably can. Although eliminative proof is impossible -- since we cannot be certain that we accurately know the implausibilities, or that we have checked all theories, or (because a true theory might seem implausible, and a false theory might seem plausible) that implausibility and falsity are linked -- we can develop a rationally justified confidence in a theory of design.
Comparative Testing: In historical science, a useful evaluation strategy is a comparative "inference to the best retroductive explanation" that asks whether known observations can be explained more satisfactorily by design or non-design.
When scientists are evaluating a theory of design by using logic that is eliminative and/or comparative, when they are assigning probabilities or comparing relative explanatory power, they are using conventional scientific methods of data collection and logical inference. Their strategies for investigating and reasoning are analogous to the methods used by detectives in forensic science. Because the status of ID theories can be changed (either upward or downward) by a logical analysis of empirical data, these theories are empirically responsive. This makes ID testable, and a particular ID theory can be rejected or accepted if its status becomes very low or very high.

Confidence and Caution
The remainder of this section takes a closer look at some reasons for confidence and caution in our conclusions about design. It examines the difficulties inherent in evaluating design, and it explains why, despite these challenges, some theories of design should be considered worthy of serious scientific consideration.
In principle, it is possible to confidently reject or accept design. Based on our estimates for plausibility and/or explanatory power, an ID theory could be assigned a very low status, low enough to warrant rejection. Therefore, ID is potentially falsifiable. Similarly, an ID theory could be assigned a very high status, high enough to warrant acceptance.
In reality, however, there are reasons for caution, and it may not be possible to reach a scientific consensus about the major theories of ID in the near future. We may not live to witness either a widespread acceptance or rejection of design, due to the uncertainties in evaluation and the wide diversity of views about the philosophy of design.

Five Logical Possibilities
Should we consider all possibilities? For a rational open-minded scientist, the answer seems obvious. Yes, during any intellectually rigorous attempt to explain the origin of an observed feature, logic demands that scientists should consider all possibilities. The feature could have been produced by: 1) undirected natural process, 2A) natural design and construction, or 2B) supernatural design and creation. These three possibilities are described in Section 7A.
To acknowledge the uncertainty in our evaluations, we can expand the list to include epistemological possibilities for "what we can know." Perhaps the feature was produced by undirected natural process and 1a) a current theory proposing an undirected natural process is true (or is at least approximately true) or 1b) a future natural theory will be true, or 1c) no natural theory that seems plausible will ever be constructed because the process was too complex or unfamiliar or improbable for our minds to propose and accept. Or maybe the feature was produced by 2A) natural design and construction, or 2B) supernatural design and creation.

I've constructed this list in order to emphasize two main points.
First, doesn't it seem intellectually rational to consider all possibilities? (of course, these include design, in 2A or 2B)
Second, even though the logical relationship between design and non-design is relatively simple (they are mutually exclusive, in the reality of "what is"), in the domain of "what we know" things are more complicated. Yes, due to their mutual exclusivity, the status of design increases when the status of non-design decreases. But there are three types of non-design (1a-1c), and scientists usually evaluate only current theories (1a) but not the other two categories, so "non-design as a whole" -- which includes current design (1a) and noncurrent non-design (1b, 1c) -- is not being evaluated. Therefore, options are available if new evidence indicates that current theories (1a) are less plausible than previously believed. A scientist can choose to increase the status of either design (2A, 2B) or noncurrent non-design (1b, 1c). There is plenty of room for interpretive flexibility. In addition to vigorous debates about the status of current theories (1a), there is uncertainty about future theories (1b) or no theory (1c). Because there is so much flexibility, could the adoption of any design theory ever be logically warranted? It will be easier to examine this question if we focus on a specific type of theorizing, our efforts to explain the origin of life. This begins soon, following a few comments about the list of five possibilities.

{ comments: Each possibility in the list is actually a category that contains a number of sub-possibilities which may differ in important ways. For example, an appearance of "undirected natural process" could be due to chance in an undesigned universe, or design and natural process in a designed universe, or design-directed action that is empirically undetectable, as explained earlier. There are also other possibilities, such as "supernatural design and natural construction," so this list is not complete. But we can say that "Scientists should consider at least five types of possibilities, and maybe more." }

An Opportunity for Design and Denial
An opportunity for an empirically plausible theory of design is the origin of the first carbon-based life on earth. Why? Because all current theories for the natural chemical evolution of lifeless chemicals into a living organism seem highly implausible. After evaluating numerous theories for a natural origin of life, could scientists rationally conclude that carbon-based life did not originate by undirected natural process? Yes. Even though certainty is impossible because we can never propose and test all possibilities for natural process, scientists could achieve a rationally justified confidence that their search has been thorough yet futile, and no promising approaches remain unexplored. Or they might discover scientific principles, analogous to the principles of thermodynamics indicating the extreme improbability of a perpetual motion machine, that indicate the extreme improbability of a natural process ever producing carbon-based life.
But tenacious critics of ID can always find ways to avoid this conclusion. They can use methodological naturalism as an excuse to bypass the process of science and declare that a naturalistic non-ID theory must automatically be the scientific conclusion, whether or not this conclusion is supported by scientific evidence. Or they can point to possibilities for a future theory (1b) or no theory (1c) in order to forever deny a conclusion of design, no matter how advanced is the state of knowledge. For example, imagine a scientific community composed of trillions of super-intelligent space aliens (IQ = 20,000) each with a life span of a billion years, devoted to science (and using high-speed travel to explore a wide variety of environments throughout the universe) for the past 5 billion years, who have not yet constructed a plausible theory for a natural origin of life. Even in this situation a denial of design would be possible, but would it be rational?
In the near future, however, the actual state of humanly generated knowledge will remain much less advanced, and critics of ID will point out, with some justification, the reasons for cautious humility when making claims for design.

Current Probability and Future Variability
The current state of human knowledge is admittedly incomplete, compared with "everything that could be known," so there is a logical basis for defending non-design. In fact, a perpetual denial of design seems possible because, when evaluating a claim for non-design, we use plausibility estimates that have two dimensions: current probability and future variability.
Imagine that, for a particular feature, our current estimate for the probability of non-design is 1%. Since non-design and design are mutually exclusive, the probability of design is 99%. This sounds impressive, but is it conclusive? No. Even if proponents of non-design agree with these current estimates of probability, they can point to the potential for variability in our future estimates. If our current state of knowledge in this area is fairly low -- if we don't know much compared with what could be known (or even if we're only missing a few key ideas or pieces of data) -- then in the future our probability estimates might be very different. A defender of design can claim that the potential for future variability is large, so in the future the probability of non-design might be anywhere from 0% to 100%. Perpetual denial is possible because, no matter what the plausibility of non-design is now, we don't know with certainty what it will be later. { This probably sounds familiar, because it's just a different way to look at the four types of non-design: the current probability is for 1a, and future variability depends on the total probability for 1b + 1c. }
If variability is unlimited, anything could happen in the future. But in a realistic extrapolation of science into the future, we must place logical constraints on our estimates for variability. If an estimate of future variability is to be practical and useful, it should be based on scientific estimates of what is probable, not wild speculations about what is possible. For example, scientists can look at the specific properties (the unfavorable chemical equilibria, the biological requirements for metabolism and reproduction,...) that make a natural origin of carbon-based life seem implausible, and try to imagine specific ways in which future advances in knowledge might change our views of each property. They can ask, "How likely is each change, and how would it affect our evaluations for a natural origin of life?"

Improving our Estimates of Future Science
When we are estimating the potential for future development of non-design theories, we need better ways to estimate thoroughness (of the search that produced our current state of knowledge) and possibilities (for future improvements in knowledge). To do this we must predict the future plausibility and explanatory power of current natural theories after they are more fully developed, predict the new theories that might be developed, and consider the possibility that we may never develop a natural theory for a natural event that actually occurred. The goal is to estimate the total plausibility of non-design, including both current non-design and (in our extrapolations into the future of science) noncurrent non-design. As described above, we should think in a disciplined way, in an effort to estimate what is probable in reality, not just what is possible in our imaginations.
An epistemological extrapolation to predict future scientific progress will require extensive effort with a high degree of creativity (in imagining what could be) and criticality (in making realistic extrapolations). Doing this well won't be easy, but it should be a high priority for those who are serious about exploring the strengths and weaknesses of a particular design theory, who want to avoid the simplistic extremes of insisting that in the future either "nothing new will happen in science" or "anything could happen."

Two Possible Effects of Knowledge
Will increased knowledge reduce the plausibility of a design theory? Maybe. Future developments in science might lead to new evidence supporting non-design, or it might confirm our doubts about non-design. Both possibilities have occurred in the history of research about chemical evolution:
In 1953, plausibility estimates for a non-designed natural origin of life were higher than in 1952, because the Miller-Urey experiments showed that inorganic chemicals could be converted into small biomolecules such as amino acids. Many scientists assumed that we would soon discover how, on the early earth, a reactive atmosphere had produced large biomolecules that transformed themselves into a simple living cell which could then evolve and increase in complexity.
Since then, however, the warm glow of optimism has been cooled by the harsh reality of improved scientific knowledge. Now we think the early atmosphere was much less reactive than originally assumed, and we recognize the many difficulties in a pre-life synthesis of biomolecules, so visions of a "rich organic soup" (or an "RNA World" or...) have been crushed. And due to rapid advances in molecular biology, now we know more about the remarkable complexity of a "simple" living cell. What seems possible (for the formation of biomolecules in a pre-life environment) has decreased, but what seems necessary (to produce the simplest living organism) has increased. Therefore, the distance between what is possible (before life) and what is necessary (for life) seems much greater now than in 1953. An increase in knowledge has strengthened the scientific support for a theory of design. In the future, if our level of knowledge continues to improve, and if our estimates for the plausibility of a natural origin of life remain as low as they are now, a claim for design will become even more strongly supported.

A non-design origin of life seems less plausible now than in 1953. For similar reasons, new knowledge has also strengthened theories of design in other areas. For example, claims about irreducible complexity are built on the foundation of our increased knowledge in molecular biology, which has opened the "black box" of the cell for in-depth critical examination. And most of the evidence for a designed universe has been recognized in the past few decades.
Although advocates of non-design usually imply that future science will support their own claims, this is not necessarily true. There is nothing automatic about the effects of future science. New research might reveal how a feature was produced by non-design, or it might reinforce our criticisms of non-design. Which of these two scenarios is more likely? This question should be the focal point of an intensive "extrapolation into future science" that effectively combines creativity with criticality, that begins with an open mind and tries to objectively evaluate all possibilities.

Some Reasons for Humble Caution
During an evaluation of design, rational disagreements among conscientious scientists can occur in three areas: A) estimating the probability that a particular non-design process could produce a feature, B) deciding a threshold for a "low probability" of non-design that warrants a claim for design, and C) deciding whether it is reasonable to claim "a rationally justified confidence that the search [for natural explanations] has been thorough yet futile, and no promising approaches remain unexplored."
Until our knowledge and skill has improved in each of the three areas of disagreement, it seems wise to adopt a humble attitude in making claims about what we know and how surely we know. For example, bold pronouncements that a natural evolutionary development of all biocomplexity "is a fact" or "has no support" would not seem justified. Instead, a humble recognition of uncertainty should provide a reason to express our conclusions cautiously and hold them lightly. I think the main goal of science is to find truth, but claims that we have found the truth should be made with cautious humility, at least in areas (like biological evolution) where there is some evidence for and against each of the competitive theories.

An Inference to a Reasonable Conclusion
A theory of intelligent design is empirically responsive and can be falsified. By contrast, a non-ID theory claiming "ID is wrong" could not be falsified in the mind of anyone who, with rigid tenacity and unbounded optimism for future science, would never (under any circumstance, no matter what the evidence was) abandon a claim that "They can't prove a natural cause is impossible."
But proof is not necessary, because among scientists the standard for acceptance is a rationally justified confidence, not certainty. Instead of asking if design can be proved using formal logic, we can ask "Is design a plausible explanation?" We can compare the status for all competitive theories, whether they postulate design or non-design, and instead of feeling it is necessary to make an immediate yes-or-no decision either to adopt ID as the best explanation or to reject it, we can think flexibly and rationally by accepting ID as one of several medium-status alternatives, as an inference to a reasonable explanation.
Yes, in the near future, scientists will have widely varying estimates for the plausibility status of design. But this is not a cause for concern, because disagreement among scientists can be a healthy way to stimulate thinking and research by advocates for different points of view. Empirically based logical analysis shows that some design theories are worthy of serious consideration, that further development of these theories is logically justified. Therefore, it is rational to conclude that the potential of design theories to make valuable scientific contributions should be recognized and welcomed.

[]

7C. Can a design theory be scientific?
( the title has been changed to "Can design be scientific?" )
Even though a claim for design can be supported by the methods of science, by a logical analysis of empirical observations, some people claim that a design theory is inherently non-scientific. Why? This section examines arguments for considering design theories to be nonscientific, along with counter-arguments.

Why is design controversial?
Imagine that, as in Contact (the novel and movie by Carl Sagan), scientists observe a signal containing a series of prime numbers, and they propose a design theory to explain the signal's origin. Nobody would dispute the scientists' right to propose this theory, and there would be no controversy about whether or not their proposal is authentically scientific.
In most ways, a design theory to explain prime numbers (or the stone faces on Mt Rushmore) is logically equivalent to a design theory to explain the first life. In one way, however, there is an important difference, which explains why one theory is calmly tolerated while the other is a topic for intense debate. From experience we know that human intelligence and technology can produce signals and sculptures, so for these features design-and-construction by humans (or by space aliens with adequate intelligence and technology, or...) seems plausible. But if there was no intelligent life in nature to design and construct the first life, a theory of design seems to imply design-and-creation. This is a cause for concern among some scientists and nonscientists, for reasons discussed in the rest of this section, even though design-and-construction theories are not controversial.

What are the limits of science?
When deciding how to define science and its methods, we can choose to restrict the freedom of scientists by requiring that scientific theories should postulate only natural causes. This proposed limit is methodological naturalism (MN). A second limit follows logically: If MN is accepted, and if an event actually does involve a non-natural cause, then any scientific description of this event (in terms of only natural causes, as required by MN) is guaranteed to be incomplete or incorrect. This logical conclusion is humility about methodological naturalism (MN-Humility).
Each limit answers a question. For MN the question is "What are the limits for what can claim to be science?", and the answer is a proposal. For MN-Humility the question is "What are the limits for what a restrictive MN-science can claim to explain?", and the answer is a logical certainty.

Should we accept methodological naturalism?
In 1998, I was willing to support either of two options: 1) reject MN and include design in science, or 2) adopt MN but also adopt MN-Humility by explicitly acknowledging the limitations of MN-science. Here is how I summarized the limits of what MN-science should be allowed to claim:
"We can view a restricted MN-science as one aspect of a broader 'search for truth' that considers all possibilities without imposing metaphysically biased restrictions on theorizing. In this open search for truth, what is the role of MN-science? It can be a valuable resource that should be respected as an expert witness, but it should not be the judge and jury when we're defining reality and rationality." (Rusbult, 1998)
The remainder of Section 7C explains why, two years later when I began to revise this overview, my views had changed. Now I think that -- instead of accepting MN and then apologizing for its logical deficiencies -- we are behaving more rationally if we simply reject MN. If we are serious about searching for truth, it seems wise to adopt an Open Science in which scientists always begin with MN, but do not insist that it is logical to always decide, before looking at the evidence, that we should end with MN.
One reason for rejection is that -- in the scientific and educational communities, and in the public media -- there is a consistent disregard for MN-Humility. Instead of explaining the logical limitations of MN-Science, there is a strong implication that the conclusions of Modern Science must be true because a non-scientific theory does not deserve serious consideration in a modern society. This implication is widespread, despite the logical inconsistency of using MN to bypass the process of science and then claiming the authority of science as support for the unavoidable conclusion required by MN (that "according to science the history of nature was all natural") even if this conclusion would not be supported by the process of science, by a logical analysis of observations and a critical evaluation of all competitive explanations.
But the main reason for my change is practical utility: I've become more thoroughly convinced that, since design theories can be scientifically useful, science should include these theories instead of automatically rejecting them, as demanded by MN.

Can design theories be scientifically useful?
A theory of design can be scientifically useful in two main ways: by improving our search for truth, and by stimulating ideas and experiments.

Is design useful in a search for truth?
Our evaluations of scientific utility will be affected by our definitions for the goals of science. In the short term, scientists are motivated by the exciting intellectual challenge of puzzle solving and by practicalities such as obtaining research grants, publishing papers, getting and keeping a job, making a profit, improving a product, or controlling our environment. But for most scientists in the past and present, an important short-term goal, and the main long-term goal, is to construct accurate theories about nature, about what is happening now and has happened in the past. In other words, an important goal of science is a search for truth, and an activity is scientifically useful if it helps us make progress in our search for truth.
Excluding design from science may hinder our search for truth. How? If, as required by MN, all scientific theories must conclude (before the process of science begins, whether or not this conclusion would have been supported by scientific logic) that everything in the history of nature occurred due to natural causes, and if some events in history occurred due to non-natural causes, then some scientific conclusions are guaranteed to be wrong, yet there is no way to escape these false conclusions. This does not seem rational. On the other hand, if the history of nature has included only natural causes, and if a design theory seems to claim, either explicitly or implicitly, the operation of a non-natural cause, we can use empirically based logic to evaluate this theory and then reject it if this seems justified.
To help us think about the question, "Is MN always the best way to do science?", Paul Nelson asks us to imagine two possible worlds: one world has a history of nature with all events caused by only natural process, while the other world has a history of nature that includes both natural and non-natural events. When we ask, "Which type of world do we actually live in?", we hope our science will help us, not hinder us, in our search for the answer. But in one of the two possible worlds, a science that is restricted by MN must inevitably reach the wrong conclusion. By contrast, in either world a non-MN science will allow (although it cannot guarantee) reaching the correct conclusion.
a summary: If MN is accepted, thereby producing a restrictive Closed Science, it may be impossible to avoid false conclusions. But if MN is rejected, an Open Science is free to reach conclusions based on a logical analysis of observations, and this is what science should do.

Is science a search for truth?
A strong argument for Open Science is a claim that theories of design can be useful in a search for truth, since it seems irrational for scientists (if they are searching for true theories) to ignore theories that might be true. In an effort to deflate this argument, critics of design sometimes appeal to anti-realist philosophies of science, which propose that scientists are not searching for theories that accurately describe reality, that a reality-oriented search for truth should not be a goal of science. Instead, they want to consider only utilitarian and methodological criteria (these are discussed later in this section) and they claim that "even though design might be true, it can't be science."
This is an interesting perspective (which I explore in more detail in another web-page), and rational arguments can be made for and against a claim that truth should not be a goal. But when we look at real people and actual motivations, it seems clear that for most people, both scientists and nonscientists, truth is a goal of science.
Of course, truth is not the only goal. Scientists are also motivated by the intellectual stimulation and satisfaction of solving problems, and by practical benefits such as obtaining grants, earning salaries, publishing papers, gaining respect from scientific colleagues and from nonscientists, and developing science-based technologies that will bring practical benefits like improved health care or new consumer products. Yes, all of these are motivations, but usually scientists also want to construct accurate theories, theories that match the reality of what is happening in nature.
Despite this, arguments against "truth as a goal" (and even against "truth" as a concept!) have gained popularity in academic circles, especially among scholars who are excited about postmodern theories of radical relativism.

Can design help improve non-design?
An activity is scientifically useful if it helps us make progress in our search for truth. As discussed above, an Open Science can prevent unavoidable error if design-directed action really did occur during the history of nature. In this case a theory of design can be scientifically useful because, compared with theories of non-design, it is a more accurate representation of reality and is thus more true.
But even if a particular design theory isn't correct, it can promote a critical examination of non-design theories, encouraging a more accurate evaluation of these theories and their plausibility. When a design theory improves the evaluation of other theories, our search for truth is advanced.
An additional benefit of Open Science, with science evaluating both non-design and design, is that we could place more confidence in a conclusion of non-design. Why? A theory would be granted more justifiable respect if it had earned acceptance due to a comparison, not just with other non-design theories, but also with design theories. We would be more impressed with a theory that had not eliminated part of its competition by insisting that we must bypass the process of science and move directly to a "scientific" conclusion of non-design that is automatic, unavoidable, and not necessarily based on scientific merit.

Does it stimulate action or stop science?
A design theory can stimulate experimental and theoretical research by advocates of design and by its critics. Intuitively, we expect that when the range of scientific theorizing is made wider by including design, it will stimulate a wider range of scientific thinking and experimenting. But is our intuition wrong? Can the pursuit of knowledge be hindered by a claim for design? A common practical concern is that a theory of ID will be a "science stopper" if the response to a challenging problem is to say "it is useless to search for a natural non-design explanation," thus discouraging research in this area of science.
This is a legitimate concern, but the potential chilling effect of design is greatly exaggerated. Most areas of science are not affected at all, because current design theorists are selective, making claims only for occasional events in the history of nature but not for most historical events, and not for the normal ongoing operation of nature. More important, many scientists will persevere, even after they hear a claim that "perhaps this feature wasn't produced by natural process," because they think a natural explanation exists and they can find it. Instead of giving up, "true believers" will be inspired to construct arguments defending their naturalistic theories, gather new data supporting their theories, or make revisions to improve their theories.
In the near future, debates about design will continue, and this can stimulate action. For example, Michael Behe (in his 1996 book, Darwin's Black Box: The Biochemical Challenge to Evolution) asks whether a natural process of step-by-step evolution could produce systems that seem to be irreducibly complex. This challenge has motivated creative thinking and experimenting by advocates of evolution who want to show that Behe's claims are wrong, and by proponents of design.

When we're wondering if scientists who propose design are "giving up" too soon, we should think about our motivations: Do we want scientists to search for truth about nature, or do we want science to be only a game in which the goal is to explain everything by natural process, even if scientific evidence indicates that these explanations are probably wrong? { For example, consider a theory proposing that life arose from nonlife by chemical evolution. Should this be "the scientific explanation" despite its scientific implausibility, simply because it's a natural explanation? }

Is design destined to fail?
In principle, theories of design can be scientifically useful by stimulating action (in response to Behe's questions, for example) and by advancing our search for truth (when criticism leads to a more accurate evaluation of non-design theories, or when a design theory seems plausible and might be true).
In practice, have design theories actually been useful in our search for truth? Does the history of science provide a basis for doubting the utility of design? In the past, some claims for design (and divine action) have seemed foolish in retrospect when MN-science, although temporarily unable to explain a phenomenon, eventually found a natural explanation. Should we conclude, by inductive extrapolation, that claims for design will always fail, now and in the future?
Although inductive logic does not guarantee a correct conclusion, it usually indicates "a good way to bet," so failures in the past should provide a reason for caution. But this should be balanced by a recognition that we can learn from history, so theories can improve. A current design theory should be evaluated based on its own merit, not the weakness of superficially similar theories in the past.
Judging each theory individually will help us avoid two extremes. First, we should not assume that a claim for design is always justified whenever current science cannot explain a feature in terms of undirected natural process. Second, we should not assume that a claim for design could never be scientifically justifiable or that design has never occurred. These extremes can occur only if we refuse to learn anything from history (so we are not cautious in making claims for design) or if we refuse to acknowledge that we can learn from history (so we assume that design theories cannot improve and therefore don't deserve to be evaluated based on evidence).

We should also remember that historical judgments can be reversed. The most famous apparent failure is now being revisited, but with increased knowledge and more sophisticated analyses by Behe and others, to ask whether Darwin really did refute the main claims of Paley. In the future, historians may look back on 1859, not as the beginning of an enduring Darwinian Paradigm, but as the starting point for an interlude, a period of temporary evolutionary confidence that faded when scientists began to explore more deeply and to demand that neo-Darwinian theories should meet higher standards for explanatory detail and empirical support. / And for other important questions, such as the origin of the universe and the first life, MN-science has never offered answers that have seemed even temporarily satisfactory.

Should we go with the status quo?
Another extrapolative argument claims that we should extend scientific methods from the past into the future. If we define science by what scientists do, and the majority of current scientists practice MN, does this make MN an essential part of science?
In evaluating this argument, the basic question is simple: Is it wise to assume that current methods are necessarily the best of all possible methods, in all situations? If we say YES, then it is rational to decide that we should go with the status quo, that "what is" can tell us "what should be." If we say NO, then we can think more carefully about our science, and we can ask "What are the best methods?" and "Is MN always the best way to do science, in every situation?"
The question of MN can be approached in two ways, by logic or power. We can use logic to evaluate the strengths and weaknesses of MN, to consider the benefits of an Open Science. Or we can ignore the question, thus letting it be answered by those who have the power to define "what science is" by making important decisions: which views will (and won't) be expressed in scientific journals and textbooks, at conferences and in the public media? what types of research, by which scientists, will be funded by grants? who will be hired and promoted? and who will determine the policies of scientific and educational organizations?

Should MN get credit for everything?
A common argument for the status quo is that "modern science is MN and is making wonderful progress in gaining knowledge and solving problems, so we should retain MN because it is so effective." But this conclusion isn't justified if, as we have good reasons to expect, a tolerance of design would not hinder the progress of science:
Most areas of research would not be affected. In areas where design is being proposed, many scientists would continue to work on non-design theories. And this work would be approved by design proponents who think, for example, that a neo-Darwinian approach is useful for exploring a wide variety of phenomena, and that many of its claims (but not all) seem scientifically justified. In fact, Mike Behe is challenging scientists to do more, not less, in searching for evolutionary mechanisms at the level of biochemistry. And an improved knowledge of non-design can increase the plausibility of design, as explained in Sections 7B and 7D.
As discussed later in this section, methods of logical analysis are similar in design and non-design, except that with design the range of acceptable conclusions is expanded. Overall, design would have very little impact on the practical productivity of science, so an argument that "a rigid MN is necessary for the progress of science" is not justified.

Is science a game with rules?
Some critics of design view science as an intellectual game played with a set of rules, which include MN, that have been established by tradition, approved by consensus in the scientific community, and enforced by funding agencies, journal editors, and hiring committees.
This is an interesting perspective. In terms of sociology, regarding interpersonal dynamics and institutional structures, it is certainly an idea with merit. But it becomes much less impressive and less appealing when we turn to philosophy and think about functional logic and the cognitive goals of science, when we acknowledge the distinction between games and reality.
The practical function of restrictive rules is different in a game and in reality. To illustrate, consider the Strong Man contests televised by ESPN. During these competitions, I've seen a man tow a semi-truck with a rope, and carry a refrigerator on his back.
For the game, if one competitor wanted to hook the semi to a tow truck or strap the refrigerator to a two-wheeler, this would be cheating. It would provide an unfair advantage and would not help in achieving the goal of the game, determining who is the strongest man. In this context the rule about "no mechanical help" is useful.
But for reality, for accomplishing a practical goal, the same rule might not be useful. If the real-life goal of a business is to move vehicles or refrigerators quickly, over and over throughout the day, using tow trucks or two-wheelers is a more effective strategy than asking a person to do all of the work.
It is obvious that a restrictive rule which is useful in the context of an artificial game -- such as requiring that a heavy object must be moved by a human without extra help -- may not be useful in real life for accomplishing practical goals. When this principle is applied to science, it seems more rational to view science as an activity with goals, rather than a game with rules. Then we can ask whether the restrictions imposed by MN will make scientists more effective in pursuing and achieving the goals of science. More specifically, we can ask "Is MN a useful strategy in our search for truth, in our development of increasingly accurate theories about nature?"

Is "natural science" a logical argument?
A claim that "in natural science, natural phenomena and natural history should be explained by natural causes" is just a fallacious pseudo-argument. On the surface, the logic seems impressive. But when we look deeper, it vanishes into thin air. By carefully examining each use of "natural" we see a shift of meaning that hides (but only for awhile, until we discover the verbal illusion) an illogical circular argument, produced by using the same word in two different ways.
One meaning of "natural" -- which is used throughout this overview -- is normal-appearing, in contrast with miraculous-appearing. In the sentence above, this meaning is used once, for natural causes. A second meaning, pertaining to nature, is used for natural phenomena (phenomena that occur in nature), natural history (the history of nature), and natural science (science that studies nature).
But verbal ambiguity doesn't even require a sentence. All by itself, the term "natural science" is ambiguous because it could refer to either "a study of the natural" (in Closed Science, restricted by MN) or "a study of nature" (in Open Science, liberated from MN), since "natural" has two meanings. But there is an implicit argument when we say "natural science" because we're implying that these two words belong together, by definition, that they are inseparable and form a logical unit, that science is natural (with "normal appearing" as the assumed meaning) so a naturalistic approach should be accepted without further critical thought. To avoid this pseudo-argument, we must improve our verbal precision. Instead of saying "natural science" we should use terms that are more general (science) or more specific (physical science, earth science, biological science, social science,...).
a reminder: In this overview, "natural" always means "normal appearing".

Could science survive a miracle?
Would a miracle mean the end of science? One methodological concern is that science would be impossible if miracles occur, because the logic of science depends on consistently reproducible results. This objection is based on a sound principle, but it loses practical validity when it extrapolates from USUALLY to ALWAYS. Yes, a world that is "usually natural" is necessary for science, but we don't need an "always natural" world. Science would be difficult, if not impossible, if we lived in a world with frequent "Alice in Wonderland" surprises and no reliable cause-effect relationships. But if, despite occasional miracles, the universe usually operates according to normal natural patterns, science will be possible and useful.

Do theories of design propose miracles?
Theories of intelligent design (ID) assume the universe usually operates according to normal natural patterns, with consistently reproducible results. ID makes claims only for occasional events in the history of nature, not for the normal operation of nature. And a basic design theory does not explicitly propose miracles in history, because it claims only that a feature was produced -- not by undirected natural process -- but by either natural design and construction or supernatural design and creation.
Does a basic "design only" theory violate methodological naturalism? Maybe. It is difficult to answer with a simple "yes" or "no" because with ID there are two possibilities and also two interpretations. ID allows supernatural design-and-creation as a possibility, which seems to violate MN. But ID also allows natural design-and-construction as a possibility, so a design theory does not require miracles and therefore does not explicitly violate MN. What a design theory does explicitly acknowledge -- and this is where it differs from a creation theory -- is the practical difficulty of scientifically distinguishing between supernatural creation and natural construction. But if critics of ID move beyond what is actually proposed, to include what they think is the total content (both explicit and implicit) of an ID theory, they can claim an implicit violation of MN.

Logical Methods and Design
A defense of Closed Science often begins at a superficial level, with an assumption -- that "natural science" is natural -- which isn't questioned. At this level, the only question being asked is whether design violates a tradition of methodological naturalism.
But we can move to a deeper level by thinking about scientific utility, by asking "What would be the practical effects of an Open Science that includes design?" and "Does a rigidly enforced MN help or hinder our efforts to gain an accurate understanding of nature?" Although some aspects of these questions have been discussed above, other important concerns about scientific utility -- concerns involving the logical methods used in science -- are examined in the remainder of this section. I hope you'll find some useful ideas in the following discussion.

Observable Effects
When we're evaluating the methodology of design, an important principle of cause-and-effect is that an unobservable cause can produce an observable effect. If we examine a feature and observe distinctive signs of design we can logically conclude that design has occurred, even if we have not directly observed the agent who formulated the design or performed the directed action that produced the designed feature.
During the history of science, this principle -- that an observable effect can let us scientifically infer the existence and action of an unobservable cause -- has been debated, and its logical validity and scientific utility have been accepted by scientists and philosophers. Laudan (1977, 1984) describes a conflict between beliefs that resulted in a significant decision about the methodological foundations of science. In the early 1700s, some interpreters of science claimed that Isaac Newton had constructed his theories only by inductive generalization from observations, and had refused to speculate about unobservable theoretical entities. A claim that Newton's method of theorizing was based on a commitment to "no unobservable components in theories" is challenged by modern historians and philosophers, but for awhile this methodological strategy was influential in science and philosophy. Scientists were inspired to mimic the methods they mistakenly thought Newton had used, so they tried to develop theories that -- consistent with the positivist models of knowledge being developed by philosophers -- did not include unobservable causal entities. But by the 1750s it was becoming apparent that many successful theories, in a wide range of fields, in operations science and historical science, depended on the postulation of unobservable entities. Thus, there was a conflict between positivist goals for science and the actual theories of science.
Instead of giving up their non-positivist theories, the scientists and philosophers "sought to legitimate the aim of understanding the visible world by means of postulating an invisible world whose behavior was causally responsible for what we do observe. ... To make good on their proposed aims, they had to develop a new methodology of science,... the hypothetico-deductive method. Such a method allowed for the legitimacy of hypotheses referring to theoretical entities, just so long as a broad range of correct observational claims could be derived from such hypotheses. (Laudan, 1984; p. 57)"
Using this logical methodology, modern scientists often propose that observed effects were produced by an unobserved cause. In hypothetico-deductive reasoning, the only requirement -- even if a cause cannot be directly observed -- is that effects can be observed. This requirement is fulfilled by a theory of design which claims that a cause (the design-directed action involved in producing a feature) can be inferred if, when we examine a feature, we observe distinctive signs of design. {more about positivism}

Historical Science
Theories of design are proposed in historical science (to study events in the past) but not in operations science (to study ongoing events in the present).

In both types of science, the logic is similar.
Scientists use a logical process of deduction when they infer from a proposed cause to a predicted effect by asking an if-then question -- If this was the cause, then what effects should we observe? -- that produces a theory-based prediction (made before the observed effects are known) or postdiction (made after the observations are known). A prediction and postdiction are logically equivalent, if each is obtained by valid deductive logic.
Scientists use retroduction when, after observations are known, they infer from an observed effect to a proposed cause by asking a reversed question: These are the observed effects, so what might the cause have been? During retroductive inference, scientists try to find a theory (by selecting an existing theory or inventing a new theory) whose postdictions will match the known data.
Retroduction is one type of hypothetico-deduction, which is a general process of logical inference that uses degree of agreement (do observations agree with predictions?) to evaluate a theory, and uses degree of predictive contrast (what differences occur in the predictions of different theories?) to compare and evaluate competitive theories. { For details, check Section 7G. }

But in the two types of science, data is different.
In operations science, in experiments (and sometimes in field studies) observations can be repeated and variables can be controlled.
But in historical science, repeatability and controls are impossible (except for decisions about which phenomena and characteristics to observe), and a deduction must be made after an event has occurred. { But deductions can be made either before or after event-data is known, to generate predictions or postdictions, respectively. } Since these data limitations occur in all areas of historical science, including astronomy, geology, and evolutionary biology, they pose no special problems for design.

Regarding the relations and timings of deduction and inference, our expectations are similar for historical scientists and for other historians. In all areas, including science, we expect historians to construct descriptions of what happened in the past and to propose explanatory theories for how it happened, but we don't expect them to predict what will happen in the future.

Personal Agency
In both historical science and operations science, in some situations we must consider the effects of personal agency because "what happens" depends on the decisions and actions of individual agents. In these situations our ability to make precise predictions will usually decrease due to the unpredictability of individuals. But similar methodological problems exist in design and in other scientific theories that postulate action by agents.
In the production of a designed feature, there are two opportunities for agency: in the design, and the design-directed action. The agents for these two phases could be the same (as implied in Section 7A) or different. For example, engineers could design an airplane, then factory workers actually build the plane. Or, God might design a "plan for action" and then use a human to carry out the action. In fact, this type of dual agency -- combining the supernatural and natural, with God and humans working together -- plays an important role in theology and (I think) in life.

Supernatural Agency
In historical science, supernatural agency and natural agency are methodologically similar. In each case a past occurrence of design-directed action can be inferred when careful examination of a feature reveals observable signs of design, even though (since it occurred in the past) the agent and action have not been directly observed. In each case it is difficult to make precise theory-based deductions (as either predictions or postdictions) but it is possible to scientifically justify a retroductive inference that "design-directed action by an agent has occurred." The logical process of inference is similar, whether the agency was natural or supernatural.
If an event involving agency was not observed, there are two possibilities. Perhaps the agent could have been observed, but was not. Or maybe the agent, if supernatural, could not have been observed, even by an eyewitness. In each case, the agent has not been observed, but observable effects -- which are the foundation of scientific logic -- could be produced by either type of agent. Therefore, it is not methodologically useful (at least it isn't useful in the "ancient history of nature" situations for which intelligent design is typically proposed) to make a distinction between agents that are unobserved and unobservable.
But is supernatural action possible? An atheist and deist and theist, due to their differing beliefs about the existence and activities of supernatural agents, will have different views regarding the possibility of supernatural action. Since there is no metaphysically neutral way to decide between these views, it may be wise to adopt the non-restrictive policy of Open Science by saying "maybe supernatural action is possible, and maybe it isn't."

Section 7G contains a deeper analysis of historical science, personal agency, and supernatural agency. It examines cause-effect principles and hypothetico-deductive inference. It describes how the process of retroductive inference is affected by predictive accuracy and precision, which in turn are affected by contingency and complexity, empirical and theoretical knowledge, mechanisms and agency. It explains why, although in historical science there are reasons for caution due to inherent limitations in the available data, scientists can develop methods for reducing the practical impact of the limitations. These methods should be critically analyzed, but we should not automatically eliminate historical science (whether it proposes design or non-design) from being authentically scientific.

Evidence and Testing
Careful observation, combined with intuition and analysis, can lead us to recognize distinctive signs of design, such as special characteristics (like prime numbers) or purposeful functionality (like a camera that makes a photograph). Del Ratzsch describes design in terms of the counterflow that occurs when events "have been pushed in a direction contrary to the normal flow of nature." He uses a diesel bulldozer as an example of an object with clearly observable counterflow marks that "cannot or would not be produced by nature." These signs of design provide evidence indicating that the bulldozer was produced by design.
When is it logically justifiable to make a claim for design? There is evidence for design when production by undirected natural process (by non-design) is not a plausible explanation for a particular feature, when it seems more likely that the feature was produced by design-directed action. A feature was produced by either design or non-design, which are mutually exclusive. Therefore, when the evaluative status of non-design decreases, the status of design increases; evidence against non-design is evidence for design, so we can use the predictions of non-design theories to test and evaluate a design theory. This relationship between predicting (by non-design) and testing (of both non-design and design) is unusual, since most theories are tested by using their own predictions. But this methodology -- of supporting a claim for design by gathering evidence against non-design -- is logically valid because the correct theory must come from within one of the two mutually exclusive categories, design and non-design.

Testing and Falsifying
Can design be proved or disproved? No. Section 7B explains why, since we cannot falsify all possible theories that claim non-design, we cannot prove design. Similarly, design cannot be disproved. This is not a cause for concern, however, because it is impossible to prove or disprove any scientific theory by using formal logic. But even though we cannot formally falsify design (or any other theory), it is possible to scientifically falsify a design theory, to develop a rationally justified confidence that a theory of design is false. Or we could be scientifically confident that a design theory is true, or that design is the best of the currently available explanations. Or we might simply conclude that design, as a potentially plausible explanation, is worthy of further development.
Critics of design tend to focus on falsifiability, with two contrasting concerns that form an interesting combination. First, they worry because design cannot be formally falsified. But non-design also cannot be falsified, so they worry because design cannot be proved. Somehow, the lack of symmetry in their own concerns (with unfalsifiability in design being criticized, while unfalsifiability in non-design is praised) doesn't seem to be a cause for concern.
When we're thinking about the testability of design, focusing on falsifiability is not the most practical approach. Instead, it is more useful to ask, Can the evaluative status of a design theory be changed (either up or down) by empirical data? This question, which is a sensitive and logically valid way to determine whether a theory is empirically responsive, can be answered YES for many theories of design, so these theories can be scientifically evaluated based on empirical data. Critics confirm the testability of design when they try to show, using empirical evidence, that a particular design theory is less plausible than is claimed by its proponents. {example} If the design theory was not empirically responsive, if it was immune to evidence and could not be tested, such criticism would serve no practical purpose.

Design uses Scientific Logic
Design theorists oppose the restrictions of MN-science, but they use the logical methods of MN-science. They begin with MN, but don't insist on always ending with MN. They use conventional methods of scientific analysis, such as the hypothetico-deductive reasoning (done by comparing theory-based deductions with observations) that serves as a "reality check" for empirically evaluating the plausibility of a theory. They think that -- if scientific evaluation shows theories of non-design to be implausible -- it is rational to reject these theories instead of insisting (as in MN-science) that a non-design theory must be accepted anyway.
Conventional scientific methods, applied with an open mind, are sufficient to provide logical support for design. In fact, in some areas of current MN-science, logical methods for detecting design are now commonly used. For example, theories of design are being proposed and tested when forensic scientists investigate crimes, and when astronomers search for radio signals designed by intelligent extraterrestrials. Similar methods are being used, and further developed, by design theorists.

Mechanisms and Matching
For theories that propose a detailed causal mechanism, we can make detailed predictions by using deductive logic, by asking an if-then question, "If this theory is true (so the proposed mechanism is operating), then what should we observe?" But a basic theory of design does not propose a detailed causal mechanism; it claims that a feature was the result of design, but makes no claims about the designing agent or the directed action that produced the feature.
Scientists generally prefer a theory that includes a precisely detailed mechanism, but this is not required for scientific acceptance. And a mechanism is not required for scientific utility. As discussed above, even though design theories don't propose a detailed mechanism, empirical evidence that is analyzed by conventional scientific logic can support claims for design.
There is no need for detailed predictions, unless in addition to expecting that a design theory should be tested for what it does claim (that design occurred), we also demand that it must be tested for what it does not claim (that it can explain the details, the "how, when, why, and who" of the design-and-production process). This extra demand is not needed to evaluate a claim for design, and it is not reasonable. Why? For the same reason we don't demand that officers in a police department should never turn a case over to the Homicide Division until they already know the details and have identified the murderer. An exploration of design in any area (in homicide, radioastronomy, origins,...) is a two-step process: first ask "Was there design-directed action?" and then investigate the details.
We should recognize the limited claims made by a design theory, so we can evaluate design based on what it is, not what it isn't but never claimed to be. This recognition is important, since it will help us think about testability in a way that is logically appropriate, that achieves a match between the claims made by design and the methods used to justify these claims.

Supplemented Theories
As explained above, a basic "design only" theory should be considered scientifically acceptable, whether it's proposed in astronomy (where a signal containing prime numbers would be recognized as design, even if nobody knew who the designers were) or to explain biological origins. But if we want an origins theory that is more complete, a basic design theory can be supplemented with details about the designer's identity and actions.
For example, a theory proposing design for the origin of increasingly complex life on earth could be supplemented with proposals for old-earth creation by God, young-earth creation by God, creation by another supernatural agent, or construction by natural extraterrestrial agents. Each of these five explanations (the basic theory and the four supplemented theories) can be evaluated independently on its own merits, to generate five different estimates of plausibility, one for each theory. Supplementation can significantly affect the scientific content and plausibility of a design-and-creation theory, as discussed in "The Many Meanings of Creation" in Section 6B.
Supplementation also affects our judgments about how "scientific" a theory is. Some arguments for "design as science" also apply to "creation as science" but others do not apply. And the question being examined in this section is whether design (not creation) is scientific.

Design and Creation
What is the relationship between design and creation? They are related, but are not the same. They are logically connected, but only partially, since a theory of origins design can be (but doesn't have to be) supplemented to form a theory of theistic design-and-action. They are theologically correlated, but only partially, since most advocates of design (but not all) are Christians. And design is not the same as "young earth" creation, since two different questions (Has design occurred? How old is the earth?) are being asked. { Also, claims for design are based on scientific evidence, while defenses of yeC, especially within the Christian community, often rely heavily on Biblical interpretation. }
Two important principles -- 1) each theory should be carefully evaluated based on what it claims, no more and no less, and 2) design can involve agents and actions that are either natural or supernatural -- are discussed in this subsection:

1) When theories are compared, we should notice both similarities and differences. For example, when we compare typical theories of old-earth creation and young-earth creation by carefully examining the individual components within each theory, we see some similarities (e.g., both agree that irreducible complexity seems to exist) and some differences (e.g., they disagree about young-earth flood geology). Each of these components, irreducible complexity and flood geology, should be evaluated separately and independently. An overall theory of creation that combines many components, that makes many claims, should be evaluated with a wide focus that includes all of the claims, as explained in Section 6B. But a theory of design that makes one claim, such as "the design of irreducibly complex systems," should be evaluated with a narrow focus that considers only this claim.

2) What are the implications of design? If a feature was not produced by undirected natural process, what are the alternatives? An agent of "design and directed action" could be a nonhuman animal (as in ant hills, bird nests, and beaver dams), a human, an extraterrestrial space alien, or a supernatural being. For each major area of origins, design doesn't necessarily mean creation:
a) For biological evolution, the "production scenario" most commonly imagined by naturalistic design theorists is for directed panspermia, with the evolutionary development of life on earth being stimulated and guided by natural extraterrestrials who became highly evolved before the advent of complex earth life.
b) For the first life, current design theories claim that a natural formation of carbon-based life (involving DNA, proteins, water,...) is highly implausible. What are the alternatives? Maybe God designed the universe so it would naturally support life, but would not naturally produce life, so God miraculously created life on earth. Another design theory, nontheistic and naturalistic, might propose that life did arise naturally, but it was life of a type (and in a setting) we cannot now imagine; then this life evolved to a level where it could design and construct the familiar carbon-based life inhabiting the earth.
c) For our universe, with its life-allowing properties, a design theorist with a pantheistic (or panentheistic) worldview might propose that a universe can somehow evolve into a unified super-consciousness that is intelligent, active, and powerful. Perhaps this happened in the distant past, and a previous universe has already designed and constructed new universes, including the one we now inhabit.

A basic design theory -- before it has been supplemented in ways that are theistic, naturalistic, or pantheistic -- is limited to claims that can be scientifically evaluated. Michael Behe clearly explains the limits of his "design only" claims, in a summary of ideas from pages 245-250 of his 1996 book, Darwin's Black Box:
Although I acknowledged that most people (including myself) will attribute the design to God -- based in part on other, non-scientific judgments they have made -- I did not claim that the biochemical evidence leads ineluctably to a conclusion about who the designer is. In fact, I directly said that, from a scientific point of view, the question remains open. In doing so I was not being coy, but only limiting my claims to what I think the evidence will support. To illustrate, Francis Crick has famously suggested that life on earth may have been deliberately seeded by space aliens (Crick and Orgel 1973). If Crick said he thought that the clotting cascade was designed by aliens, I could not point to a biochemical feature of that system to show he was wrong. The biochemical evidence strongly indicates design, but does not show who the designer was. {from Philosophical Objections to Intelligent Design: Response to Critics}
Here, Behe is explaining why -- even though his theory claiming that "design has occurred" can be scientifically evaluated -- he is not claiming "creation has occurred" because, based on the evidence he is considering, this claim cannot be scientifically evaluated.

Can a design theory be scientific?
What is science? In the past, attempts to define demarcation criteria -- which claim to capture "the essence of science" and therefore to clearly distinguish science from nonscience -- have not been satisfactory.
Despite these difficulties, some critics of design want to use methodological naturalism as a decisive demarcation test. They ask one question, "Does it violate MN?", and consider the case to be settled. But is it really that simple?

This section has examined the main arguments for enforcing MN and excluding design from science. Initially, these arguments may seem credible, but when we inspect more closely the logic is less impressive, and it seems less reasonable to exclude design from science.
Here is a brief review of the first part of the section: a Closed Science, restricted by MN, might guarantee that science will reach false conclusions; an Open Science allows, but cannot guarantee, reaching correct conclusions; a theory of design can be scientifically useful (because it might be true, it can promote a more accurate evaluation of non-design theories, and it can stimulate creative thinking and action by proponents of either design or non-design); design is not a science stopper (in the actual practice of science), should be evaluated based on what it is now (not what other theories were in the past), and should not be excluded by citing the status quo (by assuming that current methods are necessarily the best methods) or by the verbal equivocation of using "natural" with two different meanings; there is no incompatibility between science and occasional miracles; and since design does not require miracles, it does not even explicitly violate MN.

When examined closely, a few of these arguments (such as those about "natural science" being "natural" by definition) seem rather trivial. Other ideas, especially those encouraging us to think about the practical effects of an open science and a closed science, are more worthy of serious consideration, if only because they can stimulate productive thinking and interactions. Some concerns, involving scientific methods of theory testing and evaluation, seem especially important and interesting. These methodological concerns have been discussed in ten subsections: Observable Effects, Historical Science, Personal Agency, Supernatural Agency; Evidence and Testing, Testing and Falsifying, Design uses Scientific Logic; Mechanisms and Matching, Supplemented Theories, Design and Creation. Here is a summary of the main ideas:
because an unobservable cause can produce an observable effect, scientific theories can postulate the existence of a cause that cannot be directly observed;
in logical methodology and use of data, a design theory is similar to other theories in historical sciences; strategies for coping with the uncertainties of agency are similar, whether a personal agent is natural or supernatural;
the main evidences for design are observable "signs of design" that probably could not be produced by undirected natural process; since design and non-design are mutually exclusive, their plausibilities vary inversely, and testing for non-design is a way to test for design; although design cannot be proved or disproved using formal logic, scientists can develop a rationally justified confidence that design (or non-design) is the best currently available explanation, or is at least a reasonably plausible explanation; a theory of design is empirically responsive, and is therefore testable, if its evaluative status can be changed (up or down) by empirical data; when critics try to empirically refute a design theory, they are confirming that the theory is scientifically responsive and testable;
scientists who propose design use the conventional logic of MN-science, but question the wisdom of restricting science to the conclusions required by MN; in some areas, such as forensic science, methods for testing design theories have been developed and are commonly used;
a basic design theory does not propose a mechanism for the directed action that produced a designed feature, but a mechanism is not required for scientific utility; a design theory should be tested for what it claims (that design occurred), not for what it doesn't claim (that it can explain the details); a basic design theory can be supplemented (with details about when, how, why, who,...) in many ways (theistic, naturalistic, or pantheistic), and each theory (basic and supplemented) should be evaluated on its own merits; theories of design and creation can be related, but are not the same, and design does not necessarily imply creation; a basic design theory is limited to claims that can be scientifically evaluated.

Conclusion
Can a design theory be scientific? This section begins with reasons to say YES, and then explains why reasons for saying NO do not seem logically justified. The more closely we examine arguments for enforcing a rigid methodological naturalism in science, the better "design as science" looks. This suggests that, instead of using MN in a futile effort to separate science from nonscience, for each theory of design we should ask, "Is it scientifically plausible?" and "Is it scientifically useful?" These questions are explored in Sections 7B and 7D.
7B: Based on their analysis of observations, some scientists are wondering whether a theory of "intelligent design and design-directed action" might be a plausible explanation for the first life, complex life, or the universe itself. Even though proof is impossible, can we develop a scientifically justified confidence about the plausibility of a design theory?
7D: In our search for truth, is MN always helpful? Should we reject a theory of design before looking at the evidence, or should we consider the possibility that design is a potentially reasonable explanation that is worthy of further development? What are some benefits of an Open Science that has been liberated from the restrictions imposed by MN?

[]

7D. A Proposal for Open Science
( the title has been changed to "The Freedom of Open Science" )
Section 7B concludes that "the potential of design theories to make valuable scientific contributions should be recognized and welcomed." This section looks at some responses to design -- Has there been a warm, gracious, enthusiastic "welcome to our house" reception, or is the door being jealously guarded by zealous gatekeepers of knowledge? -- and explains why an Open Science would be scientifically and philosophically productive.

A Goal and Strategy, Problem and Solution
A Goal: A major goal of science is to search for truth. When we ask, "Has the history of the universe included both natural and non-natural causes?", we hope science will help us find the truth.
A Strategy: Science should answer this important question without doing any science.
A Problem: This strategy might force science to reach wrong conclusions.
A Solution: Change the strategy that is causing the problem.

According to a strategy of methodological naturalism (MN), science must conclude -- before the process of science begins -- that everything in the history of nature happened due to natural causes. But if some events have involved non-natural causes, some scientific conclusions are guaranteed to be wrong, yet there is no way to escape these false conclusions. Since we want science to help us in our search for truth, instead of leading to unavoidable false conclusions, this does not seem rational. {details}
The essential weakness of MN is logical inconsistency: MN claims to be logical, yet it demands that all scientists should always assume that some possibilities don't exist, even though logic demands that scientists should consider all possibilities. Logically, an observed feature might have been produced by any of three causes: undirected natural process, natural design and construction, or supernatural design and creation. Consistent with this logic, an Open Science recognizes all three possibilities. By contrast, a Closed Science, restricted by MN, demands that a theory of intelligent design (ID) must be excluded from science because it permits either natural design (allowed by MN) or supernatural design (not allowed by MN). In an Open Science, with intellectual freedom not restricted, a scientist is free to conclude that a particular feature was produced by either undirected natural process or design.

Should science always use the process of science?
In open science, a scientist can always use the process of science -- a logical analysis of observations -- to reach a scientific conclusion.
In closed science, a scientist -- restricted by a rigid MN -- must bypass the process of science and declare that, no matter what is being studied, "it happened by natural process." This conclusion is immediate (since we won't be delayed by a process of careful scientific investigation), automatic (it is logically unavoidable due to MN), and certain (if no debate or doubt is allowed by the enforcers of MN). Reaching a scientific conclusion without doing any science is certainly quick and efficient, but is it wise and effective? Or will our search for truth improve if a Closed Science is liberated to become an Open Science?

DESIGN THEORIES TO EX