|
Author
|
Topic: Is ID theory falsifiable?
|
Christopher D. Beling
Member
Member # 723
|
posted 21. November 2005 11:00
In a recent news article Teach our children well by Dr. T. Huffman (M.D.) that takes quite an anti-ID in school science class stance there was just one statement that challenged me. It was this
quote:
ID is not a scientific theory, for there are no set of facts that would disprove it. It cannot even be called a hypothesis since it cannot be tested.
So two things are being claimed here, namely that ID cannot either (i) take theory status or (ii) take hypothesis status.
I have some ideas on this subject. For example it seems to me that ID predicts that bio-information can only decrease with time (in contrast to neo-Darwinian theory that predicts a steady increase)- and such opens itself to possible falsification. There is also the idea from the philosophy of science that science works more by verification of hypotheses than by negation of wrong hypotheses - for example William Paley's design argument was never negated it was just superseded by what seemed at the time like a better theory.
I am interested in hearing others thoughts and insights on this issue - specifically:
(i) Is there common agreement on the rudimentary elements of what constitutes ID-theory?
(ii) If yes are any of these elements falsifiable?
(iii) Why is it that we talk about Darwinian or ID-theories rather than the Darwinian or ID hypotheses? Are we just using the word "theory" to mean "all-embracing hypothesis"?
(iv) Is falsifiability that important anyway?
If these things have been discussed previously I thank you for pointing to the relevant threads.
-Chris
[ 21. November 2005, 11:03: Message edited by: Christopher D. Beling ]
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 22. November 2005 14:36
A framework establishes a threshold for a design inferrence. Find a natural object that meets the criterian of the framework, and the framework is falsified.
IP: Logged
|
|
David L. Hagen
Member
Member # 323
|
posted 22. November 2005 16:27
Falsifiability or Practical Explanatory Power?
Good questions Chris. Why Falsifiability? Can the scientific basis or worth of origin theories actually be evaluated by falsifiability? (Contrast nomological(law like) theories such as gravitation, electromagnetism, and relativity.)
I highly recommend:
Stephen Meyer, The Scientific Status of Intelligent Design
The Methodological Equivalence of Naturalistic and Non-Naturalistic Origins Theories.
e.g., Meyer observes: "Other semantic criteria emerged, such as Sir Karl Popper's falsifiability. According to Popper, scientific theories were more meaningful than nonscientific ideas because they referred only to empirically falsifiable entities."
"In 1974, Sir Karl Popper declared neo-Darwinian evolutionary theory “untestable” and classified it as a “metaphysical research programme.” . . .
"Attempts to locate methodological “invariants” that provide a set of necessary and sufficient conditions for distinguishing true science from pseudoscience have failed.10 Most philosophers of science now recognize that neither verifiability, nor testability (nor falsifiability), nor the use of lawlike explanation (nor any other criterion) can suffice to define scientific practice." . . .
"Darwinists have long defended the apparently unfalsifiable nature of their theoretical claims by reminding critics that many of the creative processes to which they refer occur at rates too slow to observe."
Instead of demarcation criteria such as "falsifiability", Meyer proposes:
"A rational historical biology must not only address the question
"Which materialistic or naturalistic evolutionary scenario provides the most adequate explanation of biological complexity?"
but also the question
"Does a strictly materialistic evolutionary scenario or one involving intelligent agency or some other theory best explain the origin of biological complexity, given all relevant evidence?"
IP: Logged
|
|
Christopher D. Beling
Member
Member # 723
|
posted 22. November 2005 19:45
Irving, when you say that "a framework establishes a threshold for a design inference" I think you are refering to the complexity-specification criterion (CSC) (as normally represented in the explanatory filter)?
So I see what you are saying: if an object which was KNOWN to have been produced by totally naturally causes (totally described by natural law equations) and yet this object scored a positive on applying the CSC then it would indicate a definite falsification of ID theory. Am I following you ? - if so I think I agree.
One concern though, is that there still is a very small and incredibly unlikely probability of some functional specification happening by chance - is this not why we talk of a design inference rather than a design proof? When one takes this into account are we still looking at strict falsifiability?
-Chris
[ 22. November 2005, 19:51: Message edited by: Christopher D. Beling ]
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 22. November 2005 20:50
Yes, you have it.
Consider that there is a spectrum of definitely Natural to definitely Designed at either ends. Independently Specified Complexity builds as one travels from the Natural end to the Design end. At some point along the continuim, say the univeral probability bound, ID theory claims a tipping point is reached in which the Design inferrence takes precedence. Considering the allowances for error ID theory takes into account in establishing such a probability bound, any KNOWN natural object meeting the criterian would falsify that point as a demarcation line.
In Popper speak, then one could say that ID "prohibits" natural objects from achieving CSC above the point established.
IP: Logged
|
|
Melvin H. Fox
Member
Member # 1684
|
posted 23. November 2005 07:25
Irving,
Suppose event A has probability p < UPB. Suppose further that event A is shown to occur naturally by chance [however that would be done]. How would this falsify the demarcation line [UPB]? The line only marks where it becomes more likely that the event was designed. Does this not mean there may exist events with probabilities less than UPB that occur naturally? If so, should the discovery of event A be surprising? Further, if no surprise how is the inference falsified?
-Mel
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 23. November 2005 08:31
Let's differentiate where we're at with CSC, from a framework concept as a whole. With CSC design is not definitive, but inferred, thus falsification is not definitive, but inferred. Or in other words, the falsification is equivilant to the degree of inferrence.
However, in a larger construct (perhaps say IC), establish a definitive design framework, and any natural object meeting such framework would falisify the framework.
IP: Logged
|
|
Terry Mullett
Member
Member # 1810
|
posted 24. November 2005 00:47
I think the distinction between natural and intelligent needs some work. It is based on two assumptions, both of them weak.
One assumption is that "natural" is somehow Platonically fixed to include certain kinds of explanations. Certainly there is a prevailing idea of what is natural, but even that would only be a statistical mean, and nothing in principle prevents it from being mutable. Clamping down nature could serve more than one program. It could be used to exclude ideas that the current establishment doesn't like, but it could as easily be used to define a straw man.
The second assumption, partially a consequent of the first, is that intelligence could not, in principle, be characterized in natural terms. I think this needs some rigorous justification, since it seems to me that it is at the heart of claims that design theory repudiates naturalism. And that justification requires saying something clear about what intelligence is. I don't see this. I've seen a couple of shots at what intelligence isn't, but really that's a big so-what. It doesn't have to be exhaustive. I'm not suggesting we ask the designer for two forms of photo id. But an intuitive "folk psychology" understanding of intelligence won't do, and neither will ignoring prior work on the matter. As Yogi Berra said, "If you don't know where you're going, you'll wind up somewhere else." Seriously, loose examples of IC and CSI may get the foot in the door, but when it's time to show them your money, intelligence is your money.
It also seems to me that any formulation of nature which does not aim to include intelligence does little more than prolong of our cognitive dissonance. I say this believing "natural" not to be Platonically fixed. I see it as a cultural construct, serving the purpose of distinguishing those matters over which we can confidently exercise some intellectual control from those which tend to leave us inarticulate. The latter class indicates the current boundary of a natural understanding. The boundary is not fixed: it is different for different people, and it has a history and a future. Treating current limits as intrinsic limits causes tedious and unnecessary thrashing for both sides of this debate. I hate to be cynical, but I suspect there are people on both sides who are invested in keeping it that way.
I think some version of falsifiablity is important in a pragmatic more than a technical sense. If there is truly no condition which could show that a statement is false, then nothing would change if it were false. In that case it may as well be false. Really, life's too short.
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 24. November 2005 10:11
Isn't that the point?
Natural processes do not design to external (non-local) requirements. An example of a natural processes designing to an environment not-local to the production would falsify the construct.
IP: Logged
|
|
Anthony Rose
Member
Member # 1822
|
posted 25. November 2005 21:08
Somewhat new to ID myself, but have read up a bit and have a BioChemistry degere, so I'll chip in anyway
"(i) Is there common agreement on the rudimentary elements of what constitutes ID-theory?"
I think so. I'd tentatively suggest that ID is an umbrella term for several arguments -based on analysis of observations- with this joint conclusion: that scientifically speaking some objects exist that could not by any means have developed by chance under the physical laws we know.
The term first came up in a 1988 conference called "Sources of Information Content in DNA", rather than being a single published theory.
"(ii) If yes are any of these elements falsifiable?"
First, all rational arguments are inherently falsifiable in theory, but we just don't always know enough yet to be able to achieve that in practise; that doesn't make it non-scientific. And it never used to. There are quite a few examples of scientific theories which have not been falsifiable and were not rejected by the community on that basis. (Eg see Noble prize winner physicist Lederman's book "The God Particle" p349).
This is a new 'bar', raised up specifically against ID. It is fairly common for an established scientific community to protect its investment in long-held theories and positions by attacking new theories, including by this method. (See "The Structure of Scientific Revolutions" by Thomas S. Kuhn punlished well before ID, or read http://www.des.emory.edu/mfp/Kuhn.html).
Second, I haven't researched it well yet but
believe that they are falsifiable, for the simple reason that irreducible complexity is: all it would take would be an experiment on natural processes producing a structure with a specific and irreducible complexity which has a statistically low enough chance of occurring that matches an example given by ID. (Note I think this is a less likely chance than the standard defined by science to be impossible by chance but I forget what those odds are).
For example, if the majority of amino acids could be produced in a chamber generating a credible state of the Earth as it was at the required time, this would be a step towards falsification.
(Actually, the biggest problem here is that one would really have to show at a minimum the same for the basic irreducible mechanism of all life: the generation of nucleotides, and their joining together to form genetic material; plus the amino acids joining to make proteins; which would have to be able to reproduce the genetic material; which would itself have to be functional to be naturally selected; in a protected eg cellular environment. So IMHO ID is not falsifiable, though it is verifiable!)
"(iii) Why is it that we talk about Darwinian or ID-theories rather than the Darwinian or ID hypotheses? Are we just using the word "theory" to mean "all-embracing hypothesis"? "
This debate is rather meaningless and really is only fired by the creationists calling evolution 'only a theory', meaning not proven (actually a matter of opinion, of confidence in it, at this level).
An hypothesis is a postulated idea which claims no evidence. 'What if....', 'Maybe', 'Could it be that'.
A theory is a matured hypothesis. It has supporting evidence and argument. 'We think so because', 'We're pretty sure since', 'We've no reason to doubt'.
Given enough time and further extensions of knowledge a surviving theory can become well enough accepted to be considered a fact (eg mass causes gravity - did you know by the way that it might not be the mass?!)
So theory is actually the perfectly correct word for both evolution and ID.
"(iv) Is falsifiability that important anyway?"
No - see answer to (ii)
Also, simplistic though it may seem I think it is worth remembering that a true statement cannot be falsified. It can be verified though.
And verification is basically whether it matches the facts.
For example, the Cambrian period saw many organisms arrive on the scene suddenly, too rapidly to leave traces of precursor organisms (if any); and too rapidly to be explained by mutation or any other known natural process. This is one verification of ID.
I hope this helps.
IP: Logged
|
|
Melvin H. Fox
Member
Member # 1684
|
posted 27. November 2005 19:31
The zero tolerance for false positives established for CSC is self imposed. Dembski’s metaphor for CSC is a net. “On the other hand, if things end up in the net that are not designed, the criterion will be worthless.” [Intelligent Design page 141] CSC is a statistical model based on probabilities. The theory of these models is already widely accepted. Also accepted is the notion that the confidence we have in the inference is never absolute. Therefore, a false positive, detecting design where there is none, does not in itself statistically falsify CSC.
The converse, Dembski’s inductive argument that CSC is perfectly reliable because it has yet to be shown to give a false positive [Intelligent Design page 142], is equally flawed. Probability is employed precisely when it is necessary to predict the unpredictable. To my knowledge, which is limited, there is no statistical method by which inference is made without error. If this is untrue, please correct me. If we are certain, probability theory is unwanted baggage. If we are unsure, we can only be confident to a degree.
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 28. November 2005 02:08
The theory, is that the probability of a false positive is so small, that it is meaningless.
A false positive falsifies THAT theory.
IP: Logged
|
|
kyle7
Member
Member # 191
|
posted 28. November 2005 02:37
Melvin,
ID is very similar to the Second Law of Thermodynamics. Experiments showed that entropy always increases for a given process. To prove this, we would have to examine all processes, which is impossible. It can be inferred that since we see all experiments supporting this, that the increasing entropy principle is true.
You state the following:
quote:
The converse, Dembski’s inductive argument that CSC is perfectly reliable because it has yet to be shown to give a false positive [Intelligent Design page 142], is equally flawed.
Why is this "equally flawed" if we accept the Second Law of Thermodynamics on a similar inference? The heart of the Second Law is based upon probability!
I would also state that the Second Law of Thermodynamics supports ID. The law is related to the order of systems. You don't find closed systems increasing in order on their own. Even open systems are limited in the ability to increase their order based upon the boundary conditions. This is why you don't see a sand castle form when sand is poured out on the beach. Even though there are millions of ordered configurations that the sand could form to make a sand castle, there are far more unordered configurations that make no sand castle. In order to get the sand castle, you need a machine that uses energy to move sand. Also, intelligence is needed to decide the pattern of the sand castle. Wild application of energy will produce nothing. You need intelligent application of energy, whether that intelligence is built into the machine or supplied by the operator.
Now all of this has application to ID. Natural processes are the most probable. The physics of systems naturally impose limits on the expectations of patterns. When a pattern has intelligence built within it to the requirements stated by Dembski, then one knows it is designed. For example, if a signal from space contained the first hundred prime numbers then design can be shown. The probability of seeing the first hundred prime numbers sequentially is less than the upper probability bound (1E-150).
The sand castle can also be shown to be designed if the shape is of significant conplexity that the probability of it happening on its own is less than the upper probability bound. This probability is based on the physics of sand. The Second Law of Thermodynamics does not allow the increase in order of the system away from that expected by the physics of the open system.
[ 28. November 2005, 02:48: Message edited by: kyle7 ]
IP: Logged
|
|
Melvin H. Fox
Member
Member # 1684
|
posted 29. November 2005 09:34
Well said gentlemen. I agree with Irving that the probability is meaningless. How is that not, just my opinion? Also, perhaps flawed is too strong a word. I guess my exception is with using logical induction.
Let me soften my position, while I raise again my concerns. From the materialist perspective, isn’t showing the probability to be meaningless the key to the whole ID argument? Can it be shown or is it subjective to the last? If no probability is meaningless, then chance evolution is possible?
In logical induction the premises only provide some degree of support for the conclusion. We may be 99% or 90% or 70% or it seems for ID better than 50-50 is good enough. Don’t get me wrong, CSC is a marvelous invention but are we justified to say it is perfectly reliable?
Suppose we are materialists. Our argument would be:
Premise – If false positives exist, then CSC is worthless.
Premise – If CSC is worthless, then ID is false.
Premise – False positives exist.
Conclusion – ID is false.
We would set up our null and alternative hypotheses as:
H0: No false positives exist.
H1: False positives exist.
Now, the only thing required to bring ID to its knees is one false positive. Keep in mind that CSC infers design for better than 50-50 odds. If no false positive is found, we fail to reject H0 but we have not proven the null hypothesis to be true.
As far as the Second Law of Thermodynamics is concerned, Kyle makes a good point. I had not considered this. Does the heart of the 2nd Law beat to the tune of probability? I must research more. From what I have read so far, the most fundamental particles in the physical world are understood through probability distributions. It is no quantum leap to then say they are driven by probability distributions. Automotive insurance companies infer the expected size of claims arising form a policy using probability distributions. There has never been a car accident caused by a probability distribution; unless the driver was analyzing one while he was driving. Why are we so quick to attribute physical properties we do not understand [can not understand?] to chance? Can chance DO anything?
-Mel
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 01. December 2005 21:49
In theory falsification, how does the failure of falsification prove the null hypothesis?
IP: Logged
|
|
|
Printer-friendly view of this topic