I believe the reply button on the OP (the blue thing, at least for me, at the bottom of the page) just puts in a comment that has no connections. If you hit the reply button to a specific post (which to me at least is not blue), you do get a connection unless it’s currently the last post.
1 Like
Prof. Behe speaks of “the appearance of design” as though it is somehow a conclusive observation.
But this is generating a specific conclusion from a generality that hasn’t actually been proven.
It’s like the dozen proofs of God (or more) that are claimed to PROVE God, but most philosophers conclude that they all fall short as PROOF.
Arguments for the Existence of God
-
Introduction (2)
-
Religious Experience (10)
-
Pascal’s Wager (6)
-
Ontological Argument (10)
-
Teleological Argument
-
General (18)
-
Cosmology & Physics (37)
-
Biology & Chemistry (12)
-
-
Argument from Reason (17)
-
Moral Argument (24)
-
Trilemma (6)
-
Miscellaneous (6)
Still obsessed with PROOF, I see. This is not science.
How would you define proof in this sense? I thought maybe you were using it in some non-conventional sense.
As far as I know, it’s not been vogue for Philosophers to consider concepts of strict proof as very useful for quite some time. Certainly, scientists aren’t very fussed over the term because the whole framework behind scientific knowledge is that it is constituted by provisional statements of truth about the world.
Proof as a concept generally implies that something ‘proven’ is known to be true exhaustively. Proof may be applicable in the case of God and other philosophical issues, but I think it’s not popular even among philosophers because of the realisation that most philosophical/logical proofs are predicated on other philosophical positions, which themselves may or may not be acceptable/controversial.
Because of this, I haven’t known proof to be a useful concept outside of a strictly mathematical setting or for settling very particular logical problems.
Aren’t you merely re-stating my point?
I can say: We have proof that the recessive sickle cell trait is beneficial to humans at the population level of analysis.
Or I can say: We have teleological proof that God exists.
The first statement is much more defensible than the 2nd statement.
I.D. is in the same category of inquiry as the 2nd statement.
Yeah you can say that but you’d be using the word proof incorrectly. You might be using it in some colloquial sense where ‘proof’ means “have good evidence-based reasons for thinking”. But then to avoid confusion with logical or mathematical proof, you should just use the word ‘evidence’.
So you might say instead: We have good evidence-based reasons for thinking that recessive sickle cell trait is beneficial to humans at the population level of analysis.
I wondered what this new thing called “statistical induction” was. I supposed that I would need to learn it. But isn’t it just what we used to call “statistical inference”?
4 Likes
I guess you just like to argue about words?
OK, I like your description I quote. Now give me the equivalent description for I.D. where it is clear to the audience that I.D. is less compelling than how we understand Evolution?
I wouldnt be surprised!
I got the terminology from a Wiki article describing the nature of Inductive Reasoning.
We need to get comfortable with this terminology!
I don’t insist on the word “stronger”. But you need to propose or formulate your own phrasing… and I will concur with you.
Clearly one kind of argument tends to be MORE CERTAIN than the other.
If all you want to do is argue that both arguments can be equally lame, I think we can see that the spirit of this conversation is lost on you.
We dont clarify anything by trying to prove they are the same. We clarify matters by showing how the best examples of both categories are different in persuasive potency.
Thanks, but I’m already comfortable with the old terminology.
2 Likes
I would simply say, “Your opinions and beliefs about design are not scientific.”
For me, that is the major weakness in his argument if we are focusing on science. “It looks designed” is not an empirical measurement. It is not a scientific observation. The appearance of design is a subjective human opinion. That’s not science.
A non-scientist will say that a rock looks heavy. A scientist will measure the rock’s mass. That’s the difference.
3 Likes
Statistical Induction is NOT a proof, even in the sense of probability. It is evidence against the Null hypothesis (in the statistical Frequentist sense) or a comparison of the likelihood of two hypotheses (in the Bayesian sense). A Frequentist hypothesis could give evidence against evolution (reject the null), and suggest an alternative hypothesis should be considered. Bayesian methods require both hypotheses to be defined clearly.
AND never in the history of ID has the Designer been defined in a way we can form hypotheses about the designer.
That said, causality can be shown by proper xperimental design, and statistical induction can give evidence suggesting proof. We do not show causal effect with observational data.
(Except we do, when natural experiments can be found, but Causal Inference in Statistics is not our topic today.)
* In the simplest form: two groups, treatment and control, with pre-and-post treatment measurements.
2 Likes
Also,
The point being, when you find a correlation, or a “significant difference”, you might suspect a causal relation. That’s NOT proof, but it might inspire you to roll your sleeves and go looking for more evidence.
4 Likes
Certainly I agree it is not a proof.
But I dont think it is even conclusive against the Null Hypothesis.
YECs would LOVE to hear that it is effective with countering NULL. So maybe you can re-think exactly how you want to describe Behe’s use of the term INDUCTIVE!
Your cartoon and discussion is the best Ive heard so far. Maybe we can smooth it out into a coherent counter-argument to Behe’s self-confessed use of induction.
I am certainly not going to pretend to lecture a biostatistician on stats, but there are a few lessons I have learned over the years.
First, hypotheses made after an experiment a much weaker than hypotheses made before an experiment. With big enough data sets you can find a statistically significant correlations, including false ones. If you are forming your hypotheses after you gather the data, then you have, at best, found candidates for further testing. I have chased after weak correlations from big RNA expression data sets only to see the correlation fizzle out in focused experiments. At the same time, other correlations have panned out.
The secret sauce to good science is independent lines of evidence. If different methods and approaches arrive at the same correlation then you are on much, much firmer grounds. You are effectively multiplying probabilities which greatly reduces the chance of false correlations, if you do it right.