The Failures of Mathematical Anti-Evolutionism
part of the discussion precisely as Dembski presents it, and then we
Download 0.99 Mb. Pdf ko'rish
|
The Failures of Mathematical Anti-Evolutionism (Jason Rosenhouse) (z-lib.org)
part of the discussion precisely as Dembski presents it, and then we will discuss what it means: Given a reference class of possibilities , a chance hypothesis H, a probability measure induced by H and defined on (i. e. P( · | H)), and an event/sample E from ; a rejection function f is detachable from E if and only if a subject possesses background knowledge K that is conditionally independent of E (i. e. P(E | H&K) = P(E|H)) and such that K explicitly and univocally identifies the function f. Any rejection region R of the form T γ = {ω ∈ | f(ω) ≥ γ } or 142 5 probability theory T δ = {ω ∈ | f(ω) ≤ δ} is then said to be detachable from E as well. Furthermore, R is then called a specification of E and E is said to be specified. (Dembski 2002, 62–63) This is written very much in the style of track two mathematical reasoning, so let us now try to translate this back into track one. We can illustrate the idea with a classic experiment described by Ronald Fisher in his 1935 book The Design of Experiments (Fisher 1935). Apparently Fisher attended a party at which people were drink- ing tea with milk. One of the women at the party claimed that she could distinguish, by taste, a cup in which the tea was poured into the cup before the milk from one where the milk was poured before the tea. Fisher was skeptical and quickly designed an experiment to test this claim. Eight cups of tea were prepared, half with the milk poured first and half with the tea poured first. The woman was chal- lenged to identify the four cups in which the milk was poured first. The question is this: how many cups does she need to get right for us to conclude that she is not just getting lucky? Does she have to get all four? Or would we be impressed if she only got three or two of them correct? Fisher reasoned that the key was working out the probability of getting a particular result by chance. If that probability is below a certain threshold, then we conclude that the result is not due to random guessing. Fisher took 5% as his threshold. In other words, he decided ahead of time that if the probability of getting a given number of correct answers by random guessing was below 5%, then he would conclude that the woman really had the ability she claimed to have. In this context, the probability calculations are not too difficult. In fact, they can be carried out using only a small variation on the counting techniques we discussed in Section 5.4. Fisher found that the probability of getting all four right by random guessing was roughly 1.4%, below the threshold. On the other hand, the probability of getting at least three of them correct was roughly 24%. This probability is too high for us to discard random guessing as a viable 5.7 is the flagellum complex and specified? 143 hypothesis. The 5% threshold was arbitrary, but it seems to work well enough in practice. We can quickly generalize Fisher’s approach. Speaking infor- mally, suppose we have collected some data, and we want to know if something other than pure chance is at work. We start by putting forth the “null hypothesis” that nothing beyond chance is happening. We then work out the probability of obtaining the data under the null hypothesis. If the probability is below a certain threshold, typically taken to be 5% or 1% in modern statistical practice (at least in the social sciences), then we dismiss the null hypothesis as not credible. In statistics jargon, probabilities below the threshold are said to fall into the “rejection region.” Applying this method requires that we have a grasp on the appropriate probability distribution to use, but this is possible in many practical situations. Return now to our very technical quotation of a few paragraphs ago. Dembski’s intent is to provide a rigorous mathematical foun- dation for his notion of “specification.” He attempts this by adding a minor gloss to Fisher’s approach to hypothesis testing. The first Download 0.99 Mb. Do'stlaringiz bilan baham: |
Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling
ma'muriyatiga murojaat qiling