A well-accepted characteristic of a scientific hypothesis is that it must generate predictions about the world against which tests can be run – confirming or falsifying those predictions and thus supporting or not supporting the hypothesis in question. Understandably then, intelligent design proponents need to demonstrate that ID can produce predictions if it is to be taken seriously as a scientific hypothesis and alternative to evolutionary theory.
Historically (as much as I can say “historically” in the context of such a new movement), this has not been a major goal of ID proponents. Most of their previous efforts have gone into either arguing against evolution (see Michael Behe’s Darwin’s Black Box) or trying to justify ID in a pseudomathematicophilosophical1 way, bypassing the normal scientific process (see William Dembski’s design inference/explanatory filter). Neither of these routes have been very successful, as much as ID proponents will try to tell you otherwise, and combined with a substantial amount of scientific and philosophical attack against the ideas of the ID movement, they have begun to change tactics a little bit, and talk about predictions and testability has crept back into discussions about ID, both online and in books.
Stephen C. Meyer’s epic tome Signature in the Cell, which I am currently finishing reading and formulating a series of posts about, is one such book, and contains an appendix where predictions of ID are the focus. The reasons as to why that particular topic couldn’t be worked into the main body of the book still eludes me, as I see it as on par with, if not more important than, the other topics touched upon. Did Meyer think that his audience would not care to read about it, or does he himself not care? Either way, it’s clear that he didn’t find it of very much importance.
But the predictions in Signature is not the focus of this post – I’ll deal with it separately: instead I want to talk about a recent post on Uncommon Descent by Jonathan M, entitled, appropriately, “Does ID Make Testable Scientific Predictions?“, which is basically a “brainstormed” list of fourteen predictions that he thinks ID makes about the world. He notes that, as the post was jotted down rather hastily, the list is probably non-exhaustive and there may be other possible predictions out there, but judging by what I’ve seen of other such lists both online and in Signature, it covers pretty much all the same points as they do. If more predictions exist, nobody in the ID movement has thought of them yet.
So, what are these predictions that Jonathan puts forward, and how valid are they? As will hopefully become clear fairly quickly, predictions need to be grounded within the hypothesis and must flow from it naturally if they are to be productive and allow for a way to test the hypothesis they are being linked to – and ID proponents seem to have a hard time doing this.
(Note that I won’t be touching on the two astronomical/cosmological predictions Jonathan gives, as they don’t have anything to do with biological intelligent design, which is what I focus on. Cosmological ID and biological ID are not related and shouldn’t be conflated, as doing so makes things confusing.)
ID predicts the presence of specified complexity in living systems.
This is almost a non-prediction, as it’s probably impossible for life to exist without being both specified and complex in some senses of the words. It’s analogous to trying to explain the origin of a rock and saying that your hypothesis predicts that the rock is made of atoms – it’s already assumed and is unhelpful as a test of that hypothesis. If this turns out to be the only “valid” prediction in this list, it shows that ID has virtually non-existent predictive power.
ID predicts that, as scientific research progresses, biological complexity will be seen to increase over time, and information will have a more and more central role in the governing of life’s operations.
This, again, is a fairly basic semi-prediction, but it’s the first one that makes me ask: why? Why does ID predict this? As I mentioned in a recent TWiID involving Casey Luskin’s attempt at some ID predictions, if an analogy between human design and “ID” design is being made2, it’s fairly clear that humans, as intelligent agents, don’t always make things as complicated as possible or increase the complexity of objects in every circumstance. Of course, it’s possible that Jonathan isn’t making an analogy to human design, but if he isn’t – how is he justifying this prediction? Where is it coming from?
ID predicts an increase in evidence for the non-adequacy of the DNA-centric view of living systems.
I didn’t know that there was a non-DNA-centric view of living systems to compete with genetics3, but even if there was, why does ID predict that it is accurate? Again, this isn’t justified.
ID predicts that complex molecular convergence will happen routinely.
Again, why? What part of the ID hypothesis lead Jonathan to this conclusion?
ID predicts the presence of irreducible complexity with respect to macromolecular systems and organelles.
I’m feeling like a broken recor- recor- recor- record, but again: what part of ID predicts this?
Okay, it’s possible that he’s drawing upon an argumentative strategy employed by Stephen C. Meyer in Signature in the Cell – inference to the best explanation (IBE). Under this strategy, evidence against evolutionary theory becomes evidence for ID, due to a current lack of other hypotheses to oppose ID, making it the “best explanation”. It conveniently allows for the vagueness of ID to work in its favour: since it is hypothetically able to explain any data set, any time the prevailing scientific ideas fail, ID can sneak in there and plug the gaps, exclaiming that it explains the data better than no explanation at all. It’s not hard to see why IBE isn’t a great argument.
Applying IBE to irreducible complexity yields that, if evolutionary theory cannot explain it and ID can, it is evidence of ID – in a sense. The problems with this argument are numerous and I’ll touch upon them later in my posts about Signature, which revolves around IBE as the single method of supporting ID as a scientific hypothesis. For now, I’d much prefer Jonathan justify this particular prediction without mentioning IBE: if it’s possible.
ID predicts that the prevalence of functional protein folds with respect to combinatorial sequence space will be extremely small.
This prediction is related to the irreducible complexity one above, in that it could be invoking IBE as its justification. But if not, it doesn’t fare any better. What Jonathan is referring to is the work of Douglas Axe, a pro-ID biochemist who has conducted experiments that, he and the ID movement claims, prove that protein evolution is impossible. Thinking clearly about this for a while leads one to a single question: irrespective of the legitimacy of Axe’s conclusions about his research, why would ID predict that protein evolution is impossible? Life could have been designed by a Designer with the ability to evolve – this possibility is at least as probable as the opposite scenario, that life was designed not to evolve. But without any knowledge about the Designer, how can one scenario be given more probabilistic weight than the other? This prediction, like the others above, doesn’t follow from ID.
ID predicts that evolutionary pathways to new protein functions will require multiple co-ordinated non-adaptive mutations (more so than likely to be achieved by a random process).
This is virtually the same prediction as the previous one: protein evolution is impossible. As such, it does not follow from ID.
ID predicts that DNA, which was once considered to be junk, will turn out to be functional after all.
Another appeal to the human design analogy? Possibly. If so – and I also mentioned this whilst addressing Casey Luskin’s predictions – you have to consider that human designs are not perfect, flawless things. They contain compromises and non-functionality all the time, and these characteristics are often the product of laziness and factors unrelated to the functional goal of the design process, such as economics, politics and social considerations. If ID is being modeled on what humans frequently do, the fact that junk DNA could have been part of the original design (or accumulated naturally over time due to oversights in the original design) is not outlandish. Where does that leave this prediction?
ID predicts delicate optimisation and fine-tuning with respect to many features associated with biological systems.
“Optimisation” and “fine-tuning” are words that could mean a lot of things and Jonathan doesn’t give them a rigorous definition. They could, in fact, under many definitions, be as useless as “specified complexity” when talking about predictions, as biological systems like metabolic pathways must be “fine-tuned” to turn specific molecules into other specific molecules. Could life exist without this definition of “fine-tuning”? Probably not. If so, this is hardly a useful prediction. But it remains to be seen what definitions Jonathan was using, so I’ll leave it there.
ID predicts that organisms will exhibit in-built systems which promote evolvability (e.g. front loading).
Why would this be the case? Again, no justification.
ID predicts the observed pattern of the fossil record whereby morphological disparity precedes diversity.
Is this a reference to the Cambrian explosion/radiation? Probably. But why would ID predict this? If human design is the analogy, we don’t always design in this way. Sometimes designs can come from the same source and share a common design ancestor, and gradually diverge in function over time. An random example off the top of my head would be the evolution of house plans: thousands of years ago, houses (more accurately called “structures”) would have had only one living space, combining the necessary functions for living into that space, such as sleeping and cooking. Over time, more rooms appeared as building techniques improved and the single room divided into multiple, separating sleeping, cooking, eating, entertainment, etc. Those multiple rooms were not always there in the design of the house, they came later.
Of course, designs do also happen where new things are completely added from scratch, but to claim that all design must have proceeded in that way is dishonest or just plain not thought out.
This all assumes that an analogy to human design is being made, of course, something that Jonathan never clarified.
ID predicts saltational, or abrupt, appearance of new life forms without transitional precursors.
This is so closely related to the prediction above that no further comment should be needed, I think.
And that’s it! All twelve (biologically relevant) predictions. As you’ve seen, none of them have been especially helpful in testing ID, either because they haven’t been sufficiently justified or they aren’t very useful predictions. While Jonathan’s list was self-described as not necessarily being all of the possible predictions of ID, what was there was pretty representative of the consensus of ID proponent-thought on the matter, in my experience. In order to be taken even the slightest bit seriously as a possible scientific hypothesis, ID needs to justify these predictions or generate new ones that are also justified. Whether or not this will be done remains to be seen.
- - - - - - - - -
- Let it be known that I coined this soon-to-be-widespread word in this very blog post. ↩
- An analogy that has never been justified, just to let you know. Analogy without justification is easily and legitimately labeled an argument from analogy, a logical fallacy. ↩
- Could he be talking about epigenetics? If so, that’s a misunderstanding of the implications of the field. ↩
Alright, ID proponents: start testing your predictions!
By the way, to pick one example, if ID predicts saltational, or abrupt, appearance of new life forms in the fossil record without transitional precursors, then this prediction has failed.
IMHO, many of the above predictions do not differ from those of standard (modern, up to date) Darwinian theory. Testing predictions is no use, if your predictions do not make a difference.
But surely some ID claims can be tested and tests clearly refute ID. Let's take Behe's irreducible complexity and mousetraps, for example. He really makes a triple-jump prediction:
Premise: Mousetraps (the flat snap trap type) are irreducibly compex.
1. Prediction: Therefore they can have no precursor with one part less.
2. Prediction: Therefore, they cannot have a continuous history.
3. Prediction: Therefore, they cannot evolve.
While each part of that triple jump prediction has been refuted here (http://www.springerlink.com/content/k6pv4mn47t65154x/), let me just point out that the seminal and patented precursor of all current flat snap mousetraps had one part less, because spring and striker were formed of one wire.
Likewise, independently specified complexity could be refuted by showing that it does not exist in things that are supposed to be due to a non-human intelligence. Take the bible for an example, where even atheist have to admit that the majority opinion is that it's due to a non-human intelligence. It has not been speciefied independently of the belief in question. It did not fall from heaven nor was it wahsed up from outer space. It was written by believers in the faith. Hence its information was not specfied independent of the belief in question.
I think the same can be done with bio-molecular traits. How, for example, is the information in DNA specifying a trait to be conceived as independent of that trait? Translation of genetic information specifies the trait and transmission through replication during reproduction specifies the genetic information of the offspring. Where's the indepencent specification? Natural selection closed the loop.
My hunch is that whenever ID predicts something differing from modern evolutionary theory, it will be shown to be false.
"Applying IBE to irreducible complexity yields that, if evolutionary theory cannot explain it and ID can, it is evidence of ID – in a sense."
Except that Herman Muller, in 1918, predicted that evolution would produce irreducible complexity (he called it "interlocking complexity"), so IBE doesn't help the IDers here.
http://www.talkorigins.org/origins/postmonth/sep0…
These predictions seem to be reverse-engineered. Irreducible complexity was once the criterion through which Behe tried to infer intelligent design by definition. Now, ID is supposed to predicts irreducible complexity.
As John Pieret pointed out, other theories including Herman Muller's evolutionary theory predict the same.
ID predicts that I ate pork chops for dinner tonight.
Oh wow! ID really does make accurate predictions!
An Intelligent Digester! Fancy that! :)
[...] posts to wade through. Not a lot I want to mention in this intro, particularly, except for perhaps this recent post of mine responding to an Uncommon Descent post about ID’s supposed scientific predictions. It was [...]
The flaw in the IBE argument is the following:
Unless the alleged designer and his abilities have been observed, or can be explained/predicted from established basic principles, ID is not an actual explanation. The designer plays the role of a Deus ex Machina in comedy – who can resolve any arbitrarily complex situation.
Evolution is both a factual claim—that life evolved—and a theoretical claim—how life evolved. ID has never got beyond a factual claim—that life was designed. It does not even purport to have a theory or mechanism as to how life was designed.
Theories are what enable scientifically useful predictions to be made. As long as ID has no theory, it cannot predict anything useful, and thus fails as science. Regardless of whether the fact of design may be true or false.
For simplicity, here's a scenarion under which scientists would convert to ID.
1. A surprising overservation/discovery occurs. For example, junk DNA is found to contain a signature such as: "Made by Intelligent Designer Inc."
2. The inferences to the best explanations (nowadays referred to as hypotheses) include ID. They would, of course, also include alternative hypothesis such as that the message has been put into the junk DNA by a fraudulent researcher, or that it may be due tosome artefact of method, etc. the alternative hypothesis to the Intelligent Designer.
3. Predictions derived from the alternative hypothesis clearly differ. In our example, the ID prediction would be that equal messages will be found in the junk DNA of all organisms and by all research groups irrespective of the researchers ideological bent, their methods etc.
4. Discoveries of signatures accumulate over decades.
Slartibardfast’s signature on a Norwegian glacier would do it for me!
The 'junkDNA' prediction is especially bothersome to me. I had an email exchange with John West on this very issue a year ago. I asked him who had made this prediction and when did they do so. He claimed Dembski did it in 1998 and that Mims had a letter to Science rejected a few years earlier that made the same basic 'prediction.
So I pointed out that as early as 1975, evolutionists had been speculating/predicting functions in and for junkDNA and finding it. West's retort was irrelevant to this, claiming that these researchers did this work "in spite of" the dogma of the day. He had no evidence for this, of course.
IDCs appear to take things already known and declare that they had made ID-based predictions about it. It is this utter dishonesty that annoys me.
I think you're missing the most important characteristic of scientific predictions—that they must be empirical. IOW, we do all of the interpreting to get the empirical prediction, which is confined to what we can actually observe, not what we will name it, etc.
In light of that, all of Meyer's predictions are unscientific because they give him plenty of wiggle room to redefine terms.