Darwin on Reason and Imagination
Readworthy (on Tim Burke's open letter at Cliopatria)
A Small Contribution to the Eventual Resolution of the Evolution Dispute
None of these really discuss 'intelligent design theory' itself, though. I have intended to do a post, for some time now, on how not to argue with IDers, but I never got around to it. To my comments in the most recent post above, I want to add two things:
1) It is a mistake to think ID can be refuted by labeling it as a "God of the gaps" argument; this is to mistake cause and effect. An argument can only be shown to be a 'God of the gaps' argument if the 'gap' has already been completely filled with a better candidate than God - in other words, nothing can be non-question-beggingly labeled a 'God of the gaps' argument until it has already been refuted. If it were to turn out, for instance, that the gap could never be filled by any alternative to God, that exact same argument would just be a straightforward (defeasible, but undefeated) causal inference. What makes it a 'God of the gaps' argument is that God's role gives way to a gap-filler. (The problem with 'God of the gaps' arguments is that they confuse primary and secondary causation, and treat a secondary-causation issue as if it were a primary-causation issue.) Of course, it is entirely reasonable to say, "My own view is that this or that ID argument will turn out to be a 'God of the gaps' argument or a 'designer of the gaps' argument, for such-and-such reasons," and, indeed, this is my own view of a large number of ID arguments; but what is doing the work here is the set of reasons for thinking that the gap will be filled, not its being a 'God of the gaps' argument. Likewise, it is a mistake to criticize ID by saying it is based on pessimism about the future of science, i.e., that it thinks science can make no progress on a given subject; precisely the problem with ID is that it is not structured this way. There can be entirely reasonable grounds for being pessimistic about the future of science on an issue; indeed, there are provable limits (methodological limits, uncertainty principles, and the like) which bar scientific progress of a certain sort in a certain direction. But ID is not pessimistic about the progress of science in this way; it is structured as a set of arguments for the conclusion that we can be quite optimistic about the future of science, even given certain alleged failures in current theory, because the future of science is ID. That is, IDers see themselves as proposing a new direction of progress for a science otherwise doomed to fail; and this is not a giving up on science, but a claim that science is capable of accommodating ID as a biological theory. And it is this that is the issue. IDers are not arguing from an unfillable gap to a designer that bridges the gap; they are arguing that ID is the best way to fill the gap. (I think in doing so they are usually committing a version of the 'God of the gaps' argument, in that they are often confusing primary-causation and secondary-causation issues; but this does not come from a pessimism about science, but from a scientistic attitude.)
2) It has to be understood that there are lots of different design arguments, and they fall into rather distinct kinds. The most obvious distinction is between problem of evil arguments (or most of them, anyway), which argue from a lack of design or a misdesign, and arguments that pick design, simply speaking, as their starting point. But even in the latter category there are many different types, not all of which are put forward by IDers, and not all of which are in opposition to any particular evolutionary theory. Even among the design arguments put forward by IDers there are different versions. For instance, Michael Behe's argument in Darwin's Black Box seems to be a natural inference argument. It takes the design inference as a default inference, based on the way we naturally think. Default inferences can be nullified by alternative positions that essentially say "The default is off". I will call these defeating suppositions. Behe's argument is essentially that (because of irreducible complexity) the defeating supposition for the default inference fails for certain biological machines and cascades; and therefore we are rationally required in these particular cases to fall back to the default inference. (This is very similar to Paley's argument, which is, would argue - against virtually unanimous opinion, it must be said - also a natural inference argument.) Now, one irony here is that many biologists' view of the situation is structured exactly in this way; indeed, it is so pervasive that even vehemently anti-ID people like Dawkins accept this view, and even those who cannot consistently do so often slide back into saying or implying something that would require this view to be true. If this is the case, then the whole issue at stake is simply whether Behe is right about the failure of the defeating supposition; and the most successful refutation will be to show in each particular case that the defeating supposition still holds (the defeating supposition can fill the apparent gap, thus not requiring us to appeal to design to fill it). In other words, Behe's argument is weakest at his criticism of Darwinism; it is surprisingly difficult to find anyone who consistently even rejects Behe's actual design inference, much less rejects for good reason. This is different from William Dembski's argument in No Free Lunch, which is not a natural inference argument (Dembski I think actually has more than one; I am speaking of the primary one, the Explanatory Filter Argument). In Dembski's argument, we must first eliminate a 'chance' class of events and a 'law' class of events (which are not defeating suppositions but standing defaults); and then, when we have done so, we reach a 'design' class of events. I put the words in quotes because the classification itself is not extraordinary; and a great many opponents of ID have wasted many words on the classification to no good end. One may classify things in whatever way is convenient; the question is not how we are going to divide events into classes but what we are going to be able to infer once we have done so. On Dembski's argument, the inference from the 'design' class of events to a designing cause is sanctioned not by natural defaults in reasoning but by induction, and it is here, at the design inference itself, that Dembski's argument is weakest, since Dembski spends very little time and effort on this aspect of his argument. If the two arguments are conflated, as they usually are, the only possible result is muddled confusion; because they are weak and strong at different places.
Neither of these arguments, however, exhausts the field of possible philosophical design arguments (or comes anywhere near doing so). It's also noteworthy that in their bare outlines, neither of these arguments gives particular warrant to thinking ID particularly scientific. That is, it is entirely possible to deny that design considerations are scientific and to accept these arguments; on such a view, we will simply have come to another limit on a particular direction of scientific progress, where other, non-scientific considerations kick in. (There is also nothing in these arguments that makes it inconsistent to think of ID as scientific; the whole issue of whether ID is scientific or not is not an issue of the nature of ID arguments themselves, but of the role played by design considerations as compared to the role played by scientific considerations.) If they work, that is not a sufficient reason to consider them useful for scientific biology; and if they don't, it won't be because they are unscientific. That is, this is so if we are just considering the basic formats of the arguments, and not necessarily claims made for them.
What is clear, however, is that the conclusion of the argument, while somewhat interesting if true, is actually next door to nothing; all design arguments face the major limits Hume notes in his discussion of the analogical design argument, because the analogical design argument is the most general form of design inference. That is, even if we accept them as true, they may give even less than the most modest intelligent design theorist actually claims for them. This is because IDers tend to see their conclusions as at the same level as evolutionary theory; evolutionary theory, however, is a secondary-causation issue, and tells us nothing about ultimate causes (one could, of course, hold that the only 'ultimate causes' are secondary causes, but this requires higher-level assumptions). IDers, to their great credit, admit this; but they don't typically act on it. Suppose an ID argument is sound: what do we then have? We have a bare conclusion we can do very little with and a promissory note for possible further research once we get a clearer notion of all the things that are designed. ID is in this sense like eliminative materialism: if an eliminative materialist were to claim scientific status for e.m., the 'scientific' status would actually just be a promise that one day, many decades in the future, science will have so radically changed that psychology will collapse into neuroscience, and concepts like 'belief' and 'mind' and 'thought' will be totally replaced. Likewise, at the most optimistic assessment, the 'scientific' status of ID is a promise that, one day, many decades in the future, science will have so radically changed that we will be able to do something with this conclusion that we have drawn through an ID argument. But even supposing this is all true, in the meantime it leaves us with just a very vague gesture toward what may actually be just a science-fiction fantasy.
3) My pet peeves in the whole discussion are (A) when people confuse final cause arguments with design arguments: the two types of arguments have a certain relation (design arguments deal with the peculiarities of a very specialized form of final cause), but the two cannot be equated. Aquinas's Fifth Way is not a design argument; it is an argument about the possibility of anything's being an efficient cause at all. 'Final cause' in the scholastic sense does not mean what it came to mean later. (B) when people assume that all design arguments are of exactly the same format (which I've already dealt with).
I'll end before this slides too much into a rant again.